[This Transcirpt is Unedited]
Hubert H. Humphrey Building
Room 305A
330 Independence Avenue, SW
Washington, D.C.
DR. MAYS: Let's get started. We'll wait a bit on introductions. I do want to get started. I'm going to talk a little bit about the bias of what I'm going to suggest and we can move on from there.
Looking at the agenda in terms of the item that's the review of the hearing, there are three things we need to do. One, spend some time talking about the consistent issues we heard coming up and trying to flush those out deciding what are the themes we want to focus on.
In terms of plan of action, there it's kind of like, I think we have two issues. One, continue to have hearings and do something at the end or at the end of each of the hearings, decide to move forth. I'm going to suggest that we consider moving ahead and pulling together information and making recommendations.
We had a fairly large group of people there who, Gracie will attest to this, that keep e-mailing us that they want information and to know what's going on. One of the things I heard when I joined this group, is that sometimes people work on stuff and then there's the audience that moves further and further away.
This is one of those times where it would behoove us for many reasons to seize the opportunity if we think that we are on solid ground about making some recommendations, I would suggest we do that. That's part of what I'm going to suggest. We can be thinking about it.
In terms of future hearings, I'll put on the table what I think about it and then we should have a discussion. We probably should do vital statistics in states because their issues are different than the fed issues are. The Native American and American Indian community, I think, definitely needs its own time. There have been some approaches and requests that we try and do some focus on the API population. Under each of those, those are like little a-b-c=s and all that kind of stuff, in terms of outlining it further, that I would suggest.
I think most people are here and we'll do introductions and then move into the issue of discussing the hearing itself.
DR. NEWACHECK: Paul Newacheck, University of California-San Francisco.
DR. LENGERICH: Gene Lenerich, Penn State University.
MS. BREEN: Nancy Breen. I'm an economist at the National Cancer Institute.
DR. FRIEDMAN: Dan Friedman, Massachusetts Department of Public Health.
MS. HEURTIN-ROBERTS: Suzanne Heurtin-Roberts, National Cancer Institute, member of the Committee.
MS. COLTIN: Kathryn Coltin, Harvard Pilgrim Healthcare, member of the subcommittee.
DR. STARFIELD: Barbara Starfield, Johns Hopkins University, member of the committee and sub-Committee.
MR. HITCHCOCK: Dale Hitchcock, ASPE. Staff to the committee.
DR. NEWACHECK: I want to raise one other issue before we begin discussion.
MS. GREENBERG: I'm Marjorie Greenberg from the National Center for Health Statistics, Disease Control and Prevention, and Executive Secretary for the Committee.
DR. MAYS: Vickie Mays, the Chair of the Subcommittee on Populations and I'm at the University of California-Los Angeles. You wanted to add something.
DR. NEWACHECK: It's an issue that we talked briefly about at the hearing. It might be of interest to this committee and a service to the government. That is at some point, I think it would be helpful to have a discussion of the role of contextual factors in surveys.
We talked a little bit about this at our hearing and it's an important area that we're not doing a very good job on now, but there is huge amount of potential and interest in it at NCHS, and NIH, in various research groups that Barbara and I are involved in for the American Academy of Pediatrics.
I think this subcommittee could play an important role in looking into this issue and perhaps have a hearing, it doesn=t have to be an all day thing B maybe a two hour thing or something, and then perhaps writing a letter, if we all agreed, to the Secretary saying how important this is. It would help encourage that process at NCHS and elsewhere, including those factors in surveys.
DR. MAYS: I have things that were question marks. It came up, someone actually raised it at the hearing, but then again since I have been here in DC, it is this subject of articles in the Washington Post, the diversity that also exists in the African-American, the black population, I should not call it the African-American, it is the black population, and the extent to which that might be also an issue in particular states like Massachusetts, New York, California, et cetera.
I know when that special issue of HAPH came out and there was some discussion of it then and then people came and said that everything we see says that you aren't interested in the diversity there, when people were at the table. I'll just put that on the table since we're also trying to work with the diversity also of the other groups.
MS. BREEN: I don't want to get into the discussion now, but in terms of Paul's comment about contextual factors, I recommend that we broaden that discussion to go beyond just survey data into claims data, state-based data. Actually, that is often a better source than a survey is for the contextual factors. That is something we have not done enough of in this country and we've got more potential for it.
DR. FRIEDMAN: Including in there some discussion about multi-level.
DR. LENGERICH: That is the question. We are talking about multi-level contextual variables.
DR. FRIEDMAN: The idea being, most of our population based surveys, we collect data at the level of the individual family, but nothing really about the neighborhood, nothing about the broader environment that people live in. There is a lot of interest in using the interviewers who go out in the field, for example, to actually make observations about the communities that people live in so that you would be able to do multi-level analysis of not only the individual, but the family, the neighborhood, the community, and pull in some of the factors that may be affecting health.
MS. BREEN: I wasn=t aware of that. Are they trained to do that? I have the idea in my mind, in the old days, when the Census interviewers used to decide what your race was.
DR. NEWACHECK: They do it in some surveys like The National Longitudinal Survey.
MS. BREEN: Are they trained in sociology ?
DR. NEWACHECK: No, I don=t think so. They have a checklist like are their abandoned vehicles on this block? Are there broken windows? Are there barred windows? Things that capture some feeling. It's objective. The problem is how do you do that in a rural area? It's easy to think of what you would measure in an urban area. It's a really important topic area. We're missing that when we think about health. We don't measure those things and can't look at them in models.
DR. STARFIELD: You mentioned the issue of variables other than survey data. There you talk about the area files and things like that.
MS. BREEN: The study on Glasgow with data on transportation and pollution and water resources and transportation. It's available and it's critical to understanding the infrastructure within which people live their lives.
DR. MAYS: The thing to think about is to really push this issue of geocoding and linkages. That was part of what we had at the hearing. To some extent, if I had to say, and lets just start the discussion about the hearing, what I was surprised about was the extent to which I don't think the surveys talk to each other very much.
They didn't realize, for example, things that were in various surveys. Part of that was that NCHS knows what's going on inside NCHS, but then we had a much broader set of individuals at the table. It would be useful if one of our recommendations is some overall planning. They have all these PHS groups for different things B I don=t know if you want to suggest another one, but maybe there needs to be one on surveys or something.
MS. GREENBERG: There is a workgroup of the Data Council on integration of surveys.
DR. MAYS: That's a little different.
MS. GREENBERG: The extent to which it gets heavily into content. It has on the drug abuse and health insurance. That doesn't mean that you can't make a recommendation that that group or some subgroup of it shouldn't do something. That should be within the purview of the Data Council. There is a workgroup on survey.
MS. BREEN: There was an IOM report a couple of years ago, on data integration. It was subsequent to that that they decided to use the sample from the NHIS for the MEPS which was an advance.
MS. GREENBERG: I think we heard B there has been discussion about the HIS sample has been used for HANES, it's now not going to be used because of issues of oversampling. Whatever decisions that are made need to be looked at.
MS. BREEN: I call it orchestration of surveys. We have date, we have national surveys, and right now, even the NHIS don't yield comparable data on what they're measuring, for various reasons. We need surveys that will do that. We can look at Minnesota and the nation and we can not have to adjust or we can know how to adjust.
DR. MAYS: Why don't we open it up and talk about the hearing and what you see the theme?
MS. HEURTIN-ROBERTS: I just wanted to add to the thought that I We started focusing on having not a large enough sample of certain groups. I want us to remember and think about the issue of cultural validity and tailoring of specific items and whether they're piloted or not appropriately. Some surveys do this much better than others and some of them don't do it at all. I question the validity of the data that we're getting for some groups.
DR. MAYS: Can you say specifically what groups do you worry about?
MS. HEURTIN-ROBERTS: Certain issues with language for Hispanics or Asians. There is some work that I had done in the ECA studies by NIMH. Even for African-Americans, many of the items there were taken from DSM and really weren't appropriate to the sources. That needs far greater attention then we=ve given it.
MR. HITCHCOCK: In part, it's a language issue.
MS. HEURTIN-ROBERTS: It's a cognitive issue as well.
MR. HITCHCOCK: If you take he DIS and ask Catholics if they ever felt guilty, for instance, and ask some other group and try to translate the idea of feeling "blue" to Hispanics. They don=t know what you are talking about.
DR. NEWACHECK: We have two things on the table.
We're looking at cultural appropriateness for children in different states. We have examples of activities that children engage in, i.e. Boy Scouts, and this and that People from New Mexico said these things don't make sense. You have to add working on the ranch and getting firewood. That doesn=t make sense in the urban area but it does make sense in a rural indian area. It's the cultural appropriateness of questions and testing them. There is the language translation issue and when they work right when you translate them for people who don't speak issue.
DR. MAYS: Even for people who speak English, sometimes the way we ask the question is so foreign. You can take groups who speak English, but feeling "blue", the notion. Mental health there are several things -- they're very personalized. When you ask the question in that way, they're like it doesn't register.
The whole concept of how people respond in terms of their emotions has a lot of cultural overlay to it. Which is why when we did ECA in the '80s, we ended up with African-Americans having such low rates of diagnosed mental disorders and with something like phobias are off the mark. There is really terminology thing we have to think about.
MS. BREEN In terms of following that up with a recommendation or plans of action, we should take advantage of some of the existing surveys that might have the flexibility to do some experimentation in those areas.
Paul may be working on one. I'm working with a group at UCLA, The California Health Energy Survey, which is modeled after the NHIS with a strong basis for comparison with the national survey. It's being administered in six languages, including English. They're working very hard to administer it in a culturally appropriate way.
For example, they're training interviewers to interview Native Americans who don't walk on words like we do. I might interrupt. This is not done in the Native community. You need to give them more time to answer the question and not just go on.
As you said, there are culturally appropriate ways to administer questionnaires even in English. We have a cognitive psychologist in our program who was hired to just this and is very interested. This issue of how to administer questions to different groups, language and cognitive, are very hot issues.
In his field, he would love to work in this area. There are probably others and there is a meeting coming up where they're going to be working on this. There are cognitive psychologists who would like to do this.
We may be able to make some recommendations and put some money in the pot. NCI is interested in improving the quality of these surveys to get better data and food data is one of the worst. That's one of the hardest.
DR. MAYS: The nutrition questions they were having problems with. I think also the Office of Behavioral and Social Science Research was interested. I don't know that there is money. I also know that the Office of Research on Women is interested. As a matter of fact, they have a group right now that is working on some of these issues.
I do think, again, I kind of made notes, that if we wrote a letter, it would go to the Secretary, but I think it should go to to NIH and NCHS and OMB to highlight issues in the purview of those groups. It is one thing for the Secretary to act on things, but other things where NIH can move ahead on some things.
MS. BREEN: In the federal bureaucracy, things get done if you can take it to the person who's interested who is interested in doing it rather than the highest level person.
MS. COLTIN: Another theme area that came up was a need for guidance around analytic techniques to use. I Paul raised an example. Even something as simple as when to adjust and when to stratify. There are instances when you're stuck with adjustment because of the sample size problems. In fact, that's not really what you ought to be doing with the data. That's an issue.
DR. STARFIELD: That's the concept, not really a methods issue. It is what=s your question and what is the best way to answer it?
DR. MAYS: There is a group, Ken Kopple=s group, that some of us have gone to those meetings. They actually are trying to help by writing a guidebook. When you're trying to look at health disparities, what statistics to use to understand, like for example, if you are doing two ethnic groups, three ethnic groups, it really gets down to what should you do.
DR. STARFIELD: It's not a methods issue. It's the concept. It's what it is you want to find out. If you want to control for race, you're going in with a certain idea of what you think race is. It's the idea you go in with, not just plugging something in.
MS. BREEN: I have another issue that is a little bit a long the lines of this one and I am debating on whether or not it fits. Suzanne and I had the opportunity to hear Ken Koppel, he came over to our Health Disparity Working Group. We had a pretty fascinating and illuminating discussion with him because we have been calling ourselves the Health Disparities Working Group. Why? We're interested in reducing health disparities and improving health. It became very clear by the end of the discussion that health disparities means reducing health disparities regardless of whether it improves health.
DR. STARFIELD: It might reduce overall health. It is not so much reducing it but it not might improve it.
MS. BREEN: We concluded from that meeting is that we need twin measures for measuring the Healthy People objectives. We need to make sure we're looking at progress in health and reduction in health disparities.
DR. STARFIELD: That's a value statement. If you look at the international literature, there is this argument all the time. Are we going to reduce the disparities in Bolivia or are we going to improve average health.
That is what the debate was on the World Health Report was. MS. BREEN: I would argue we take a value statement here. I think it's important that we're working to improve health and not just reduce disparities because that is a statistical issue.
DR. MAYS: You can reduce something.
MS. BREEN: You can reduce the difference between them.
DR. MAYS: Everyone can get worse but in a report, you are able to say you reduced. Therefore, you have moved along the path towards the target goal that was in Healthy People 2010. If you look at what the overall issue is, that's what the goal is for Healthy People 2010.
MS. GREENBERG: 2010 has various things. It should be this, it should be that, it should be the others, because in order to improve health, you want no more than A, B, or C. And then you also want to reduce the disparities.
MS. BREEN: If it's doing both, we're okay.
DR. STARFIELD: It's not clear you can do both. That's the issue. It's not obvious you can do both,.
DR. MAYS: You have targets.
DR. LENGERICH: It's not required that you do both. You can make advance in one without the other. I guess I would like to raise the question then, I am not real tied in with Healthy People 2010, but the whole issue of measuring disparity, is there controversy in that? I think there would be, again, when you are looking at differences, ratios, or those kinds of issues because they were the logic question and is central to our theme. I would like to ask that question? Is that being addressed adequately? We did not hear about it in our survey.
DR. STARFIELD: Look what's happening with infant mortality everywhere in the world that's decreased a lot in all population groups. A lot of people would say that's terrific and it doesn't matter about disparities because everybody moved up.
DR. LENGERICH: How do you measure the disparity? Is it a ratio or a difference that you're trying to go for?
DR. MAYS: That's what this group is doing.
DR. LENGERICH: I would like to put that on the table as something to bring and have the subcommittee hear about as well.
DR. MAYS: Maybe what would also be good, in terms of recommendations, is to think about a call for clarity on the issue of the reduction of disparities and the improvement of health. This is why we should not try to do based on this hearing, but hold this one until we have Ken=s ideas. I think we will be on stronger ground. CDC is leading this effort, so we don't want to make statements and find out that things are a little different than we thought.
DR. NEWACHECK: It would be helpful to stick with what we heard at the hearing and what we're talking about already in terms of recommendations we could put in now and then talk about are we actually going to do a report with recommendations or letter?
Two recommendations worth thinking about is one, ensuring all of our major federal surveys include oversamples of important minority populations. The MCDS survey does not do any oversampling at all. I was shocked. It seemed like it wasn=t even thought about there. We're spending millions and millions of dollars on that survey and we have no oversamples of minority populations at all.
The other is like HANES which is half oversample now. The wite population is a minority in that one. We have these extremes for no apparent reason. Doing some kind of careful review of who does oversampling and who doesn't and ensuring there is minimum numbers in each survey, I think would be important. That was one thing.
The other thing B and this is sort of an analytic issue, is that the data that we have available in our surveys allows us, to at to document but not understand disparities. We don't have indicators that would allow us to go beyond describing disparities. Going further to be able to improve survey design in ways that allow us to have the variables that we need to understand the causes behind disparities is important.
DR. MAYS: Let me add one thing thaqt came up during the course of getting speakers and that is about the oversample. One person pointed out to me, who actually used to work at NCHS, that you have to be very careful about oversample because what happens in oversample, specifically for minorities, is we get a very skewed group. When you do that, the group that you're getting may not be as representative as the statements that you want to make.
What they began to raise were things like you had to go back to the survey and see where they get them. Typically, oversampling is done in large cities. You're weighting heavily toward a very urban type of individual. I was hoping that we would get more discussion about some of those issues.
But people didn't get as technical as I would have liked. That may be one that it would be nice to bring someone in and have them talk about it so that we're clear, if we call for that, what the pros and cons were. I used to think that was great and then this person said, no, you need to think about what it means and if, for some populations, the issue of an oversample, if it gives an urban group, that may not be ashelpful in terms of Native Americans. They went through some thinking and I thought it was very interesting.
DR. NEWACHECK: Part of that, too, that it does reduce the power of the survey to do oversampling. That is, overall survey end decreases effectively when you do oversampling. Your ability to generalize to the whole US population diminishes. So there are costs to is as well as benefits, and potentially the benefits may be less than we think. So I think having a recommendation, if we agree with it, with a discussion of the strengths and weaknesses of that approach or the cost and benefit of it would be a good value.
DR. MAYS: We may need to get a little more information on it then.
DR. LENGERICH: One thing that came out of that meeting was that the committee made a recommendation about increasing sample size. I was looking through this document here, it is listed in the 1998 report on minority statistics. The question is not really to oversample, but what happened to this recommendation?
MS. GREENBERG: Money.
MS. BREEN: Several times it was brought up, and when I was reading through the report of the committee that Carolyn Clancy headed B there was awoman named Perel, who had written the report, she had brought it to our attention B one of the recommendations which hasn't been widely implemented is to collect information on a person's first language. That is probably something that we would want to recommend that should be routinely asked in all the federal surveys.
DR. MAYS: There is some NIH group that has a very nice discussion about language and language proficiency. It's NICHD. We can see if we can get copies of that for the committee and then piggyback on that. The discussion is very nice.
MS. HEURTIN-ROBERTS: Related to what Paul was saying on the issue of oversampling and not having a large enough sample size, one of the strategies that we might want to discuss further, or recommendations that we would like to have people consider, is the idea of doing smaller, more regional samples. In which case you don't have to oversample. The sample is there. The question is then whether it is generalizable across the country. But still, that's one approach that can be really useful.
MR. HITCHCOCK: It doesn=t really need to be generalizable across the country. There are some places with 80 percent or higher of the population that you're looking at. We also talked about doing those studies on a regular planned basis, get into the budget would be one aspect. Have some money in there that would be available to study a certain aspect.
DR. LENGERICH: We've approached this from national surveys and the question is, what is the place for the state data collection process and our recommendations to those? If you're advocating for more regional, that almost becomes a collection of several state opportunities.
MR. HITCHCOCK: I don't think we're talking about region. That could be an administrative concept the way it gets used in our surveys. If we're looking at a specific group, Asian Pacific Islanders, I think you're wasting your time doing a national study to gather data. If you wanted to do a study of Cajuns, would you do a national study?
DR. LENGERICH: No, but just as the California Health Survey is doing, they're taking the matter into their own hands and are going to address the populations within their state. I guess we have to balance this question of what's the role of a national survey versus what is the role of some sort of subnational survey.
MR. HITCHCOCK: One of the things that most concerns us is the expense of doing these state surveys. I don't know what size survey we're talking about in the California HIS?
DR. MAYS: 55,000.
MR. HITCHCOCK: You need a national size survey to get the decision.
DR. MAYS: Remember one thing. We'll have a hearing. State issues are a little different than the federal issues. It might be hard at this point to make state comments. We ought to hear about vital statistics, we ought to hear about state issues because they come at some of the issues very differently and they want the diversity, et ceterea.
DR. NEWACHECK: We do have surveys federal surveys with state components. We shouldn't neglect those.
MS. BREEN: The federal interest in the state surveys have to do with getting better national data which this committee is very interested in. I thought that Reinard put it nicely when he said parallel survey, where the national survey seemed the best way to go in order to get information on some things, a more local survey was better for other things depending on maybe the population or even the condition or utilization that you want to look at B HMO penetration in different parts of the country B various things like that.
So I think, in that sense, it's of interest to us. It makes sense to table it for the purpose of looking at state surveys perhaps, but I think it is important to keep it in the broader context of orchestrating data so we end up with a better understanding of what's going on in this country.
DR. FRIEDMAN: I think the notion of a state-based hearing makes sense. I suggest we put it in the context B and i think this is consistent with what you are saying, not so much of vital statistics and other information systems, but other state public health information systems and include in surveys, surveillance, administrative data. Most states we think of vitals as one particular area in the expanding(?) system, and we try not to think of it as something in part.
MS. COLTIN: One other thing that came up was the topical surveys that periodically, to the NHIS in particular, that are there are different topical surveys that are done at different intervals. When we talk about oversampling, it's also important to address it around topical surveys, because it may be more important for some topical surveys than for others as well.
MR. HITCHCOCK: They call them supplements.
MS. COLTIN: If you were doing one on diabetes, for instance, you would want to oversample. One thing we didn=t hear about in the hearing were the provider surveys done at NCHS. There was a meeting about two years ago about the provider surveys. There were a number of recommendations. It was a panel that NCHS put together. One recommendation was that they have rotating topical modules as well.
For instance, in any given year, if a patient came in and had diabetes and that was the topic, you asked additional questions to get information about that. I don't know if there has been any coordination such that if there were a topical module for NHIS in a given year that it would be the same topic in NAMSES so that you could learn something from multiple sides. So if you were going to do cancer, that=s next for the NHIS or whatever, it would be interesting to pick up information related to that topic in the provider surveys.
DR. MAYS: Do you know if that particular panel was part of the data user's meeting?
MS. GREENBERG: It was a group that's been meeting over a period of years.
MS. BREEN: This meeting was a one-time meeting. I was there.
MS. GREENBERG: Is that the meeting that Linda Demlo pulled together? She's the branch chief for those surveys, isn;t she?
MS. BREEN: They came to NCI at one point and talked to our program. Amy Berstein also works there. She is a health services researcher. They were interested in looking at episodes of illness. Right now, we don't find those surveys very useful because they're event based. You have no knowledge of what individual had this and you know nothing about follow up or precedent. It's just right there in time and some number of people had "x" from a doctor's point of view.
It provides some information, but not as much as given the resources that are used. We had a discussion about improving those data. I don=t know if it was related to these other things or not, but it hasn't gone anywhere since. I checked recently and nothing has happened. I don't know where that's going, but it would be good to beef those up.
MS. COLTIN: There were interesting ideas thrown out that had to do with if, in fact, you had a supplement, that the supplement could collect information, at least some historical information about the patient as well as what was happening today. It would have to be minimal because of the nature of the survey. But, you could get a little bit more than what you would in a single point in time.
The other possibility was to use it as a way to identify a sampling thing for the patient, that the doctor could hand to the patient a survey and ask if they would be willing to respond and mail it back in so that you were able to identify these patients in the context of that survey, but actually get more information directly from the patient as well. Just piggybacking it.
DR. MAYS: Maybe what we can try and do is to see if we can find something about that. I have a feeling that the way we are going with some of these recommendations B some are sooner, some are later, so some will require just gathering more information so that we don't jump off a cliff and they tell us they've already done that.
MS. COLTIN: The issue that came out of that panel was an issue that would be very relevant to this topic which was sample size. The National Hospital Discharge Survey has this enormous sample and the National Ambulatory Medical Survey has this really inadequate sample and there isn't a lot you can do with looking at subpopulations with the NAMSES. And yet with the NHDS you now will be able to get more information from administrative data, do you even need that survey as much and you did before? Shouldn=t the resources should be shifted away from that and more towards the ambulatory.
MS. BREEN: Another thing we didn't talk about that we should probably keep in mind and talk about for surveys would be response rates, which are falling even for in-person surveys and people really don't know how to raise them for telephone surveys. I was reminded, when you mentioned the mail survey, because the response rate could be dreadfulB what you were talking about there, which would undermine its representativeness.
DR. STARFIELD: One of the disadvantages of being on this committee so long is that you remember the issues of the oversampling. It's all been talked about before. There are a few new things we are talking about B the disparities is new and the context is new. We should think about whether we want to focus on new things and not revisit the things we weren't able to resolve before unless we've got a brilliant idea about how to solve it.
DR. MAYS: I think there is one thing at least. Debbie Ashley helped to go through the archives. You have another piece of it in there. I'm not always convinced B oh, there is Debbie down there. It is like we talked about it before but I don=t know that we got someplace. DR. NEWACHECK: It's like national health insurance. It comes up every 20 minutes(?) and you talk about it and hope that it's going to happen this time because there is some interest. There is a little window now.
MS. BREEN: With Tommy Thompson and George Bush, don't worry, Barbara.
DR. MAYS: Someone just told me that the Secretary is very interested in health disparities again.
MS. HEURTIN-ROBERTS: At the cultural cancer council.
DR. MAYS: Some of my federal colleagues also requested something in some of these reports some reports. They wanted it documented. I do think issue of there is a window, and we can dress this window up and say we sent this to you in 1988 or something like that and we sent it again in and 1994? It may help. I think that those things that we can highlight, given that we want to send it to a couple of different places, we should do especially if already have the hearing and the recommendation, let's pull them out.
MS. HEURTIN-ROBERTS: I get the impression that some places, people are ready to act, but if they have something to back up their actions, they'll be able to do it. They need some evidence to make it possible.
DR. MAYS: I don't want to try to do everything.
DR. STARFIELD: We won't pick on the things that we worked on and didn't solve and is no reason to think we're going to solve it now.
DR. MAYS: That's where history will help us. If you say, look, we stood in the Secretary's face on this and it didn't go anywhere. When we amass all these recommendations, what's going to fall out is that some will have better evidence that will allow us to move ahead than others.
Those are ones that we have great passion about, but we can't make a great case for, will either be dealt with at a subsequent hearing or we need to say that. That's nice was who was in the audience. There were a lot of professional organizations, that if they got wind of this, would be very happy to help.
Can I ask a question about the response rate issue? Is it correct that all the federal surveys have to have a certain response rate?
MS. BREEN: OMB requires a certain response rate and if you want to get your survey through OMB, you tell them you guarantee an 80 percent rate B you lie.
DR. MAYS: They don't check and come back and say you don't get the money next year if you have less than that.
MS. HEURTIN-ROBERTS: I want to say something about response rates. This is similar a problem that we see in clinical trials, community trials and getting participation. One strategy that's being talked about more and more is a participatory approach, a paticipation of the various stakeholders involved. Somebody from the floor at the hearings -- and I don't know how you could bring that together. But that's one strategy that is being used..
DR. MAYS: CDC has a whole little branch that they devoted to this and there is more discussion of it..
MS. HEURTIN-ROBERTS: We are talking about it.
DR. LENGERICH: NIH has calls out now for programs and projects.
MS. HEURTIN-ROBERTS: It's almost an oxymoron B a participatory national survey B because when we do national surveys, we think of top down in structuring it hierarchically. Having a grass-roots participation will be difficult to pull off.
DR. MAYS: I think CHIS can give you some sense of how to do that. It is more expensive, but CHIS is doing about this in the sens eof when it started and how it formed its technical advisory groups. Now it=sliving up to the responsibility by going back and having the dispersion of data with various specific community groups as well as how it's putting the data on its web site.
MS. HEURTIN-ROBERTS: It also relates back to the issue of response rates and participation and the appropriateness question that we talked about earlier.
DR. MAYS: If there are any more issues B and we have a lot of them B but if there is anything specific to the hearing, because we are also generating other things we have passions about B well maybe not passions, but anything specific to the hearing that you want to make sure we get on the docket?
MS. BREEN: One other point since you're talking about beefing up these surveys in a number of ways is something that Reinard brought up and that is to professionalize the interview staff. A lot of the health surveys have professional staff. At least the ones I met in the test runs for the NHIS. Many of them had been working in the field on that survey for a decade or more and they were very professional. The Census, for example, every ten years, gets together millions of people, trains them quickly, and sends them out into the field. So I think we end up with a combination of those things.
DR. NEWACHECK: The bulk of the national surveys, like MEPS, NHIS, HANES B they're all professionals. NHIS uses trained Census Bureau employees who are full-time workers, and that's all they do. They're not like the students earning money on the side or something like that. I think that only occurs at the decennial censuswhere you need a whole army of people to be out there on April 1st to do it. I don't think that happens in any of the NCHS surveys.
DR. MAYS: When you move outside of NCHS to the NIH surveys, I don't know if it's proper to say they are investigator initiated. For example, the health and retirement survey, some of those. They have less control. I think he was actually referring B he kind of made a comment about NIH B I remember cringing and thinking, oh, no, please don=t start imposing on my surveys B but there are some surveys that are national, like the National Drug Abuse Survey that is done by Michigan, the Health and Retirement Survey. There are probably about five or so of those, and I think that is part of what was coming up. Their staff may not necessarily be at the same level. That may be one we need to explore.
Let's talk about a plan of action.
I think that now that you have generated what you want to do, there are several options to think about. Each one has pluses and minuses to it. One is that there could be a letter, kind of succinct and to the point, that says these recommendations come based on a hearing and anything welse we can pull out that we have to substantiate it.
You're looking at 3-4 pages and you move on. There can be a report from the hearing in which case we abstract from the testimony and use some of the data that we were given by individuals. Did we get everybody's presentation in some form or another?
So we have that that can bolster what we're going to say. We have things in archives and the web and so on. That will take a bit longer. There are combinations which is, do a report to one person whereas with someone else you might do something shorter.
I would say we want to think about what is the goal? What kind of impact do we want to have, and where do we want to have that impact? My suggestion is we think about something that goes to the Secretary, we think about something that goes to OMB since they are involved in guidance about race and ethnicity. We think about NIH, we think about NCHS, and we think about the Agency for Health Care Quality Research. I picked those because NCHS maintains a certain number of these surveys. The Agency for Health Quality Research maintains certain surveys, and NIH does.
MS. BREEN: You might want to add CMS.
DR. NEWACHECK: SAMHSA, too, probably.
DR. MAYS: We didn't actually hear from any of the SAMSHA surveys. SAMSHA came and wanted to be a part of these committees.
MS. GREENBERG: The committee's recommendations always go at a minimum to the Secretary and the co-chairs of the Data Council. You could specifically ask for those, but this is the standard procedure also is I would think you would want whatever you report, than just a letter, to make a presentation at the Data Council where all of these agencies are represented. Jim Scanlon is rthe exec sec of the Data Council and he will send it out to all of the agencies and ask them to respond, comment on the recommendations. I don't know that the committee needs to send it to all of these different agencies. It will be sent to them and they will be asked to respond.
DR. MAYS: We could do a report for NCHS because we have someone from NCHS.
DR. NEWACHECK: A letter would lack punch and wouldn't have much impact, especially if it had six recommendations that kind of came out fo the air. You know, where did these come from, they don=t have any basis. NCVHS is important, but you don=t show is why these are important. We just take it on faith and I don't think it will go anywhere.
I think a report model makes much more sense. It would be based on the hearings. Presumably there would be more than one, but if it even happened now the report could be a 10-page report describing what was learned at the hearings with recommendations. That would have much more impact than simply a letter and give us some sense of accomplishment, too.
MS. HEURTIN-ROBERTS: I was going to say a report is better not only because it has more impact, but it is much more usable. Rather than say go do this, if you have a report you can give guidelines and thought about how it should be done and why. I have a question, how does this fit in when we first started talking about whether our plan of action would be to have hearings, hearings, hearings and a report or hearing and a report, hearing and a report. I think we need to think that through some more, too. What is this report going to encompass? We talked about some more hearings in the future. We need to sort that out.
MS. COLTIN: I would go with the report as well, for two reasons in particular. One is, I think heard that there are a lot of issues. It isn't just race/ethnicity. It's contextual factos and a whole lot of information related to measuring disparities in health care. It seems to me it would be a little different from what's been done in the past, which has been focused on the racial and ethnic recommendations that may or may not have gone as far as would have liked them to.
If we did a report on measuring disparities and health care adequacy of federal and state data systems, we'd be able to bring in recommendations and what progress has been made and what needs to be done in line with the race and ethnicity. But we would also be able to bring in the contextual factors, we would be able to bring in income and education and some of the other important factors that we heard about in the hearings. It would be newer and fresher if it were framed that way than another report from this committee about race and ethnicity.
DR. LENGERICH: I support that and particularly, as the comments go on I think they are getting more specific. I'm not ready to support a report after only that hearing. If we want a report, there has to be more addressing things like contextual variables and addressing issues around measurement of disparities, Healthy People 2010, what is going to be done and some of the other issues that we've raised.
MS. BREEN: I want to vote for the letter not on its own, but as a nice addition to this. I think it is also useful to ask people directly to do things or to say this is what we need as opposed to simply making recommendations. It's all too easy to say someone, somewhere will do that and it won't be us.
I think this sounds like a great report. I like the way Kathy framed it with the nice exec summary at the beginning for easy reading, and a letter laying what we think some of the top priorities are and actually asking people if they would be willing to participate, would be a nice approach, an effective approach.
MS. GREENBERG: A kinf of joint effort could be a single report and then send it to the different agencies with a tailored letter.
DR. NEWACHECK: I agree with what Kathy said. We can make this a more interesting report if we integrate things like that into it so there is sort of new stuff as well as old stuff. But I also feel that not only in this context, but more broadly, because we need this for lots of social research and health research, not just for disparities analysis. It reallywould help us to understand health in general as opposed to the narrow, individualistic approach that we take now.
Whatever we do in that area, I would hope that as a subcommittee, we could do a separate letter or report or something on contextual factors. This is really moving away from the current focus B which we have one focus right now B to a measurement of race and ethnicity in surveys, but as a secondary focus of the subcommittee=s efforts, we can do something that will make a contribution there as well. It would help this effort and would be more broad in terms of value.
DR. MAYS: Now that I have heard this, let me put some things back out to you so we can see where we want to go. One is, Suzanne=s question about hearing/report or hearing/hearing/hearing and report, we should visit. The reason I am saying that is for several reasons. One is to have a sense of when is enough enough? How big are we going to get because we are widening this B and our constituency. I want to say we had a lot of people in that room and we have a health disparities agenda and we want to have relevance. So I want us to thhink about how much we want to do at what time.
That is number one. Number two, I want us to think about if we start growing these areas, then we have to grow some way to accomplish them. If we start making this broader than health disparities, which is what the hearing was on, and start talking about social determinants then we're probably going to need to join in with Dan's group.
You've done part of this work. I don=t think it means we should go off and do hearings. I think it means we need to have a relationship with you all to get that piece done, because you are almost finished. We have to think about everything does not necessarily have to be just us. But we'll do a partnership. You thought you would be out of business because you are almost finished, but I think that may be a way to get that done, which would answer the need that Paul has, which I think is really good to think about, to stay focused to some extent on how to do what we need to do and how to do some of these broader things.
DR. STARFIELD: The visions report does the framework and we could look at the framework and say we're going to take out these pieces because we think they are valid.
DR. MAYS: That might be one way to deal with this. Everybody doesn't get the information from that group, right? That might be helpful.
DR. NEWACHECK: Everybody got it today.
DR. MAYS: Okay.
DR. NEWACHECK: Do we have a writer?
DR. MAYS: We had one who took a great position, but we have another who is an award winning writer who knows this area. I had three people and put the first one that I thought was the best one. Columbia offered her something. We lost her to Columbia. The other is B I have to revisit the document(?) before I put his name out there.
DR. NEWACHECK: One issue that as we move along that persons is not B we do have transcripts and such B but they are missing a lot of the discussion and context.
DR. MAYS: I have their vitas. I have already talked to them. We thought we were going to have the writer at this meeting. Believe me, this just happened. While I was here, I got the phone call.
MS. BREEN: This is really comprehensive. This is a big help.
DR. MAYS: We will get you copies of it.
MS. HEURTIN-ROBERTS: I would just like to second what I think you're saying. I think we do need to come up with a report focusing on health disparities in a reasonable amount of time. I think people in this area are used to hearing a lot of promises and then nothing ever gets turned around. I don't think we want to do that. We have heard stuff over and over again. We heard the themes repeated and I think there is enough to do a focused report.
DR. NEWACHECK: Plus the NCVHS procedures -- it would be at least six months even if we start writing it right now before we have our report, and if we have more hearings it will be longer.
DR. MAYS: I can't function like that. It's like having a baby. It's too long.
DR. STARFIELD: I heard wonderful things about the hearing. If you have another hearing, it will go beyond the focus of this one. I think you ought to do something now.
MS. BREEN: The only problem with not having another hearing is if we're going to talk about state and national in a single report, we really need to hear from the states before putting it together. This was all just national surveys.
DR. STARFIELD: Why not just have a report on the federal?
DR. MAYS: Why not do it in stages and do the national?
MS. BREEN: We could do the federal, that was included.
DR. NEWACHECK: I think your doing a staged set of reports would be a lot to ask. We've got to go through the subcommittee approval process, committee approval process, the Secretary. It would just be an enormous amount of work. If we had one more hearing soon.
DR. MAYS: We can't have a hearing that fast. If you know what it takes to do a hearing like that, we can't have a hearing that fast.
MS. BREEN: How fast can we have a hearing?
DR. MAYS: Six weeks. Where are we now?
DR. NEWACHECK: Let's say it was the end of April where we filled in the gaps that we now have in some of the national surveys we didn't hear about, we cover some of the issues that we didn't cover in depth, like contextual factors and some issues related to response rates or language issues that were only hit tangentially, so we could have solid recommendations around each of those points. We could do a report right after that pretty quickly.
DR. MAYS: What I would say is we have to sit with the staff and see how quickly we can do it. That hearing took us a lot of work. Maybe half of the work is done because we have that kind of framework.
It's not like just finding a person. We had to go through like eight people. They send people and they will just talk and it's worthless. It will require either Suzanne talking to them or I talked to them personally and shaped them to present. Otherwise, they were pulling something out from something and it was going to be not okay.
MS. JACKSON: They organized the two layers and it was just magical how that happened, but it took so much communication with each person and group, finding just the right person and getting that person what provision and statement we needed, and then as time went on, they would still wonder if they were the right people and give us other people. It was amazing.
The core concept for the panel was orchestrated for the September meeting which didn't happen. By the time we got to it, it started to reshape itself based on needs. Vickie found a collection of information from the Data Council. We needed to be able to get to the core of what this subcommittee was trying to get across.
It really took all the time we had to get it together from a presentation to the full committee , to a solid well orchestrated layered presentation that we had.
DR. MAYS: I'm willing to see what we can do, but the only thing I'm going to ask is that you have to give us the time frame in which we can do it in order to get what we need.
DR. LENGERICH: Let's say we didn't do it and did a report right now, do we have enough material from one hearing to write a real report as opposed to a letter? I don=t think we do.
MS. GREENBERG: Do you have a sense of where the holes are so that one could start fairly quickly to put together another hearing?
DR. MAYS: It sounds like what you want would be what are the other national surveys? Let=s go through this. We want contextual.
MS. BREEN: We want state that focuses on things that we don't usually get.
DR. MAYS: Give me an example.
MS. BREEN: American Indians and Alaskan Natives, Asians, high income and high education Blacks and Hispanics, Foreign-bornorn by race, persons who speak other than English, people without health insurance, rural people.
DR. MAYS: That's New York, California, Minnesota. Some of those have the kinds of stuff that we're interested in and have in their offices, either the state office of Offices of Minority Health where they have begun to look at some of this. It's possible.
DR. LENGERICH: What I was thinking of what we were going after in states was the systems that are operated fairly consistently from state to state, such as vital statistics and risk behavior survey. That's what I was thinking about when we were saying state, not sub-national efforts that really targeted a particular race or ethnicity.
MS. BREEN: If we were to pick some, I would pick based on these criteria. I would try to pick states where they were getting at populations that we don't get at very well at the national level or because they have particularly well functioning BRFSS or something that could give us a sense of what we can achieve with that data system and some of the laggards.
DR. LENGERICH: We did hear from BRFSS and we already would have some information about how particular states can address groups within it.
MS. BREEN: I missed the BRFSS presentation. Was that a federal person?
DR. MAYS: What you got was an overview of BRFSS, but what you haven't gotten, because we were actually working on that, was to have a BRFSS user. We thought we wee going to get Massachusetts, but it turned out that that person said they weren't appropriate.
From what I can tell, the state issues you are talking about are going to be a little different from the federal issues we talked about. That needs to be organized a little differently. I talked to some of the people at the states. They began to tell you that they're less interested.
They want as many people as they can get. If it's population based, it's great, but they have got to respond to the needs of their state. California is about the only one that's done such a huge survey if we go beyond just the federal.
MS. BREEN: It is the only one.
DR. NEWACHECK: I'm worried about getting into the state area because it is a whole different can of worms. This committee is usually focused on federal surveys although the ones you mentioned are partially federally supported. I'm worried that we might get distracted if we start going into that area.
There is a lot of other stuff that we haven't really covered well enough to write a report. We talked about things like how to design culturally appropriate questionnaires. It was raised as an issue, but we had no discussion of it.
We had no experts tell us what processes you need to do. Do you need focus groups or testing, cognitive laboratories? We have none of that information really. We don't have the depth. We just know that it is an issue. To have a couple of people talk about that issue from NCHS's cognitive laboratory or some other source.
DR. MAYS: Let me ask a question because I am confused at this point. That is, the difference between depth and the difference between pointing out what needs to be done and the agencies will do it? So I just want to ask the question, because I am getting a little concerned here in terms of this. For each of the issues that we're going to talk about, and again remember I have not done this before, you need to help me understand how much depth are we talking about in order to be able to write the report versus having enough certainty that something is an issue and knowing that there are experts out there to address those issues?
DR. NEWACHECK: That's the difference between a letter and a report. The letter you can just simply have recommendations without supporting them necessarily. The report would be designed to provide the documentation as to why these are important issues and therefore provide a basis for the recommendations.
If you're going to say you're going need to design survey questionnaires in a culturally appropriate manner, we could do that in a letter without any justification of it. In a report, we would probably have a page or half a page that we would talk about why this is important, what kinds of work is being done in the area, and provide a basis for the recommendation.
DR. MAYS: That's an enormous amount of work.
MS. COLTIN: He's not saying that we make a recommendation that we do it, simply that there is evidence to support that it needs to be done and that there are some potential solutions.
DR. MAYS: You're comfortable if you hear from one person, that's enough. There are lots of issues that people are having big fights about.
DR. NEWACHECK: If we have one or two people that we can draw on their comments in the report, in a section on cultural competence, saying Dr. SO-and-SO made the following comments, it would provide the basis for the recommendations that are going to come five pages later. We talked to the experts, it shows we did our homework as opposed to we just thought this was an important issue because we're smart people.
DR. MAYS: Even in a letter, we would say throughout the surveys, ten of the major surveys don't have language. There is something to it.
DR. NEWACHECK: We didn't know, for example, which surveys do any kind of focus groups or testing in minority communities to test the cultural programs. I have no idea.
MS. BREEN: I have an idea of that. There are a lot of people around the table. Your knowledge base times 15, maybe we have that knowledge base and maybe we should test it before we decide to do or not to do.
Back to the argument you were making, I agreed with everything you said and therefore thought we should go ahead and look at the states rather than not look at the states. If you want to know the country, we have to look at the states and localities because that's where we're going to get the contextual data. The neighborhood in a national survey is not very helpful.
You really need a much smaller unit of analysis if you want to get any information on context and these special populations that we are talking about looking at, they are all local, too. So I agree with everything you said, but feel that that's a reason to go after the local stuff.
I also think we should pool our knowledge based on where we think some models are. In addition to the California Health Interview Survey, which is a model, and California also has a BRFSS going which can be compared with that to see what the pros and cons are, and a comparison of costs, and how much more bang are they getting for the buck with CHIS? King County, where Seattle is located, has a very sophisticated data system that they have set up using claims data, local data, survey data and they have a very good handle on social determinants as well as health outcomes and utilization of health services in their areas.
SO I think we've got some models out here that we could draw on and that we could pull together a reasonable hearing. They have a nice web site, too. We could pull tegether a reasonable hearing that would fill in some of these gaps and help us make an even stronger argument. I don't think we're going to get where we want to get in terms of understanding our country unless we do include some of these smaller units, not just the nation.
MS. COLTIN: That raises an issue. I think we should outline the report and then design the hearings to fill in the gaps. We didn't have an outline for the Quality report and we kept going with a lot of hearings and now I'm trying to pull it together and work backwards and I really wish that we had done it the other way around.
DR. MAYS: I very much appreciate lessons learned. I'm trying to go down the paths that seem like the ones that we will have the outcome that we're really seeking.
DR. LENGERICH: I fully support that, but I do want to respond to Nancy. If we consider the King County and the California, I think we have to consider Alabama and Tennessee and the poorer states that aren't going to be able to mount those kinds of efforts because they just don't have that kind of resources to address their populations. Those two are very unique.
MS. BREEN: That's where the federal government comes in.
MS. HEURTIN-ROBERTS: We keep talking about states as though we should focus on individual states. When you are talking about California, it is true you could and should B but not all states are California. There are little states together that alone could not do anything. Maybe we shouldn=t be focusing on states as political entities, but on regions or areas.
DR. LENGERICH: Or populations?
MS. HEURTIN-ROBERTS: I don=t know.
MS. BREEN: We need to be open. Maybe we want to look at New York City, maybe we want to look at Kings County, maybe we want to look at California, maybe we want to look at the southeast, maybe we want to look at Louisiana, because it is so different from other states. MS. GREENBERG: This is compatible with these parallel studies as well as the vision for the designated population HANES, where you would go out to specific areas, it would not necessarily be a state. It could be an urban area, it could be a rural area, but generally a smaller geographic area. Acknowledging that over sampling is only one technique, and as you were saying earlier, if you want to learn about certain populations, you need to go to those populations as opposed to trying to find them through some massive screening. It depends on what kind of data you are looking for, If you are talking about a lot of data that's descriptive, but not that much that tells you why, then I think that's where community studies can give you some of those beginnings of why.
DR. MAYS: Because I know we have another group that needs to meet, I want to bring this to a close. Here=s what I am going to suggest that we do. Grace, do you know when we get our transcripts from the hearing?
SPEAKER: Usually ten working days.
DR. MAYS: The transcript of the hearing should be coming to us soon. I will try to organize these notes of today, send you the transcripts through e-mail, and then we'll have a conference call.
In the interim, what we'll try and do is make some sense of this. I like Kathy's idea also, let=s try and see if we can make some sense of this, have an outline. Part of what I am going to do is B some of you have skills in very particular areas. You may get assigned that part. I do think that if we're going to have a set of hearings, we're going to be back to needing to talk with lots of people.
If we want to do it quickly, we will have to engage more people to help us come up with individuals. You know what the deal is. People can come in and read from a piece of paper or else people can come in and give us what we really need. That's what we want.
MS. BREEN: We need to do what you did last time. That was an excellent hearing. It was very targeted. The information we got, maybe you wanted more, but it was very good background for what we needed to talk about. It really helped move the thing along. I think we should do that again.
MS. GREENBERG: We probably need to poll for a date before the June meeting. The sooner we can get a writer identified and someone working on this, on the outline, the better.
DR. MAYS: You don't have to convince me of that. They could write this instead of me if they were here. I'm very concerned about getting one. Thank you, everybody. I apologize. We were down here late to those of you came from the outside. I'm glad you were here.
(Whereupon, the subcommittee was adjourned.)