[This Transcript is Unedited]
Renaissance Hotel
999 9th Street, NW
Washington, DC 20001
Participants
DR. FRIEDMAN: Why don't we get started. I'm glad the Rob has joined us. Ed Hunter is on vacation and opted not to come back from the beach to join us. Let me just spend a couple minutes bringing folks up to date on the process since the last meeting. And what would be really helpful is if we could go through the high level schematic outline and then what would be really helpful is if we could go through the more detailed outline.
DR. STARFIELD: Can I just ask, what's in the packet is the same as we got in the mail? So if we made notes on it, we can use it?
DR. FRIEDMAN: Yes, absolutely. Since the last meeting, we've had Ed Hunter and Rob and Gibb Parish and I have been having regular conference calls and we got together once or twice since then, once in Atlanta and we're also going to be getting together in early July here.
And we've been essentially moving ever expanding the detailed outline which is now 20 pages long. Sooner or later we're going to have an outline which is no longer an outline, but a draft of the report. We revised the structure of it since the last meeting and I think that it's clearer. We're starting out with a summary and we're not clear yet whether it's going to be essentially a simple executive summary just recapping chapter by chapter what's in the report or something that focuses a little bit more on the highlights.
Paul has been very helpful in insisting that we really need something that up front makes the main points. We've put report purpose and stakeholders and overview of the report development process (Lisa entered) in the preface because we wanted to make it easy for folks to skip over the less essential parts of the report.
Chapter 1 is meant to be extremely brief and provide an overview of the vision and perhaps a few vignettes to illustrate current problems and value added by achieving the vision. As I said, we're thinking of that as quite brief and we really have done nothing to expand upon that. Any suggestions you have for that would be appreciated.
Chapterr 2, what are health statistics. Chapterr 3 and 4 are based upon the two models that are at the end of the detailed outline. In each of those chapters, we're going to be laying out the model thendevelopingg a matrix based upon each model to describe the current state of health statistics.
Along one axis of the matrix will be the components of the model. The other element of the matrix will describe national, state, local, andproprietaryy data systems. In the cells of the matrix, we would be providing illustrations of how those elements andcomponentss and whether they're being done now in the American HealthStatisticss System.
The third part of each of those chapters will be adescriptionn of gaps and issues and drawing largely upon the testimony we heard in the hearings. Chapter 5 is a reiteration of the principles. Theseelargelyyreflectt the principles that we've been working on for the past year or so.
But at Rob's suggestion, we've reorganized them and consolidated them into a few major headings. I think it will read much better and much more clearly. Thatreorganizationn also reflects the testimony we heard from several of the people, particularly from Mike Rodrey when we were in San Francisco.
Finally, the recommendations and the roadmap.That'ss another part where we have not started specifying it and when we get together in early July, we'll be doing that. What we hope to be able to do and something I'd like to discuss today is come up with a framework for recommendations and then send out that framework to a variety of stakeholders and invite them to provide recommendations.
This could be folks who have given us testimony. It could be NAPSA(?), SANATA(?), WINPASCO(?), et cetera. And that would be a way of doing the report not in a serial fashion and moving ahead with recommendations and perhaps we could send that out this summer. So that's where we are. Any suggestions or comments on the schematic outline before we move into the more detailed?
DR. STARFIELD: This definition of NH is supposed to be under F, right? It's not supposed to be the first part of the document.
DR. FRIEDMAN: Yes, the detailed outline. Also, in the detailed outline the length of what's in there now of the verbiage does not necessarily reflect its important. The NHII, we're not going to have much more in the NHII than we have in the outline and we just wanted to cut and past what we already had in there.
DR. NEWACHECK: I made this comment before that there's clearly a strong emphasis on health status outcomes. That's appropriate given this is a population's subcommittee effort and that's the view around the room. But I also feel that there is a strong thread in health services research that's focused on healthcare itself as an outcome and not health status.
I feel that healthcare doesn't get as much emphasis on here. If you look at NCHS data systems, half of those are healthcare statistics, they're not health outcomes in terms of health status statistics. There's a large number of people in the health services research community and potential users of this report that really view healthcare as the outcome that they're looking at: organization, delivery or financing of healthcare. They're interested in those aspects and not health status or population health.
I feel that potentially for that group, this report might look marginal or not be appropriate for them because there is a strong focus on health outcome and population health. I wonder what your reaction is to that?
DR. FRIEDMAN: It's something that we've discussed. It's something that Ed Hunter is more sensitive to than I am. We've tried to respond to it more in this outline and I think we need to continue to work with it. My own feeling is that the outcome of interest if health and function and well being. What we need to look at is all the different possible influences on that including healthcare. Any specific suggestions you have in places in here where that could be.
DR. NEWACHECK: I wonder if there could even be a chapter on healthcare statistics opposed to population health? There is a large user community out there interested in expenditures or quality or organization delivery as a final outcome not as an intermediate outcome as it relates to population health.
I think this report focuses on population health, but I think you don't want to lose a lot of people who aren't interested in that outcome, but rather in healthcare itself related to statistical systems related to healthcare and not health outcomes.
DR. STARFIELD: We have in the outcome, we have not only the population, but also the distribution of population. Of course, you see healthcare is a major determinant in that. I think that the health services research community hasn't kept up with the times. It's got to move if it's going to have any credibility.
If you look at all the modern models even starting with a decade ago, it doesn't focus on access healthcare as we have put it as one determinant. I think we play it up as a determinant, but not as an outcome. That doesn't mean we can't play it.
DR. NEWACHECK: I don't think it's a matter of keeping up. People are just concerned about different aspects of the healthcare and some are just concerned about healthcare itself rather than population health. We lose people potentially. It probably depends on how the ultimate report comes out. I'm just concerned about that.
DR. STARFIELD: We actually talked last night and I think we should pursue a committee as a fact that the American Public Health Association is just not on-board with any of this health information stuff at all, not at all.
MR. STENBERG: Because of my report back to CDC that was one of the main things I was going to highlight.
DR. STARFIELD: That's a big limitation.
MR. STENBERG: That was my biggest take on this from the meeting last night,.
DR. LUMPKIN: With this issue, it just seems to me that as I was reading into it, unless I missed it, are we purposely staying away from the determinants of health?
DR. FRIEDMAN: I started with determinants and then I backed off from using that partially because on one hand, it's a theoretical model. At the same time, there's been some pretty articulate arguments that have been made saying that we should not have deterministic models of population health and we should think of the input variables since we don't have empirical data. We should think of them as influences rather than determining factors.
DR. STARFIELD: I think the thought was that they are so interactive that it's hard to think of a deterrent. That's why there are no arrows in this diagram.
DR. LUMPKIN: Okay. I was just thinking of that as a way, intro to how you address the health services data sources which are important and the issues of access to quality are appropriately within the domain of public health. We in public health see it that way, but the rest of the world doesn't see it that way. As long as that schism is there, we're not thinking of advances. I'm resonating with some of the comments and see how it fits within the model of why we're dealing with these health data.
The comment that they're behind the times should be a bright light to us that that's an area that we need to push the envelope. That's NCHS, not necessarily this particular document.
DR. FRIEDMAN: It's definitely in the model and both in terms of health policy as well as healthcare, but those are also only two elements of the model. Paul, I suppose you're right. One of the things we are trying to do is emphasize the need in health statistics systems for additional attention on other ecological types of variables some of which we do not pay much attention to, but much less so in the day to day surveillance systems.
MR. WEINZIMER: Also would be a part of the model is statistics development, the second model outlining current data systems and how they contribute to closing gaps.
DR. LUMPKIN: Would it be of value to have this specific discussion of this issue? I'm not exactly sure where, but the discussion of health services research with the scheme of health statistics?
DR. MAYS: Looking at the document, I can see the desire to have healthcare and have health services research a little more prominent, but it takes it somewhat in a different direction where it's not just putting it in, but it seems that you would have to redo the whole document so that if what you really want to do is change people's thinking, then it should be throughout the document. And then where possible, feature or highlight it opposed to just having it in one small area.
I think having a discussion about it is fine, but it doesn't do justice to what you say you want to do for those group of people. Having it where it is is fine so that the document isn't becoming all things to all people. You have to decide to what extent you're going to do it. As it's set up, you really should go back and weave it in the whole document is what you want to achieve is your goal.
DR. LUMPKIN: My thought is that it is to some extent, but if what we want to do is speak to who should be interested in it, there is that one group where something should, maybe in the introduction, something that says what does this mean? Those in public health knows exactly what it says.
Sometimes it's better to say what it does mean to people in public health and to try and say that, I think it's not all that written out. It's important to people outside public health to know how important this is. The assumption is that somewhere along the line you have to explain what in public health is important to them.
DR. FRIEDMAN: John, how about one possible place being the preface where we list the stakeholders? We could have a discussion of the implication of the document for each of the stakeholders.
DR. NEWACHECK: Is there going to be discussion of specific surveys and data systems or is it going to stay at the high level?
DR. FRIEDMAN: This is a real issue. Gibb, who developed a matrix like this as a chapter for a book edited by Toish and Churchill(?) on public health surveillance systems in the second edition. He developed something similar to this.
We've been going back and forth on how detailed. On one hand, we clearly can't have something that's really comprehensive and detailed, but I think we will at least be mentioning individual surveys and data collection systems, but it has to stay high level. This is 20 pages long.
DR. NEWACHECK: I was just thinking Ed Hunter did a table that had all NCHS data systems as well as some of the other ones like MEPS identified and the columns were population based and different characteristics. It might make a nice appendix to this report. I'm just thinking about whether the healthcare data systems are going to be highlighted in here.
DR. FRIEDMAN: The intent is that in Chapter 3(b) where we map the current system to the first model, we would describe healthcare data collection systems at a high level, but national, state, local, and proprietary. At the same tim, we're not talking about 10 pages on that. We're talking about some entries in the matrix.
DR. NEWACHECK: How long will the report be total?
DR. FRIEDMAN: I would guess 40-50 pages. Our intent will be to do a couple of drafts and then professional editing. We need to do the first few drafts ourselves before we hand it over to somebody.
MR. STENBERG: I agree. If you're going to do the work in compiling the data systems and then cutting it down for Chapters 3-4, I think you should put an appendix where we can list. This report is going to go up on the Web and this will allow a way for people to have access to those data systems.
DR. FRIEDMAN: Complete list being NCHS, ARHQ, CDC. I think that's fine. We'll do that. One other thing we need to do is we need to concentrate on some of the main stuff instead of getting diverted. One of the things that Gibb and I have been obsessing on is the glossary. I said to myself let's forget about that for now.
DR. NEWACHECK: That's one way of giving a little more emphasis to the healthcare systems.
DR. FRIEDMAN: This is something that Ed is in more agreement than I am.
DR. STARFIELD: It was Ed actually that filled out that little box.
DR. FRIEDMAN: Should we go through this document section by section? We've talked about expanding the stakeholder section in the preface to include a statement about what this document means for each of them.
This is a little bit buried in there, but you'll see in Chapter 1(a), we have too general vignettes. The second was put in by Ed in response to your and his concern. Any suggestions here, we would appreciate about how to structure this as possible vignettes and so forth.
DR. NEWACHECK: My comment here would be to engage the public health services research community. It's just on healthcare as an outcome. It wold be to have that vignette. Not healthcare as a determiner of health, but healthcare as an outcome in itself.
So you would have one vignette about determinants of health and another about healthcare statistics and how they could be used as opposed as an determinant. It doesn't necessarily fit the model, but engages a different audience. What about heart disease, bypass surgery, ratio difference?
DR. STARFIELD: That's why we have the distribution of health as average health which is generally what we publish, but also distribution of health which gets at the disparities of equal importance.
DR. LUMPKIN: What I was thinking of as a vignette pointing out health statistics and defining that particular connection with a good health statistics system and it's hard to find what the disparity is and how to address it and find solutions.
DR. FRIEDMAN: I think some of that is based upon discharge data, VA data and I think that would really be a good one.
DR. STARFIELD: The intent is to have a definition of health statistics in the preface; is it not? It's not here.
DR. FRIEDMAN: We now have it at the beginning of Chapter 2.
DR. STARFIELD: It's got to be in the preface.
DR. LUMPKIN: Folic acid and --
DR. STARFIELD: I don't mean a big long discussion, just one sentence.
DR. FRIEDMAN: Any other suggestions on these would be welcome and feel free to send them to me.
DR. LUMPKIN: You may want to look at the MMWR series of articles on the 10 Greatest Achievements. You may be able to pull one out of there.
MR. STENBERG: Picking up on the themes, healthcare delivery like shifting and where people go for healthcare, changing roles of emergency departments. I don't know if it fits or not.
DR. STARFIELD: It would be 150-page report.
MR. STENBERG: I'm just taking up on the theme of stretching to healthcare itself.
DR. STARFIELD: I know you don't mean it this way, Paul, but none of the model is going to be in the -- to talk about healthcare as an outcome. I know what you mean when you say outcome.
DR. FRIEDMAN: These vignettes are 2-3 sentences and we could put something in on changes in inpatient, move from inpatient to outpatient treatment and how we know what's been happening on the inpatient side because we've got inpatient discharge data. But we don't know what's happening on the outpatient side became we don't have any population based outpatient. That can be 2-3 sentences.
MS. GREENBERG: That's what happened with eye surgery when NCHS didn't have an ambulatory.
DR. STARFIELD: USA TODAY. That shows the difference between errors and adverse effects. Those aren't errors. There is a big front-page article on 5 percent bad outcome from laser eye surgery.
MR. STENBERG: The other one about JCHL picking up on making errors a part of accreditation.
DR. STARFIELD: Did you see they defined errors as unintended?
DR. NEWACHECK: One model that might be interesting to try to capture for healthcare and health outcomes would be something about S-CHIP. Right now, there are millions of dollars going into public health services research, looking at outreach, retention. There are intermediate outcomes, basically getting kids enrolled in the program and keeping them enrolled.
Ultimately, those same researchers will look down the line at health outcomes, but right now, they're looking at intermediate outcomes in the healthcare system. Is it working properly? Is it getting kids enrolled? Are kids getting services?
Maybe there could be a two-stage thing that would be a vignette that shows the value of health statistics in helping us to do formative evaluations of a program like S-CHIP and then ultimately to look at health outcomes from that program.
Right now the focus is on just the intermediate outcomes of how many kids are enrolled, are they being retained in the program and that sort of thing and there's a large amount of investment going into surveys like this, just looking at that issue alone.
You could do it as a two-stage vignette that illustrates the value of health statistics for healthcare outcomes, these intermediate outcomes, and then ultimately for population health outcomes for that population.
DR. STARFIELD: The vignette will point up what we're getting and we still need?
DR. NEWACHECK: But it would also illustrate how the Congress asked for it. They didn't ask for health population outcomes. It asked for evaluation data on re-enrollment and retention basically. That's what it asked for. A lot of money has gone into that, $20-30 million worth of federal dollars just to get bound for their five-year evaluation. HCFA is involved, AHRQ and a lot of agencies are involved in collecting that data and sponsoring surveys.
MS. GREENBERG: The nice thing there is without robust health statistics system that picks up whether the health status is improving, you don't know whether you're achieving getting everybody covered, but are you really achieving the public health goals, popveing health status. You need to follow that over time which is what we do very little of.
DR. STARFIELD: That gets at the disparities issues by narrowing the gap.
MS. GREENBERG: Is the intent going to be to have different vignettes than the NHII report, complimentary vignette?
DR. FRIEDMAN: We haven't thought about relating the vignette directly to the vignettes in the NHII report.
MS. GREENBERG: I know there will be community health vignettes in that.
DR. FRIEDMAN: We could think of doing that, but we haven't gotten that far with this section.
MS. GREENBERG: And then we commission some papers which could contribute to these. There was one on children health, on health reform, are any others?
DR. STARFIELD: They haven't gotten lost yet?
DR. FRIEDMAN: Any additional thoughts on this would be very much appreciated. This has been very helpful. Chapter 2, we keep redefining health statistics. This is very much a definition in process was an attempt to broaden the focus of the definition just from population health outcomes to make more explicit that it also includes influences.
MS. GREENBERG: It's a link to the model.
DR. STARFIELD: But it's missing the distribution aspect.
DR. FRIEDMAN: Good point.
DR. LUMPKIN: Question. Are health statistics the infrastructure --
DR. STARFIELD: Give us an example of what else is infrastructure so we can decide whether it's similar to that or not.
DR. LUMPKIN: Workforce issues.
DR. FRIEDMAN: I would think that would go under the healthcare box and influences, but that's not a particularly effort.
MS. GREENBERG: Community resources?
DR. FRIEDMAN: Yes, maybe community resources.
DR. STARFIELD: It's everything in the model. INFORMATION is just everything in there, but it's an infrastructure of our national data system not just health data.
DR. LUMPKIN: When one describes healthcare, you have to talk about issues related to workforce issues that become an input variable, a statistic you have to measure. We don't do that on the non-healthcareside. How many sanitariums are there? I'm not saying it has to be in here. I'm just wondering if it's a piece that might be missing in the concept of what we describe in health statistics.
DR. FRIEDMAN: I think it's a good point and one we should make. I'm not sure where we would put it in, but I think it needs to be in there.
DR. NEWACHECK: One thing on the definition in the last sentence. We have a list of people, health services researchers so that they'll feel like they're part of this.
DR. LUMPKIN: In the last sentence it says "properly organized and communicate health statistics to enable to access local or national health." Okay. Mobilize implies allocate resources?
DR. FRIEDMAN: It would to me, but perhaps we should make that explicit.
DR. NEWACHECK: Mobilize has a nice ring to it because it implies action. Allocation doesn't have that same meaning.
MS. GREENBERG: It sounds like Susan.
DR. FRIEDMAN: That's what I was going to say. That particular sentence has the ring of something that Susan wrote.
DR. STARFIELD: A couple of things under point to health statistics. The first sentence as individuals and population subgroups and again to just keep. That's the average level of this issue. This third sentence, "health statistics enable us to view the health of the American population and subpopulations". I don't think we need as wholes.
DR. FRIEDMAN: If there aren't any other comments on this, Chapter 2(b) uses of health statistics. One of the things we need to add there are examples. It's not intended to be lengthy or comprehensive, but I think having some real life examples throughout that section would be helpful.
DR. STARFIELD: Prototypes or uses of health statistics is the Kaiser Commission on Medicaid which everyday we get tons of material. Very expensive I'm sure. Is that a use of health statistics? Most of them say what we've known for 20 years.
DR. NEWACHECK: I'm hoping people will forget the articles I wrote 20 years ago.
DR. STARFIELD: Is that a use, what's a use?
DR. FRIEDMAN: We've tried to categorize different types of uses here. The first is research, the second policy development and assessment, et cetera. The third is more programmatic purposes.
DR. STARFIELD: What would the criteria for something being useful? I'm having trouble trying to figure out what would be an example. It's like we're always looking for something other than taking the lead out of gas. That's the one, interestingly enough, but the health statistics in another country.
DR. FRIEDMAN: We would talk about infant mortality, subgroup differences. In public health practice those are the information that's used in developing programs.
DR. STARFIELD: It doesn't have to be a utility in improving health.
DR. FRIEDMAN: Those are the data that are used, allocating resources to decide where to put new resources at very specific geographic levels.
MR. WEINZIMER: There are local meetings in Albuquerque and New Orleans where people discuss meeting the data and making decision.
DR. STARFIELD: The distinction between that and the lead poisoning. It actually has an impact on improving health.
MS. GREENBERG: Those are the hard ones to come up with. These intermediate ones are easy.
DR. LUMPKIN: Here's an example. We had to address the issue of infant mortality and our numbers had a slightly upward lift. In Indiana, it's reaching a plateau. But when we analyzed the data for women who participate in our family case management, women who don't participate its 35.
There is an example where resource allocation would say you don't need to allocate more resources for case management. You have to allocate them for the jobs of women. So outcome should become the next priority for the program.
DR. STARFIELD: Did you publish it? We should use that kind of thing.
MS. GREENBERG: You must have had large populations. Is that the whole state?
DR. LUMPKIN: Over 70 years or whatever. So that's a nice little things.
DR. NEWACHECK: It would be nice to have examples to that it could come out in headlines of USA TODAY so that there would really be things people could grab onto whether they're experts in public health or not. I'm trying to think of some things that have been in USA TODAY or whatever.
DR. FRIEDMAN: That's a good idea. I wouldn't want to use colorectal screening as an example, stage of diagnosis of colorectal cancer as an example of something that we use specific geographic areas. There's a clear connection between stage of diagnosis and mortality.
DR. STARFIELD: That's an example of a disparity issue.
DR. NEWACHECK: An example of public health services research too.
DR. LUMPKIN: The last point there looks like it ought to be -- It doesn't seem to be consistent.
DR. FRIEDMAN: Enabling information or within that?
DR. LUMPKIN: Yes.
DR. FRIEDMAN: At one point it was. I agree.
DR. LUMPKIN: That would be a natural reference to NHII?
DR. FRIEDMAN: Yes, the reason why we removed it as a separate number and put it back was that we did not want to put too much emphasis on health statistics as used for individual health decisions. Not as if many people will pick up NCI survival data and say, I'd better go get my colonoscopy.
MS. GREENBERG: It's truth. I have a close friend who's in the middle of chemotherapy for breast cancer, who is going for two more treatments because the survival data shows that those two extra treatments give you a 2 percent better survival.
MR. STENBERG: That was exactly the point we were making at the NHII meeting that people want these statistics to make decisions. I don't know if I told you, but it's a key point to capture.
DR. FRIEDMAN: Chapter 3(c) all those individual bullet points should be prefaced with "need for". If that's a little hard to understand, I apologize for that.
MS. GREENBERG: This is just a technical point on Chapter 3, components, number 4. I think we could clarify the aggregate and ecological, but I have some wording to suggest.
DR. FRIEDMAN: At some point, we had a definition section in here and we decided that it would be diverting and we would put in only a few key definitions and move the rest to a glossary.
DR. NEWACHECK: Have you run the model past a focus group so that people get this?
DR. FRIEDMAN: That's a good point. No, we have not.
DR. NEWACHECK: When I look at this abstractly, it seems like everything on the universe is on this page, community, healthcare, but there is no connection between them. I understand the reason for not having, but I could see double arrows between the major components and the outcomes of interest population. I just wonder what people who haven't been involved in this model development would think of this page without seeing any causality at all.
DR. FRIEDMAN: Even people who have been involved in it might argue for putting them in. I like double arrows myself that shows there's a relationship, but it isn't a one-way relationship.
DR. STARFIELD: I have a model that I've been using and I publish it and talk a lot about it. I keep being pushed to either remove the arrows or make them double. The problem with making them double here is it enormously messes up the guide.
DR. NEWACHECK: If you just had some fat big arrows that would point just between each of these things. This could stay out as context doesn't need any arrows, but among these four major boxes and the outcome interests. You could have a double arrow that would be shaded and would apply there is some kind of duel relationship.
DR. FRIEDMAN: If you feel inspired to draw something, rip this out and put it in and fax it to me. that would be really helpful.
MS. GREENBERG: The assumption with population being the only oval box and being slightly shaded and in the middle is that these things are all interacting.
DR. NEWACHECK: It feels like a solar system to me. That's how I look at it with the Milky Way out there. Here is the sun in the middle and Mars, but that's visually the message I get. I don't know that birth is related to Mars or whatever. I get the point sort of, but it's a subtle point and people are used to seeing arrows and direction. It doesn't' have to have it, but I'll draw a picture.
DR. FRIEDMAN: That will be great.
DR. MAYS: If you put arrows in, you're going to get into a lot of theoretical discussions with people. Arrows would actually end up with people criticizing ways in which it's not necessary. You might say none of this needs arrows. I can tell you a lot of recent work.
If you're going to call cultural contacts and you're going to have norms and values, races and sexes which there is work coming out now saying generic biological characteristics. You're going to get into something that it's going to take away. There will be a lot of discussion about it, but let everybody see themselves in it. If you put arrows, they see themselves in a way in which it doesn't fit how they view it. Then they're going to argue about it in ways in which it loses the point.
DR. FRIEDMAN: That was why we left them out. That was the reason. Another possibility since we're describing this as a template, we can do one without arrows and one with and just make clear.
MS. GREENBERG: I was thinking of some variant on that. As part of the description, you can say that people can use this if they want to focus on causality or un-relationships they can use pieces of them. They can use arrows in various places, but that's why as a template, it doesn't have arrows.
But it could be adapted for particular uses if you want to show the relationship between cultural context or environment or the relationship between aggregate community stuff. Actually, as a template, I think it shouldn't have arrows.
DR. LUMPKIN: Wouldn't the arrows depend upon the health issue?
MS. GREENBERG: That's what I'm thinking. The arrows is an application as opposed to the template itself.
DR. LUMPKIN: Taking up from Dan's suggestion that he didn't like, we could present the template and an example using a health issue of how it looks with the arrows.
MS. GREENBERG: You could do that which might actually make it more user friendly, but the template without and then an example. Yes, I like that.
MS. JACKSON: An overlay where you could see the template on its own and then see how things can be influenced by a certain circumstance?
MS. GREENBERG: But just a particular example as opposed to --
MS. JACKSON: I would take like one example.
DR. LUMPKIN: In the '96 Highland Report on Performance and Community Health, there were a number of examples in the appendix where the field model was used to develop, but by using those matrices and other ones, you very easily plotted out reference so that you don't have to justify the arrows.
DR. STARFIELD: This is an IOM report '96?
DR. LUMPKIN: Yes.
DR. FRIEDMAN: I know the one you mean.
DR. MAYS: Can I ask about some, and I'm just catching up on this, on terms? Under 4, you talk about political context. You have a second bullet, political culture. Is it political culture or political climate? I'm not clear what political culture is.
If it's political, the culture changes so much, but political climate, you really locate based on a certain set of indicators. We talk about the political climate. We look at laws and we look at things like that, but the culture has more to do with the process of how we function.
DR. LUMPKIN: There are different ways that things get done and if we're using health statistics to figure out how things get done.
DR. FRIEDMAN: Both may be relevant and one of the criticisms that have been made since some of the other population health models has been that they are too generalized rather than being related to specific times and specific places. The definitional issue, it's a real issue which is why we were getting converted with the glossary. We were using words loosely without really nailing down what we meant.
MS. GREENBERG: Political culture is also in Figure 1 in the template.
MS. GREENBERG: Since you said the Gibb word, is he continuing to be involved from New England or did he move to New England to be more involved?
DR. FRIEDMAN: Absolutely. I don't know about that, but a good part of his time is allocated to this.
MS. GREENBERG: I was surprised when he went off to Vermont.
DR. FRIEDMAN: He's got additional responsibilities.
DR. MAYS: Let me ask about the cultural context one. You have norms and values, races, sexes and competition and cooperation. In terms of the isms, is there a particular reason to have just racism and sexism.
DR. STARFIELD: Rather than?
DR. MAYS: I'd probably put in for sure things like ageism. I think that's very critical of the population to age more if you want to look at the whole child thing. I want to go back, when you say norms and values, you're talking about cultural norms and values. I'm not clear how you're using them.
DR. FRIEDMAN: Neither am I. I mean that seriously. This is part of the discussion that we've been having in terms of the glossary. We have been non-specific in how we've used some of these words. At some point, we need to go back. Even the fact of listing words in the glossary has made just how clear how sloppy we've been and we need to nail it down.
DR. MAYS: There's an issue there and I keep looking at it in terms of the figure. The issue in terms of the cultural context is first of all, majority/minority if we're talking about population and how do we then reflect that and to say if you're really talking about collection of statistics, how does that play out?
Some people criticize health statistics for that reason. I was looking at one that says something about the ethics. It should be scientific objectivity and I said, oh, really. The problem is whose science is it. Here, in terms of cultural context, again, if we want all the audiences in I'm trying to figure out how to say this a little differently so there isn't one value that's promoted, but instead it's a reflection that there are competing values, competing. That's what I mean in terms of value here.
DR. LUMPKIN: I'm not reading that section that way. What are the values that either enable or create barriers to health? What are the norms? Is this a community like the one I live in where people are -- That is the cultural context of our community.
Another cultural context may be related to the majority and then other scattered minorities. To what extent now, if you have a major norm, do those norms have some impact upon healthcare? That's really how I read that section.
DR. MAYS: I don't disagree with what you're saying but when I look at it, it doesn't' capture the diversity of what you're saying. It seems pretty standard public health that people are not agreeing with. It's more complex than what's here. You're saying is I agree with, but either we need examples or some way in which what you're advocating --
I think it's more complex than what's here. To say that what you're advocating is how do we live in a society in which we're talking about population's health and we're worried about some of the vulnerable population which you indicate here, that norms and values are important in that context.
MS. GREENBERG: Diversity of norms of values?
DR. MAYS: Cultural contexts. I think that is really the essence of going back to different cultures and not just a context.
MS. GREENBERG: You want the diversity in this.
DR. MAYS: I want the diversity there. I want to push people to think about that.
MR. WEINZIMER: Having diversity of norms and values would be an important thing to communicate.
DR. STARFIELD: That indicates in an influence diagram that within a community there would be diversity, but basically what we're trying to think about is the cultural context or influence and you want to look at that and consider that. It's not the diversity you want to look at, it's what it is in any population group you're looking at.
DR. MAYS: And that you're clear that there are many.
DR. NEWACHECK: This implies the way it's written here is that there is one norm and one set of values and that's not true.
DR. STARFIELD: I wonder if we can just take that out, that norms and values because we don't even know what it means or how we'd measure it.
DR. LUMPKIN: There was some discussion about norms and values as related to data collection issues or willingness to participate in surveys and we want to try to get that in.
DR. MAYS: That's very different than being in this section.
DR. STARFIELD: That's below the social cohesion which is also on this.
DR. LUMPKIN: I just had a non sequitur thought. When we began discussing these issues and talking about a very quantitative system which has at some point all data measures. I just wanted to put that as a place holder as it bears some discussion. In other words, by community, you can't measure whether somebody comes from a certain culture. You have to make some sort of qualitative assessment and then we quantify those and we aggregate those.
DR. MAYS: You can't measure that quantitatively?
DR. STARFIELD: Can we measure?
DR. MAYS: I think there's a growing literature.
DR. LUMPKIN: Frequently that's called qualitative measure.
DR. MAYS: If we take the National Survey of Black Americans, there's a whole section about your ethnic identify, how close or distant you feel and they give you characteristics that have to do with the population itself, what kind of food you eat, what kind of language you use.
DR. STARFIELD: Acculturation type questions.
DR. MAYS: Is that what you're talking about?
DR. STARFIELD: The whole acculturation component of that was critical.
DR. NEWACHECK: Let me suggest that I need to go back and do a little research because at our school, there is a whole group of people who are modelling what they call qualitative measures at the School of Public Health. I'm not sure I fully can explain what it is they perceive as different.
MS. GREENBERG: This may be semantic more than --
DR. NEWACHECK: But to some extent the application or anthropological measures of public health, I think we need to address that in some way. I'm not saying here, I'm just saying this is a place holder idea that ought to be somewhere.
DR. FRIEDMAN: We have five minutes. I'm beginning to think looking at Marjorie here that what would be really helpful is would be before we come up with a draft, before we start turning this into more verbiage in a draft which we were hoping to have for the September meeting, I think it would be really helpful, we have to have the September meeting, if we could get together and continue to go through this or a slightly later version of this.
DR. STARFIELD: Are you thinking in August?
MS. GREENBERG: You mean July probably.
DR. FRIEDMAN: Even if it's this document. The discussion is extremely helpful and having people sit down around the room who are not working with it every week and it's going to be a committee as well as an NCHA and Data Council report. I'd rather get the input earlier than later. If there is any way we could take half a day or a day.
MS. GREENBERG: What I was just thinking about and it may be too late is that with the exception of Paul, John, you, Vickie and Barbara are all going to be at the executive subcommittee meeting. Coming in earlier that day and maybe Paul could come also. It's in Chicago. So it would be an easier trip for you or we could put you on the phone.
DR. FRIEDMAN: I don't know if I could come or not, but having Ed as well as Rob and Gibb there would be really helpful.
MS. JACKSON: By that time, you would have to have another iteration of this document if you're talking about something in September.
MS. GREENBERG: You could be working on the part we've gone through.
DR. FRIEDMAN: We're going to continue iterating.
MS. GREENBERG: It wouldn't be like you wouldn't do anything between now and then. I got here late because I was in Privacy so I wasn't able to make a global statement, but I think this is going to be incredible when it gets done.
DR. FRIEDMAN: When it gets done is the operative and very frightening phrase.
MS. GREENBERG: This is really rich I think. I'm very excited by it. I didn't get a chance to make that statement but having a completely separate meeting might be difficult.
DR. FRIEDMAN: That would be fine with me. I am free that day and the one caveat --
MS. GREENBERG: You're coming in, right, for the executive session.
DR. FRIEDMAN: I'm coming for the executive session.
MS. GREENBERG: Aren't we coming in on Monday night and then a Tuesday. So if we came in Monday morning. Monday morning is hard. Why don't we do it the day after, you're thinking?
DR. FRIEDMAN: Maybe I could do it the next day. I need to check.
MS. GREENBERG: The problem with the 15th is that's Data Council; isn't it or maybe not? We're having an evening dinner. Actually, Data Council is the 8th, not the 15th because the 1st is a Wednesday. Too much a good thing is wonderful Mae West, I like it. So the question is whether, Monday afternoon would have been ideal it seems. I don't know when you're coming in.
DR. MAYS: I have no idea how early I can leave to get in. It's Chicago, on the way.
MS. GREENBERG: So maybe it should be Wednesday, half day on Wednesday.
DR. MAYS: It doesn't matter to me. I'm staying with my family.
DR. FRIEDMAN: Why don't we check with Ed and Gibb as well. It would be really helpful if at least some subset of them would be here.
DR. STARFIELD: Dan, I have a few comments which I can convey to you, but there's one policy issue that somebody needs to think about whether we're going to tie the recommendations, the roadmap to NHII in any way or is it going to be linked in any way.
DR. FRIEDMAN: I'm not sure how they should be tied, but they should at least be referenced or tied somehow. I agree.
DR. NEWACHECK: I don't know if we're talking about Chapter 3 in the model, do we really have to have this ecological aggregate everywhere, both in the model and in the text.
DR. STARFIELD: It has huge implications for data. Aggregate means you can use individual level for data and just put it together, but ecological, you've got to have different kinds of data.
DR. NEWACHECK: But in the context of this discussion, it's a conceptual model. It's not really an implied model when we're describing Chapter 3 with broad concepts. In the figure, it's broad concepts too, but I find it distracting to see aggregate and ecological in parentheses everywhere.
DR. FRIEDMAN: There may be visual ways of differentiating without having to have the words. Since this will be a 4-5 color report, we can have some in red and some in blue. The iteration we're discussing in August may be sloppier than this. Thank you.
(The session was concluded.)