Hubert H. Humphrey Building
200 Independence Avenue,
S.W.
Room 325-A
Washington, D.C.
DR. FRIEDMAN: Before we get started, why don't we do introductions. Everybody doesn't know everybody. Why don't we just real quickly go around the room and introduce ourselves.
I am Dan Friedman, and I am a member of the NCVHS. I am assistant commissioner of the Massachusetts Department of Public Health.
DR. STARFIELD: I am Barbara Starfield. I am from Johns Hopkins Medical Institutions. I am a member of the committee.
MS. WARD: Elizabeth Ward, member of the committee, from the Washington State Department of Health.
MR. SCANLON: Jim Scanlon from HHS, staff to the committee.
DR. SONDIK: Ed Sondik from the National Center for Health Statistics.
DR. AMARO: Hortensia Amaro, professor at Boston University School of Public Health. I am a member of the committee.
MS. DURCH: Jane Durch, staff to the Committee on National Statistics at the National Academy of Sciences
MR. WHITE: I am Andy White. I am deputy director of the Committee on National Statistics.
MR. STRAFF: Miron Straff, director of the committee.
MS. LI: I am Rose Li. I am with the National Institute on Child Health and Human Development at NIH, and staff to this committee as time permits.
DR. HENDERSHOTT: Gerry Hendershott, National Center for Health Statistics, staff to the committee.
MR. HUNTER: Ed Hunter, National Center for Health Statistics.
DR. NEWACHECK: Paul Newacheck with the University of California, San Francisco, and a member of the committee.
DR. TAKEUCHI: David Takeuchi, Department of Sociology, Indiana University, and a member of the committee.
MR. HANDLER: Aaron Handler, Indian Health Services. I am a staff person to the subcommittee on populations, just killing time.
MR. GOLDWATER: I am Jason Goldwater. I am with NCHS.
MR. BARTLETT: I am Larry Bartlett with Health Systems Research. I am trying to figure out if I am in the right room.
DR. FRIEDMAN: You are. I am relieved that only one person amended their introduction to say, as time permits.
This is the first time that we have all been in the same room together, both individually as well as organizationally.
It is really a terrific opportunity to review what we are trying to do at the most general level, as well as specifically what we need to do in order to get there.
There were four things that we put on the agenda today, which was reviewing the status of the commissioned papers as part of this effort, reviewing the status of the focus groups, the CNSTAT workshop, and then talking about next steps.
At the most general level, why don't we start with Ed spending a minute on what he came to us and asked us to do around a year ago.
DR. SONDIK: I felt that health statistics in general, as far as the center was concerned, and health statistics more broadly, as far as the data council was concerned, seemed to be evolving without a clear picture of target, if you will. Where was it going.
I felt there were so many forces that were impinging on all of us, that we really needed to account for in our planning, and I didn't feel that we had a consensus as to how to deal with those forces.
For example, there is just the changing health care system. There is technology. There is biology coming down the line that, I am convinced, will play a major role in health statistics. There is HIPAA and the implications of HIPAA.
It seemed to me that with all of this change, there was enormous opportunity for us to define a target or a set of targets, a set of scenarios, if you will, that we could be aiming for.
So, I came and said I thought it would be very helpful for us at NCHS, but more broadly, for the field if we pulled together a group to think about these alternative futures and how we might define these futures best as far as health statistics were concerned, and how we might then, if you will, try to control our systems toward aiming for targets that we thought were best.
It seemed to me that there were three major groups that needed to be involved in this. One was this committee; the second was CNSTAT; the third was the Data Council.
I have visited all three to see if we could enlist the support of all three, and this is sort of the first time that we have really been in the room together to see just where we stand on it. That really was the goal for me.
As I envisioned it, I thought it was important to define a target for health statistics in general. Then secondly, and more specifically, but secondly, for NCVHS, a target that we could work toward within the health statistics system of HHS and the state and local health systems.
I thought it was important first to get a picture of where we saw the field going, and the institutions in that field, whether the institutions were maintaining themselves appropriately, whether they were strong enough, whether they were hitting the points they should be, how they interrelate, and that sort of thing. In a nutshell, that is how I saw it.
DR. FRIEDMAN: Essentially, what happened over the ensuing three or four months, we had a series of discussions, both internal to NCVHS as well as with NCHS, and Carolyn Rimes and some others at HCFA, regarding how we could get there.
As a result of a great many conversations and several iterations, we basically came up with a strategy consisting of three steps.
One step was a series of, for want of a better term, focus group panels that would address, although not necessarily achieve consensus on, some very basic issues about health statistics as a field.
What is health statistics, whether it is boundaries. How is health statistics different or not different from, for example, public health surveillance or health services research, epidemiology. What are its functions.
How do we evaluate health statistics and health statistics systems, both in the past and as we move forward.
One of the things we should do today is discuss those focus groups as well as we have got a list of potential members for the focus groups.
Larry Bartlett, who NCHS has contracted with to be a facilitator for the focus groups, was good enough to join us today.
That is essentially one leg of the three legged stool. The second leg was a series of commissioned papers that were intended also to address some of the very most basic issues around health statistics.
The genesis for both the focus groups and the commissioned papers was our feeling, jointly, that health statistics, in and of itself, as a field, is not something that has received a lot of definition, or as a matter of fact, any definition.
Unlike -- some of you have heard me say this several times -- unlike public health surveillance, where there is a literature defining public health surveillance and its purposes and its boundaries, unlike health services research.
There really isn't anything like that, that we have been able to find for health statistics. In fact, it seems to be defined primarily legislatively, or common sensically, but not in terms of saying what really is it.
So, the genesis for both the focus groups and the commissioned papers were both partially just trying to put some boundaries around what are we talking about when we are talking about health statistics, and is it the same or different than other things that we are talking about, like surveillance.
The commissioned papers that we came up with were roughly half a dozen, in terms of things that we felt would move the discussion forward, and move it forward to being able to help define where we should be as a field in 10 to 15 years.
These are in no particular order. One paper was on models of health statistics systems in other countries.
There certainly seems to have been some more reflexive general work in some other countries, such as Canada, about what health statistics is, than there has been in the United States.
We commissioned a paper, actually, from the Canadian Institute of Health Information, NCHS, to examine that.
Another paper somewhat related to that is on models of the health care system and models of health statistics systems.
Basically, the genesis of this paper was the notion was that it is hard to identify gaps in health statistics if we don't have a general systems model out there, particularly as we move forward.
Noralou Roos and her colleagues from the University of Manitoba are going to be doing that paper. They have already done some related work on this.
The published a very good article in Millbank a couple of years ago that was related to this.
A third paper was to focus on what we could learn about met and unmet health statistics needs from the health care reform debate.
Rick Kronick from the University of California at San Diego is going to be writing that paper.
A fourth paper is going to focus similarly on what are the met and unmet health statistics needs from the Children's Health Insurance Program.
Those topics are really starting points. In our discussions with the authors, and NCHS' discussions with the authors, I think we haven't intended to really limit it. The idea was to give them a starting point.
There is another paper that we have been trying to commission, and we haven't had any luck so far, unfortunately, which is what are the characteristics of robust health statistics systems, health statistics systems that have proven useful over time and have been around for a while.
Communicable disease reporting, for example, has been around for 100 years at the state level. Vitals at the state level has been around for 150 years. HIS, I don't know, has been around for about 40 years.
What can we learn about what has worked and how can we apply what has worked toward looking toward the future.
Then another paper that we identified a need for was essentially a meta review of reviews and recommendations relating to health statistics systems.
So, that is basically commissioned papers. Then finally, the workshop which had been the initial genesis of this effort.
So, that is essentially --
DR. STARFIELD: So, you have people for four of the six papers; is that right?
DR. FRIEDMAN: Really, four of the five. One of them, my understanding from Ed, is CNSTAT is going to be doing. The fifth, which is the robust, we still don't have anybody for.
MR. HUNTER: I think Ed talked to CNSTAT about it. I don't know that there has been a real concrete discussion of it.
This one would be sort of compiling previous advisory input and recommendations that are out there, that might serve as a starting place, but aren't necessarily implemented.
So, we have had a brief conversation about that. We haven't exactly followed through on that all the way.
DR. SONDIK: One of the first discussions I had on this was actually held at the Academy with Miron and others and Don Detmer.
I felt it was important to, at least as I was thinking about this workshop at the time, to have a set of recommendations or conclusions that carry as much weight as possible.
At the same time, I wasn't sure that the problem as I saw it was one in which it needed to be studied for a long period of time.
In that discussion -- I don't want to misstate this, Miron, so tell me if I am wrong or my memory is wrong on this -- it was really in that discussion that we thought of this workshop format.
Then it would be possible to at least record what the group said as a kind of consensus or, in some cases, maybe non-consensus, but to be able to do that in a fairly expeditious manner, as opposed to doing something which would be a longer Academy study of the topic.
In holding a workshop under the CNSTAT's auspices, the Academy's auspices, would really put the agenda of the workshop in their purview.
So, we discussed that and then there was the question about how we could get the various actors here to all be able to share ideas.
One of the suggestions at one of the subsequent meetings was that we in fact have this meeting in which we all got together here and sort of talked about our ideas for this.
I think it would be useful to hear the Academy's perception of this.
MR. STRAFF: There was general agreement for the workshop, but not necessarily in lieu of a panel study. I think we felt, let's get the issues out on the table and they might give us some ideas for how to proceed over a longer period of time.
MR. SCANLON: I think, Miron, one of the examples of such a workshop was one that CNSTAT held for HHS back in December of 1997, I think, where you brought together the welfare reform community, the health statistics community, the health reform community, where you got producers of data, and users of data.
Over the course of two days, you were able to identify a number of the data needs, even though they were a subject of formal recommendations to the Academy later, I guess. You were able to put it together sort of as a proceedings, so people could see what the views were.
DR. NEWACHECK: Dan, from the perspective of our committee, can you talk a little bit about what you see as being the ultimate product of this workshop? Would it be beyond that to something else? It sounds like it could be a committee study, too, but from our side, the NCVHS side.
DR. FRIEDMAN: Sure. This has been so collaborative that, for better or for worse, Paul, I haven't really thought of it as NCVHS distinctly.
I do see the commissioned papers and the focus groups as hopefully generating a small body of literature on trying to define health statistics and lay out the issues.
I would certainly hope that the workshop would continue to do that, but also put it more in terms of where is it going to be, where does it need to be, in 10 to 15 years.
Personally, I would be very satisfied if we could come out with three or four or five or six papers that were similar to work that has been done in the public health surveillance world.
Here is what it is and here is what it should consist of. Here is what its purpose is.
Alternatively, I think it would be a real service if there was a paper that said, hey, maybe it is really not different from X,Y and Z, but just something to sort of pin it down, I think, would be very helpful.
DR. NEWACHECK: I guess what I am getting about, what I am thinking about, and maybe this is not the right time to think about this right now, but sometimes when a project is in a workshop, that is great, but then there is no continuity, no next steps. Then it kind of gathers dust.
I think perhaps we should be thinking perhaps longer term about what we would really like to do to implement some of the ideas that come out of the workshop.
Maybe we need to first fill in some of the details about what we are actually doing before we get to that point.
DR. SONDIK: This informatics infrastructure that the committee has been addressing, it seems to me this fits in with that.
I take it that that will be a continuing theme of the committee. It clearly has got to be a theme of the department, as to how it is addressing all the issues associated with informatics, everything from the practical side to the research side. I see this fitting with that.
I think this idea of the focus groups, and particularly the idea of the definition of this, I think it is a very good idea to be a little more precise than what we currently have, which is really more of a set of functions and is not all that sound from, I think, a conceptual point of view.
I think if we went around the table, everyone would have a somewhat different conception.
It is funny, we all know what we are talking about. I think it would be the same animal, but we all have a somewhat different conception of it.
I think in the future, the future is going to present some alternatives to us that might go outside the boundary of what people have thought of as health statistics in the past, like administrative data, which I see as one of the implications of HIPAA.
It would be good to have a foundation for how to deal with that.
DR. FRIEDMAN: Dan had also used the frame road map. I do think there are a number of efforts right now that are going on with the states, with CDC, with NCHS, in other countries, parallel in other countries.
Canada is having a national effort to look at information needs in each of the provinces right now. I think it would be extremely useful, not just in the workshop, if we could specifically tie it into some set of continuing activity.
MR. HUNTER: If you go back to where Ed first talked to Don Detmer and Miron and others, I think the genesis of that was some rumblings that we needed to have some more focused kind of ongoing input into what we were doing at NCHS, and I think then elsewhere as the discussion evolved.
Then there was sort of a search for where is the right place to get that on which topics. I don't think we viewed this as three steps that end with some papers and a focus group report and a workshop report, but some sense of where you go in that broader vision process and the road map and so forth.
Then, also, what are the things that we collectively -- not just NCHS -- would benefit the most from an ongoing advisory input on, which brings it back to what the committee's role might be, and Miron's example of a possible panel study, or we once talked about a second workshop to be more direct in following up a road map or something.
That workshop ideally would end with a couple of specific paths suggested out of it that are part of a longer process.
I don't really see it ever ending, because we are really looking for some sort of ongoing process of looking at what the vision is, updated, and renewed in some road map.
I would think that the committee should keep track, as we go through it, with what it might do with whatever comes out of these things in an ongoing advisory kind of way.
DR. TAKEUCHI: I was wondering -- it may come out in other committees, in other papers -- but would it be useful to have at least some discussion about what our assumptions are about what the future is going to look like, that may have implications for these data systems or health statistics.
For example, like there is a notion that we will have greater globalization of economic continuity with different nations, and how is that going to impact the way that we look at health statistics, collect data, and will there be more sharing or less sharing.
If there is going to be more sharing, what targets do we shoot for to get more consistency across nations.
DR. SONDIK: I think that is very important. I mean, in health research in particular, it is certainly global today.
It is one of the major priorities of CDC to take a global outlook. NCHS, I can say, we in general, and I in particular, feel very strongly about that, both from sort of the substance of what we do and, even prior to that, the issue of standards, which we play a key role in.
I think that particular point is well taken. I thought during these focus groups, this would be one place where, in fact, these factors could be raised, in that kind of discussion.
One of the neat thoughts in this, I thought, was to have the paper authors attend these focus groups. In some sense, it would be sort of directed to them and other key participants in this.
This would be part of the discussion, as to what these factors would be that would be influencing the system.
Miron has -- I don't know, do you want to get into that draft that you have?
MR. STRAFF: Well, we have a draft of the agenda for the workshop.
DR. SONDIK: Yes, but that was a key part as well in that. So, in fact, you really put your finger on, to me, one of the exciting things about doing this, is talking about these various influences, and trying to get a bit better handle on them, and what their influence might, in fact, be.
There are different ways that we could do that, but I certainly thought this focus group was an innovative one to begin with.
I guess the point of the definition of health statistics was not quite so evident to me until, in particular, Dan brought it up. Then the more I have thought about it, the more I think this is an important point.
It is also another issue that we should talk about, is that there are no doubt other efforts that we need to account for and be cognizant of.
The CDC, sort of in particular and sort of in toto, wants to take a look at surveillance. I don't want to misrepresent that effort.
When I think about that, I personally think of surveillance as a subset of health statistics. I can tell you that people in surveillance do not necessarily see it that way. From at least my point of view, it makes perfect sense.
So, I think it is important to have these focus groups to look at the definitional issue. Really, the key thing, I think, is these definitional factors that are going to influence the system.
MR. SCANLON: I think the other perspective to bear in mind, I think, is when this originated, this was meant to be a perspective on the health sector in general. It wasn't meant to be the federal sector or even the state or local sector. It was meant to be the industry, public health community, research in general.
There may be trends in privatization or outsource support or not sharing data. There may be a lot of trends that may be underway in the health system generally that need to be taken into account.
The opening focus is meant to be on the health system generally. Then conceivably there will be aspects that deal with federal agencies or with state agencies or with NCHS and with other parts of HHS. It is very wide angle in terms of beginning.
MR. STRAFF: I guess I am a little bit confused by the use of the word focus group, which comes from market research, and where you really need to get at individuals and kind of probe their inner thoughts and inner feelings and get their reactions to things, and aren't just going to be superficial, and understand the reasons for it.
Here you talked about trying to get them to focus on what they think health statistics is, and then we talked about using it more as a brain storming session for where future scenarios might lie and what limitations they have.
I would suggest it is probably the latter that is more conducive to that kind of group setting. Larry, you have had a lot of experience with this. What do you think?
MR. BARTLETT: All right if I chime in? This discussion, I think, is right on point, as I was making notes as we went.
I was actually hoping we would come back to, Ed, the things you spoke about in the beginning, in terms of looking at the future and the factors and all of that.
I think that it makes a lot of sense, and practically. This, in my mind, we talked very briefly about definitions of expert meetings versus focus groups versus whatever.
This isn't quite a focus group. It is not sort of bringing in consumers -- a focus group, you are representing a segment.
You may do that, but I think you are representing people who know something, have a particular expertise.
I like the notion of beginning with what is the future, what are the implications, for a very practical reason.
The picture that I have got is stimulating discussion that is really far ranging, but yet, specific.
My vision of sort of talking about definitions and all, is someone wanting to write something up on the board and people kind if imploding, wanting to word smith and the like.
I have been sort of listening to this and saying, we talk and ask questions about what are the factors that they foresee in the future that are going to have an impact on health statistics.
If one gets them out of their box, you know, whatever sort of starting point that they have, you want to build upon that, but not have them sort of stake out turf and kind of talk about the same old same old, something that moves them ahead.
I am just sort of making notes and thinking about okay, we have a sequence of questions here, we have a sequence of questions there.
The questions that I have got in this, if I can just sort of throw them out, and they are questions that we talked about before, and they may be just me understanding the objectives, the whole process but also the objectives of the focus group, and your focus group expert group or whatever you want to call them, particularly in terms of how it links into your second step and how it links into your third step and where the whole thing goes.
I think you all are talking about that. I think this issue of who is coming to these things, it is sort of kind of a chicken and an egg.
I am not sure you have to figure out what is it that you want to do, what type of input do you want, and that is going to help shape your thinking about who comes to it.
There is a specific set of questions that, once this level comes down, you come to some general level of agreement.
Then we craft the questions. We don't go through lock step, necessarily, but there is a flow in terms of the future, it gets them thinking about that. What are the implications for health statistics.
What are the things that we are doing now that are most likely to be affected by these changes. It is just a series of questions that essentially you guys would feel comfortable if we got at them.
You want the group to be creative and to sort of flow, but at the end of the day when the dust settles, you feel that if we covered these things and we really sort of poked and prodded, we have covered the things that are interesting and are things that you want to be covered.
My suggestion is just keep going with your discussion. This is quite helpful and I guess I am sort of sitting back here and thinking, okay, when the dust is settled here, or at whatever point, and we begin to operationalize this thing, I want to make sure that you actually provoke some thinking and we will need to come back and talk -- not here necessarily.
What do you want to provide people ahead of time. What is the whole piece. How do you have them situated going into these things, what are the specific questions that you lead them toward. How does it fit into the next couple of steps.
I haven't heard anything being said here that I am sure wasn't helpful to me getting a handle on what you want to do with these things.
If you had to make a call about the picture of the meeting, I think you used the term brain storming. I think you really want to be provocative. I think you want people sort of thinking hard and being creative, but also bring them down with some specific questions to get at things, at definitions.
Don't just say what would be your ideal definition. Do it in a way that has them thinking about the 21st Century. I think that is a wonderful way to lead into this thing.
MR. SCANLON: He is talking about the focus groups.
DR. FRIEDMAN: So, starting with the more future oriented and then at some point coming back to the definitional issues.
MR. BARTLETT: If your definition is not -- again, I am the least informed here, and sometimes that is god awful and sometimes that is somewhat helpful at the margin.
If you want them -- you are likely, if you just sort of launch into it, you are going to have people talking about the definitions in health statistics now, you know, talk about things from their own perspective.
I think to the degree you can open their pores and really start thinking about the 21st Century and what is coming down the line and all that stuff, I think you get slightly different comments.
You get them being a little bit more wide ranging, and then you can kind of bring them back. I think if you start with kind of nitty gritty, it is real tough to go from here forward.
You have sort of said things, people hear themselves speak, and they kind of fold in on themselves. I like the way you let off the background in terms of why you are doing this. It is provocative.
MR. HUNTER: I was listening and I was going to ask the same question Barbara did, and that is what you meant by this meeting, by the focus group.
I think that is one of the things that makes this kind of challenging. Some of the same things might be said about the workshop.
Because there are different people involved, you may have different ideas. You may want them to address some of the same questions, but I kind of view this as maybe like a funnel in some ways, where you start off at the focus group levels and with the papers kind of generating ideas.
It is a small enough group that you can brain storm some ideas that everybody will just say, nah, and move on to something else.
By the time you get to the workshop, where it is a somewhat more structured discussion, because it is bigger, because you actually want some kind of, not a report with recommendations, per se, but something where there is a little bit more coalescing around ideas, you actually want to use the focus groups to start very broad in brain storming and kind of cull out some ideas that may be worth pursuing more directly at the workshop, and let that help shape the questions that you would lay out for the opening session of the workshop or something, but not necessarily come up with the five things that need to be reviewed by the workshop per se, because that is not really what the workshop is about.
MR. BARTLETT: It seems to me that you have increased your probability. You haven't sort of wiped out totally the notion that you will have missed an important issue or an important perspective or something that you really do want to use time at the workshop to focus on.
You may have some specific questions later in these focus groups about the workshop and how it should be structured, or what should we cover. You could go to that level of detail.
I guess my take from the earlier call, and I guess the way I am listening to it now, is that, indeed, this thing feeds in.
One of the biggest benefits is that you are more informed about how you would present, or how you would structure discussions in the workshop setting about key issues.
You would find maybe a couple of perspectives or a couple of key issues that, who knows, maybe might have slipped through the cracks, that someone flagged and said, we need to make sure that is represented there.
This is just me sort of grasping on what I have heard, one sort of flowing into the other and leaving you better prepared and more organized for your workshops.
It may be that you could ask them, well, if you had come to a workshop on this topic, what is the best way we could have gotten out of the blocks on this issue.
You could ask them at the end. Again, I put that at the tail end, how would we operationalize this if we wanted to have this type of discussion, but we wanted it to lead to some recommendations and we wanted to see if we could come to consensus.
Maybe you said you really think with a workshop of, give them some characteristics, which of these things do you really think we could draw toward consensus on. What are the other things that -- you could explore a whole bunch of things.
I just would sequence it later rather than sooner, certainly give them an idea of the big picture.
DR. FRIEDMAN: You know, Larry, as you are talking, it occurs to me that there is also another potential way of structuring it, which is we had talked about a couple of couple-day sessions with two different groups.
Another possibility -- I don't know if that is a good idea or not. One possibility is having one group that really focuses on the 10 or 15 years from now, what are the factors that we need to think about, and another group which is probably smaller and a much more down-to-earth working session that focuses on some of the definitional issues.
I don't know if it is a good idea or not. I would hate to lose the richness of one influencing the other. At the same time, I would hate to see the latter not get --
MR. BARTLETT: The question I would ask of you is this definitional thing. I appreciate what you are saying. I certainly don't know the area as well as I should to sort of know the ins and outs and the different views in terms of where surveillance fits, whether it is underneath or side by side with.
I think the big question that you may want to think about is the definition of health statistics with -- you know, it could be both. You start with the now and you go to the then, the down the line.
I think that you need to have some sense of, if you want to really -- if you want to devote a chunk of time to the definitional issue, is it in the context of what is likely to happen in the 21st Century?
If that is the case, then you need to lay out that backdrop and then sort of move into it. So, as you guys talk further, there may be some models that jump up.
We talked about two groups, possibly two meetings. I think we talked about the notion of going to the criteria in the second day.
It just never seemed to me to have enough meat to it, enough to warrant a second meeting. There may be some sequencing in terms of having vetted that scenario, having talked about these issues and how they impinge on these other things, now let's get down and dirty and talk about a definition for the 21st Century, something like that. I sure don't have the answers on that.
MR. SCANLON: What is the size of the group, the range of the size that is envisioned in terms of an expert group. It is small, nine or fewer, or is it several dozen?
MR. BARTLETT: If you don't mind, let me throw in my two cents. When we talked on the phone, you can do it any size you want and the product will vary.
You have got three factors: what you want to cover, how much time you have got and how many people, and then all the other variables about how strongly do you feel about X and Y and Z.
I think that anything above 12 to 15 is a mistake. I think certainly a half dozen or below, I would just imagine, without giving it much thought, you don't have a good enough mix to generate the kind of brain storming and different views and all that stuff.
I think you are probably talking nine to 12 ideal, nine to 15 doable, and I wouldn't go past that. I wouldn't go past that and I wouldn't go below it.
DR. SONDIK: I was just going to say that I just don't think that in the time when one could do a workshop, that there would be enough time to necessarily bring out all of these ideas.
I don't see these things that come out of this are necessarily conclusions. Like any sort of a brain storming session, things sometimes, in the heat of the brain storming, look like the world's greatest idea.
Then, on the drive home, you scratch your head and you say, my god, that doesn't really seem to make any sense at all.
You need to really sort of sleep on these things. There are so many possible factors influencing this and futures, and definition may be an issue.
I think in large part it is a definition because in any kind of endeavor we have different institutions involved.
There is a question of sort of how you divide things up, I suppose, and then definitions become really important.
What I hope comes out of this, out of the workshop, is a guide to first of all the department and the field in general, for how it should allocate its resources, in some sense.
For example, we are putting in, in some areas like NLM, we are putting in a lot of resources into informatics.
It is wonderful. It is having tremendous pay offs. In other areas of the department, we are not putting in the resources that I personally think are really important. That is my personal opinion.
What I would like to see out of this is, in some sense, a road map. It may not be of the precise quality of a AAA triptych, but it would be something that would give a kind of a guide as to where resources should be going, resources of research and development type, and in the field, as I said, in general.
This would be useful to states; this would be useful to the private sector. Then, as we said, we have talked about a second workshop that would be more specific to NCHS and that I would look forward to that and that would be even more specific guidance.
I want to just sort of put that emphasis on the idea of allocating resources. That is kind of how I see it, something that would be very practical, I think, for the field.
MR. BARTLETT: If I can just add to that, any clarity you can provide on the workshop has wonderful dividends on shaping the focus groups or expert meetings.
Hearing that, that is the sense, then, that would move me and I think move us, at least at the tail end of the focus groups, in perhaps a slightly different direction or certainly add a series of questions or just somehow move toward that.
I guess, you know, the question is, are we talking about the focus group or are we talking about the workshop.
All the decisions that you make or all the clarity that you can provide on the workshops, to the degree that the focus groups are supposed to lead into that and inform that, you will also have the back wash.
The more that you know ahead of time about what you want to do on the workshop, it will help; the more you will get out of these focus groups.
MR. HUNTER: Just to clarify. I think I am hearing you say this. When you talk about coming out of the workshop with like a recommendation on, as a theoretical, investing more resources in informatics in places other than NLM, I am taking that not as a specific recommendation that money be spent here, but the sense from the workshop that informatics is important and should be given attention.
DR. SONDIK: That is exactly correct.
MR. HUNTER: One of the things that is also hard here is to think of post-workshop things where you take that, and then maybe the Data Council or somebody else is the place where you take that and say, okay, here is the broad theme from this group of eminent people. What do we do about that specifically, in translating that into a decision within the department.
I mean, you don't put that kind of pressure on the workshop itself. You want the workshop to tell you that, no, this is an area that we need to go back home and work on.
There may be multiple activities coming out from the workshop. There might be a workshop at NCHS, there might be more focus groups where you actually take a specific thing and go back and say, okay, informatics in non-NLM places, what do we do about it, what do you all think, what are the specific kinds of ideas.
I just want to make sure we weren't being so specific that we wanted a budget recommendation, but just a kind of a thematic one.
MR. SCANLON: It is more directional.
DR. SONDIK: Biology, for example, is that the kind of -- are we considering that enough in health statistics these days, DNA, the implications of that.
The devolvement and the specific groups that can be defined in any number of ways, if we continue in the direction we are going, will we be getting the kind of information that we really need, or should we be considering some alternative strategies to give us that information; topics like that.
Ed's point, I think, is well taken. I am just taking it to the next, especially because we were discussing budget this morning; I have that on my mind. The idea is, as Jim said, of general direction.
MR. STRAFF: We would like to talk about the workshop. do you want to distribute our draft? We may need to ask you to share some copies of this.
I think our goal for this workshop to try to assess -- obviously not in a complete and comprehensive way -- where health care is going, where health is going, and what the implications should be for the health statistics system.
We have organized it along the lines to have presented the work that this committee and the Data Council are commissioning.
After an introduction, focus on your case studies, under health care financing reforms and the children's health insurance program, what needs were met, what needs were unmet.
Then it was a little bit unclear, to have a presentation of other commissioned papers. Here you have the international comparisons and some others that you mentioned. I don't know how far along you are.
There was some question within our committee that the international comparisons might not be so relevant here because our health and health care system is so different from that of other countries, how we manage health statistics.
I think we might want to have it as background. We can certainly present it, but we may not want to devote a lot of time to it. That is one of the questions.
Then we had a session that we were wrestling with, because it is impossible to make this comprehensive.
That is, where are we going with health and health care. We worked on this and we divided this up into three broad areas -- health status and the determinants of health, the organization of financing of health services, and the changing environment for health services, changes in health policy and technology and other influences on the system.
These could be ideas that you could use for your focus group or brain storming sessions as well. We would either consider commissioning a paper in each of these areas, or just having a small panel discussion, having two or three people talk about them.
The business of health status and the determinants of health would include genetics, behavior. We may want to talk about diet, smoking and the like, as well as psychosocial aspects of health. That is an area that we can push the boundaries as well.
You can imagine just two or three people, expert in that, talking about that generally, and then opening it up for discussion.
We can't look for a consensus here, about which is the most important area. You are going to have to walk away and make those decisions yourself.
It might wake you up in terms of thinking about it in different ways, or learning from what other people think are important trends in health.
I think we want to cover issues like privacy and confidentiality, and especially the changing role for federal, state and local data relationships, in particular with regard to the evolution of public health programs.
Finally, hear from two important users, one from the non-federal perspectives, and from the federal agency perspectives.
Conclude with some wrap-up discussion of where do we go from here. Then Ed Hunter's question, what the next steps might be.
This is a very preliminary draft that we are working with. I wanted to bring it here to you and get your comments about it, to see how it might fit in with your plans for your commissioned papers, the focus group discussions.
DR. SONDIK: Can I make a comment? I think in particular the -- in some sense, I suppose this is the meet of this. It is the panel discussions, particularly the non-federal perspectives. It is really important.
These are the links that somehow I feel weakest on. You know, we have all of this organization that helps us link to government and various levels of other organizations.
I feel the links to the people who are outside the government, these are the ones that trouble me. Clearly, a part of this is state and local health officials, but only a portion of it.
I think that is very appropriate, that it be there, but only be a portion of it.
MS. WARD: I think my concerns are health care provider. I mean, the care system is so complicated, I can't imagine you are having a provider commenter.
You have got providers and plans, and plans are sometimes providers and providers are sometimes plans, and plans are one huge part of the system, and they would not see themselves as necessarily represented.
The delivery system is not represented here. I think you have to just -- not that I have the answer, I just have the problem statement for you, that this is not the delivery system, that it cannot be represented in this way.
DR. NEWACHECK: That is probably true for all of these.
MS. WARD: That is true. I understand. Research is very complicated. You may have to trade off, because you can't spend three days just have commenters going on and on and on, one being this and one being that.
MR. STRAFF: For each of these topics we could conduct a different panel study, and for some of them we have.
It is just a question of being able to allow some ideas to come to the table. We will miss many.
DR. NEWACHECK: If there is good audience discussion, then there is an opportunity to bring up things that might not be brought up by the person depicted as health care delivery or health services research or whatever. I hope there is time for that.
MR. WHITE: We would do that. The key would be getting the right kinds of discussants and commentators.
Mr. STRAFF: The right kind of people in the audience. How many people. What is the minimum number of people who could represent the delivery system, just a theoretical.
MS. WARD: It may be just a labeling issue. If you took the term provider out there, and had some sort of presentation of how this relates to our new and futuristic delivery system, that might solve the problem, if you took the word provider out.
MR. BARTLETT: I think your point is well taken. I just came in off the road from doing five days of meetings around the state of California on privacy issues, and the notion that a plan and a provider would cover any of these things, the line is blurring, but they had markedly different features.
MS. WARD: And in most places they are at war with each other. That is certainly true in my state. We are sometimes friends, but most of the time, that is the biggest consumer comment you will get is, I have never seen health care as antagonistic in my life. I can feel the fight going on between my health plan and my provider when I am in the provider's office.
That may be the future that we want to change, but that is the future that people are feeling apprehensive about.
MR. HUNTER: That point applies possibly to most of these. I can't think of the person who can represent privacy. Is it the pro or the con or the moderate, if there is such a thing.
It has got to be thought through. Is the premium there on getting the constituency represented at the table, or is it to make sure that someone is there representing a range of interests in that area, that someone is there to reflect on the privacy debate and what it means, from a more neutral bystander --
MS. WARD: That is why I think if you took the word health care provider out, but talked about that as a commentator about the health care delivery or medical care delivery system, then you have really got an entirely different thing.
What is more equal to those other huge areas that are also contentious.
MR. HUNTER: It does raise something of an issue of how much we would want to have the key stakeholders of these different groups represented, as opposed to the point of view they hold reflected by someone else. That is sort of a fundamental difference in direction you could go.
If it is a summit meeting of people who have a stake in this as opposed to a group of knowledgeable people on health information or the health services area that come together to make sure they understand all these and then decide what to do about it together, it is a different meeting.
DR. AMARO: I think the other group or perspective or set of interests that you would want to include in here are policy makers.
I guess state and local health officials would kind of be seen as part of that, but probably more broadly.
We hear all the time from the elected officials who come in front of us, how the data systems always or often don't really answer the questions that they face, and they have to often make decisions about it.
DR. NEWACHECK: Part of that, I think, would be reflected in the second panel.
MR. HUNTER: Those are producers.
DR. AMARO: The users, the legislators. They really are users of the data. They have to look at it and make decisions.
MR. BARTLETT: How many participants? I am sorry, my question was, how many participants and can someone repeat the objectives?
I got the sense that you really wanted not maybe consensus, but some recommendations, something coming out of this. That takes time.
MR. STRAFF: We are not looking for recommendations. The Academy can't issue a report of recommendations from the workshop. It needs to be a more deliberative effort.
We do want to have as broad a sweep as possible, and we want to get imaginative people to give us new ideas, some of which you will toss out on the way home, but others of which may expand the boundaries of health services.
MR. WHITE: And lists of potential next steps.
MR. BARTLETT: How many people in the audience?
MR. STRAFF: Well, we always start small and we will probably go up to 60, even though we would like to plan for something like 40.
DR. STARFIELD: You just said something that I was going to say, so let me second what you said.
The more I hear the discussion about these constituents, the more I think that we don't want to go with constituents if we want imagination.
If we get constituents, we are going to get what they are missing now, and not where they think we are going to be 10 or 20 years from now.
I think maybe I wouldn't frame this by the constituency but frame it by imaginative people, like you said.
DR. FRIEDMAN: On the one hand, I would second that, Barbara, but at the same time I am coming back to what Hortensia has said.
For me, when I think about health statistics, I think of it as data driven by use for policy making and program development.
DR. STARFIELD: But the use is going to be different soon, than it is now. The system is changing so much.
DR. FRIEDMAN: That is true, but still, there needs to be some sense of what is the utility. I think that to the extent that at the state level we have had trouble selling the systems, and at the federal level, I think one of the reasons for that is that the utility isn't self evident to the people who are paying the bills.
DR. STARFIELD: It is form and function. It is whether form drives function or function drives form. I think you really want -- you want to think about it in terms of what we need to know and what kind of information we can get, and that may change the form of the system.
DR. AMARO: Even going with your suggestion of not having a representative point of view, but somebody who articulates what the needs are, I think it would still be useful to have somebody thinking about what are the needs that policy makers have, that would help us look at what should the system look like in the future.
DR. STARFIELD: We have policy makers aren't on here. This is sort of structured differently.
DR. AMARO: I agree with you, that I think if we can be a statement of what the perspective is from the different constituencies, that would probably be more useful.
MR. HANDLER: About four or five years ago, I heard a presentation made by someone from Statistics Canada.
Statistics Canada takes care of a lot of different departments in the government of Canada. It is not just health related, but it is cross cutting.
Basically, the way he described it may have been a little simplistic. What he said was, when the cabinet officers in Canada get together, there is someone from Statistics Canada in the room.
When they are discussing different things, and they need an answer, they need a number right away, they point to the guy from Statistics Canada, who is sitting there with a lap top, and he gets the information.
That is a different way of thinking. Maybe that is what we should be shooting for in 10 or 15 years.
MR. SCANLON: Harking back to the CNSTAT workshop on data needs for welfare reform and health care or market changes, you specifically invited big thinkers, I remember.
We had some folks who were really meant to be looking at what the next few years held and what the big policy questions would be. I wonder if there is a comparable group?
MR. STRAFF: I think that is how we see that section in the middle. I think yes, we will need to invite these visionary imaginative people and give them free rein, and then see, indeed, if what they are saying has implications for the health statistics system and, if so, how we might deal with it.
MR. HUNTER: Is the idea that you would want big thinkers that cover all the perspectives of the constituencies without coming to push one of their specific agendas? That is a very small group of people.
My concern, one thing I want to make sure we talk about before we go to far on this is what your sense of time table is on this, thinking also that we have to back that up to the focus groups and coordinate the papers.
MR. STRAFF: We haven't scheduled a date, but I think we were thinking of spring at the earliest, which would be around June.
If you have papers or other activities to feed into that, and if you wanted it earlier, we could.
MR. HUNTER: Could you give us a sense of backing up from that? Say it was June. At what points before that would you need to have feedback on agendas or finalized agenda, or what is your process for doing that and getting invitees named and that kind of thing.
MR. STRAFF: We really should start lining up participants three months ahead, and even earlier if we are going to have some of them commission papers.
We will have a subcommittee look over this. This is your opportunity now to make input from this.
I should say that Ed Perrin(?) is going to chair this for us. We are not sure how we might handle this, because we are not sure how we might structure this.
If we decide to do that, we might have a panel discussion and invite the people to give us background papers, research papers they have already written, that would be a basis for their presentation, or just have slides prepared and have them distribute copies.
MR. HUNTER: If we had focus groups in probably January or February before the last would be done, would there be opportunity for that to help, given your time table and your approval processes.
Would there be opportunity to feed March/April into an agenda for a June meeting?
MR. STRAFF: I think so, because we could always put it in the agenda book that we mail out of background papers.
If we have to work on it and do some more synthesis and put it into another product, that is a different story.
MR. HUNTER: I was thinking more in terms of helping shift what relative proportion of the time or focus or questions or something. I mean, the focus groups are going to generate ideas. Ideally, they could help change the agenda.
DR. FRIEDMAN: Miron, on a purely pragmatic level, is July a workshop month or is it not a workshop month?
MR. STRAFF: It is possible but more and more difficult. It is harder to get academics in July. We have been successful, but I wouldn't push it later than July, late June.
MR. HUNTER: The NCHS biennial conference is in July of this year. It will be well hidden.
MR. WHITE: Well, you need to know as soon as possible.
MS. WARD: Are there specific things that you are thinking of?
DR. FRIEDMAN: There are a couple of things in June, two or three, including this committee meeting. HSR, NAPHSIS, we may start bumping up against people saying, I just can't go to one more meeting.
MR. HUNTER: And they are the two broad thinking selfless people who are out there.
DR. SONDIK: I would like to keep it prior to fiscal year 2000, if that were practical.
DR. STARFIELD: I have another thought on this constituency. Maybe it is useful to have these points of view, but from people from without the constituency.
Is there someone who can think visionary-wise about state and local health departments, but not somebody who is from within one.
MR. WHITE: It would be nice to do that. That is one thing we might be able to add to the debate, is a non-stakeholder take on the whole situation, a room full of stakeholders in a tank full of sharks.
DR. FRIEDMAN: In terms of your process, is this now the opportunity to comment on this?
MR. STRAFF: We would like to take this back to a subcommittee. We should really like to think about what the timing is. If it is going to be in September of this year, then we will have more time.
Planning something in June, it would be nice to have an idea of the agenda for us to get the approval, and then for us to start to line up people. We haven't even decided if we will do this with commissioned papers. If we do, we would like to get started. Maybe three weeks or so to give comments. Before the holidays, the December holidays.
DR. FRIEDMAN: My guess would be, just based upon the conversation -- I can't speak for anybody else -- I think it is a real good outline.
I wouldn't be surprised, at the end, if it looked different, but I do think, from my perspective, it would be helpful if folks could have a chance to digest it and think about it and put in some additional comments; that is all.
MS. LI: Can I ask a question about the scope of this? One area that seems like it is not covered are federally funded data collection activities that collect health data.
There are some fairly major national data activities. I am wondering to what extent that gets brought in.
Out in the field, you hear a lot of frustration, sometimes, of the government being too slow to collect the data that you want, or the government is uncomfortable asking these questions.
So, principal investigators come in with investigator-initiated grant proposals and they fund big data collection activities, and then actually go and do these things.
Then sometimes they are co-opted into the government and become government surveys. The NSFG is one example of a survey that began in private hands and, over time, landed at NCHS.
I am just wondering how those issues can be brought in. It seems to me that that is an important part of collecting health statistics, that doesn't always get integrated with health statistics.
MR. WHITE: Is that something that could be part of the last session, where you discuss next steps and try to come up with this fuzzy road map for research?
The question is, could some of the things discussed in the workshop bring to mind that mechanism as the best way to collect data.
DR. STARFIELD: That is something that came up in the committee meeting discussion this morning, that there is an increasing number of these non-government foundation surveys, just lots of them.
Not only are they directed at the same thing, but in a different way, with different approaches, different questions. I mean, you can't put any of it together.
MS. LI: I guess I put into one kind of area the ones that are nationally representative, large data collections, that lead to public use data tapes, and that are meant for public analysis or other researchers using the data very quickly after the data are collected.
DR. STARFIELD: I am talking about those, even. There are just lots of them.
MR. SCANLON: That is why we want to focus on sector wide. We don't want to focus specifically on federally sponsored data collections.
If one of the trends in the future, as in the past, as investments pay off for the foundations, if they want to pay for more data collection, that is a trend that ought to be discussed.
If that is something that we really see as lasting and desirable and is serving the public good, that should be part of the discussion, I think.
The other issue of whether the Federal Government supports a research survey through a grant, I don't really see that all that much different, in terms of the public good nature and the public use of it as a direct federally sponsored one.
I see that as part of the research, part of the portfolio of information that people might think is useful, and part of the trends that we see.
DR. SONDIK: That is a really interesting point. That goes to that paper that we haven't identified anybody to write yet, the robustness of the institutions, or the components.
There is a lot of discussion of the role of the foundations. Part of that discussion is, well, they are in it today and then they are out of it.
The same is true of many of the grant supported activities, that are there now. In some sense, there are different purposes for those people doing those grant supported activities versus the ones that we do at NCHS.
There is a big overlap, but there is a significantly different purpose on the part of the principal investigator sometimes.
Look at the nurses study. That is a really interesting one, the Harvard one. That is supported by a grant for now, for how many years? A hundred. Twenty.
I think this would be a good thing to put into the mix for the brain storming sessions, and as well to make note of here. It is hard to even kind of categorize them.
I think it is an excellent point and it really speaks for making this thing broad to begin with.
DR. STARFIELD: Maybe you also have a lead on who should write the robustness paper. Maybe it ought to be someone like Steve Schroeder or someone from a foundation that is putting a lot of money into a survey, and ask them to commit themselves to how robust they are.
MR. HUNTER: There was actually a very interesting session. I was part of the Public Policy Association meeting last year. Henry Aaron chaired, talking about with the growth of philanthropy to get through the task laws, how many new foundations were being formed all the time and how much money was in the foundations.
This discussion was sort of a brain storming thing, but are these crowding out more serious, long-term efforts that might be robust because we can't go and say, we want a survey of this, because somebody else is doing it. Once we can't go, we can't go back.
One thing, I guess, that I take from this discussion is that any place where it implies we are talking about specific data sets or specific solutions, we probably ought to change that to be somewhat more generalizable about the field.
It does really also suggest that in our focus groups, at least, we bring in some of these foundation issues and non-traditional collectors of information.
MR. STRAFF: The regard the federally funded and the federally collected surveys as part of the purview of health statistics.
I think, you say, we cover them in broad areas of health related activities that we are discussing and directions and ideas will naturally arise.
We are not going to reach decisions at this particular workshop as to whether data should be collected by grant or contract or research or foundation.
Also, keep in mind that we are thinking about another follow-on workshop which would be directly specific for NCHS, that would look to help them prepare better for the future for those activities that are within their purview.
So, when we do focus on that in that workshop, it will just be for the NCHS, and not other agencies. This first one, we are talking about health statistics.
MR. SCANLON: In the stage setting part of maybe the first one, just as we would want to have people talk about and think about the future health care system and so on, we may want to have them think about future directions in data collection as well, like foundations, like potentially the industry itself.
They may or may not be trends, but they would be something for people to think about.
DR. NEWACHECK: If they predict the stock market.
DR. FRIEDMAN: There are two or three relatively operational issues that I think we should spend a few minutes on, one of which is what -- you would like some coordinated comments from us?
MR. STRAFF: However I can get them.
MS. DURCH: Including ideas for people that you think might be either participants as panelists or to attend.
MR. STRAFF: Jane is our study director for this. Jane, maybe you should give them your e mail address, in case anyone wants to send you suggestions.
MS. DURCH: It is jdurch@nas.edu. I would be happy to hear from any of you who would like to contribute your thoughts.
DR. FRIEDMAN: So, that is one operational issue. The second operational issue is our remaining paper that we had addressed, that we had tried to entice Dorothy Wright into writing. She unfortunately refrained.
Ed Sondik and Ed Hunter and I and several other people had had a conversation several weeks ago, where we talked about the differences between that remaining paper and the workshop and the focus groups.
We had talked about the possibility of if we were going to focus specifically on robust systems, speaking with George Van Amburg about it.
George is not around. I don't know where or when he will be returning. So, we are really where we were several months ago, other than we have crossed several people off the list. We need to find somebody. So, any suggestions would be appreciated.
The third operational issue is, I think we are talking about January to start the focus groups. In addition to narrowing down or specifying a list of questions, we clearly need to also start talking about starting to invite people, and talk about dates, et cetera, et cetera.
DR. STARFIELD: Which one do you want to take first?
DR. FRIEDMAN: How about the remaining paper.
DR. STARFIELD: I guess maybe we need to think about a different kind of person than we have thought about.
I just had those thoughts about a person from a foundation doing it, not to give us the foundation point, how they perceive the need for robustness.
I think if you get the right person, you might get a good paper.
MR. SCANLON: Is the goal here to try to describe health statistics system that have sort of proven the test of time and utility?
DR. STARFIELD: First of all, how important it is to do that. We are assuming it is, and I am sure we will come out with that, too. That had to be defended.
MR. SCANLON: Is the idea --
DR. FRIEDMAN: That was my idea. My idea was basically, hey, we have got some things that have been around, and we have some things that have proven useful, and that have proven useful for a long time.
We have got some things that have provided data over time, where 150 years ago the people inventing the systems were probably not saying to themselves, in 150 years, I want to have data, you know, comparative data on infectious disease incidence for the next 150. Somehow or another, those systems have worked.
That was my idea, basically saying, what is it in those that has given them the strength to survive and adapt and still prove useful.
DR. NEWACHECK: Is it the measures or the systems?
MR. WHITE: The utility and the output are the two that jump to my mind. Infectious reportable diseases, the utility is immediate in preventing epidemics.
Vital statistics, the utility is essentially legislative. People are required to identify themselves and birth certificates and death certificates are a real way of doing that.
DR. NEWACHECK: It is also because they are responding to people like PIs like Paul Ginsberg who want to do the studies.
I am not sure that Steve Shroeder, for example, really thinks about health statistics. He uses them occasionally, but it is not something that he focuses on.
DR. STARFIELD: Suppose you gave him a topic that said, do we need robust systems and what would they look like.
DR. NEWACHECK: I think he could probably comment on what sorts of information needs he would see the future of the health care system needing in a very broad way.
Someone else would have to translate that into data systems or statistics. I don't think he thinks in those terms. It would be very difficult for him to do that, or the staff would have to do it for him.
DR. SONDIK: We should get a Moynihan.
MR. BARTLETT: I am not sure this is why we need the workshop on the definitions. I think I am hearing you say, Barbara, that you think it is a debatable assumption whether a system has to be robust.
I am almost hearing you say that a survey or a specific data collection mechanism, when some people may be thinking -- me included -- may be thinking that the system needs to be robust.
Whether you are relying on the same data source and the collection methodology is a different question. For a system to be robust, it may need to change the way it collects information.
That is how infectious disease surveillance has evolved. It is no longer just a physician going into the laboratories, and electronic laboratory surveillance. It is doing other things, to make it available and vital, to use the wrong definition also.
DR. STARFIELD: I think you put your finger on why we are having a problem getting someone. Maybe there aren't those kind of people around. We have to think of these subsystems.
MR. BARTLETT: On you invite list, it seems to me that in some parts you are looking historically at systems that have been robust.
I like the notion that they continue to exist because they are able to adapt. Again, if you are looking at this in terms of, what does this mean going into the next century, someone who is used to looking forward might be able to take a look back, what systems might need help, but might give you an interesting objective look at how they have endured, how they have adapted, and what that means in the context of what may be coming down the pike.
I don't know who that would be, but maybe as you look, it sort of broadens your fields in terms of where you want to look.
DR. NEWACHECK: How about somebody like Cliff Coss(?), who has been in this a while? You don't think so? He is outside the --
DR. STARFIELD: He is so wedded to a particular data system.
DR. NEWACHECK: How about David Eddy?
DR. FRIEDMAN: That would be great. I don't mean that lightly; that would be wonderful. Is that likely to happen?
DR. STARFIELD: It is so far afield from what he does.
DR. SONDIK: Except that he uses this information all the time. It is possible.
DR. FRIEDMAN: We should assign Ed to call him.
MR. SCANLON: We could get some people at UCSF.
DR. SONDIK: The more I think about it, the more I think someone who is a really sophisticated user of this information is maybe what we want.
DR. STARFIELD: Of multiple kinds.
DR. SONDIK: Yes, and who is able to not see this thing solely in terms of federal activities; see it across the board. The points are brought up about grant supported or wherever the information comes from.
DR. FRIEDMAN: I do think that David Eddy would be wonderful and absolutely terrific.
MR. HUNTER: Plan B would be?
DR. FRIEDMAN: We have to come up with a lot of Plan Bs here, Ed.
DR. MOR: Jim House at Michigan? He is a big user.
MR. SCANLON: I wonder if Plan B should actually be maybe get someone who is knowledgeable enough, but have them identify some cases that were successful.
In other words, you don't need a senior statesperson to write this. You can actually have someone who is very knowledgeable and sophisticated in statistical areas, but have them talk to people about what they found, maybe three or four very good successful, responsive statistical systems that have held up over time.
DR. FRIEDMAN: I think you are right. We also considered that, and have difficulty coming up with someone. Any suggestions from anybody here would really be appreciated.
Not only are we not on plan B any more, but we have been through a lot of alternatives here. I personally think it is a really interesting paper.
MR. HUNTER: Another option might be, George may still come home from wherever he is. If we don't find somebody to give us specific examples, I think something Miron and Andy can think about is, is this worth thinking of this in terms of other statistical systems.
In some respects, we are talking independent of the subject matter. We are talking about the concepts and the nature of the system itself. There may be examples from other parts of the statistical world that could be brought to bear on that, too.
I mean, there are federal, state cooperative ventures at BLS, there are different models of using private sector data in different agencies. The health accounts are a different kind of model.
MR. STRAFF: Every agency differs. If you just look at federal and state qualitative efforts in data development, they are all different.
MR. SCANLON: That is a whole area in its own right.
MR. STRAFF: You wouldn't get the guidance from them, I don't think. You can talk about general principles.
MS. GREENBERG: I don't have plan X, but I was sort of thinking the same thing that Ed was. It may only make the paper even more difficult to write, although it may call for Jim's scenario, of not just focusing on health, certainly, but having one or more examples outside the health field.
If we took Jim's approach, then we could think more in terms of the consultant, knowledgeable writer.
MR. SCANLON: I think I know someone who actually worked at -- we were thinking of getting him connected to the committee for some other work as well -- I don't know if I should mention the name -- Dan Melnick(?).
Do you know Dan? He was the head of applied studies at SAMHSA. He worked at the CBO, statistical policy decision, before that.
He is knowledgeable about the economic surveys. I think you guys know Dan. He would have to rely on people giving examples of robust systems.
I don't know if you think that is a good way to proceed. He is retired now. He has set up an individual consulting company. He could actually do consulting work for us.
I could ask him, if you think this is a good idea. It is plan B, you are right. It won't be an expert per se writing about their own views and measurements.
It would be someone who is knowledgeable about a lot of statistical systems, who has had a lot of experience, also worked on the Hill, and in a health agency.
He would probably have to talk to others about what do people feel are successful systems and what are their characteristics, something like that.
MS. GREENBERG: Some of these systems, we have some literature out there about them, so we can draw on that as well.
DR. FRIEDMAN: What do you think? Would you rather start with yet another plan A, as in Eddy, and then make a quick run at it, and then if it doesn't work, then we could --
DR. SONDIK: I could certainly do that.
MR. HUNTER: It seems like it would be hard to get somebody -- maybe if we had gone back to the original plan A and told Dorothy that if she had the ideas, then Dan Melnick would write the paper, the same thing might hold with somebody like that.
You might suggest that we will provide staff support with a knowledgeable person like Dan, and he will do the leg work and put this together, if she can help generate some ideas and give some direction.
MR. SCANLON: I think Dorothy's problem is she is continuing as the chair of another panel at the Academy. I don't think she is going to have, under any circumstances, the time. I assume she is continuing.
DR. STARFIELD: In any case, we would have to suggest to them people to talk to, which could include these people.
MR. HANDLER: I am formerly from the Bureau of the Census. They are wrestling with the same kind of problems, I guess, that we are here. What data do we collect. Who is our user population. How do we disseminate information. How do we coordinate activities with state and local people. How are we responsive to their needs.
Maybe you can get someone from the Census Bureau, from their perspective. They are wrestling with the same type of issues, but not health related specifically. That is just one suggestion.
MR. SCANLON: I think if they are part of the same statistical system, we might want to look at them. They have had some successes.
I think I would rather have someone else maybe use them as examples, but not necessarily have them write it. It just complicates matters, if another federal agency does this.
DR. FRIEDMAN: I do think we have a second plan. You will talk to David Eddy and see what he says and then move on.
Focus groups. Larry, when do we need to nail down dates and invite people?
MR. BARTLETT: If we are talking about doing it in June, in January and February. I believe we have had a number of iterations. I think certainly before you go out to anyone, we need to have some reasonable expectation -- this is two days that we are asking them to participate.
You have a number of different permutations on this, two groups, two days. We talked through this, Dan. You know, back to back, a break, make a decision whether we need to go back to them for a second day, whether you do four groups one day apiece.
I think you can do an awful lot in one day, a kind of 10:00 to 3:00 kind of thing, particularly if it is organized well.
Getting someone for one day is a lot easier than two back to back. I think the sooner the better, but certainly you want to give them at least 30 days and you want to have back up in terms of a list of people that you call in terms of representing different perspectives. I think that is the issue.
The first issue is who do you want. Do you want some of your end users. Do you want people familiar with working with the data sets. Do you want statisticians, researchers, other type of end user folks. I think those are some of the questions that come up.
I think putting your list together and then probably right after Thanksgiving trying to nail something down, because it is going to take a little bit of time, and then stopping when you feel that you have got the mix that you want.
DR. FRIEDMAN: So, it is essentially one day with the renewal option. You are likely to get no takers.
At this point, I think you are going to have a tough time getting someone to sign up for two days in a row.
My first thought would be to see if you can be really highly productive in a day. That still gives you the option of going out to a third and a fourth group or even combining the groups or doing something along those lines.
DR. FRIEDMAN: I think that is a good idea.
MR. HUNTER: Does that suggest dividing the meeting, maybe having three or four different types of discussions, or having the same discussion twice with two different groups?
MR. BARTLETT: With a group that size, and if you are trying to do it in a day, and I have just a fuzzy notion in my mind of the scope that you want to cover, from the big picture, what is the future looking like, perhaps far more specific questions, you know, if you were making this decision and you had these resources, where would you direct them, or something that doesn't get into budgetary issues, but nonetheless gets to prioritization, I think that is a pretty full day and a lot of different things.
I think two of the sessions should be the same general material. With 15 people, you guys are statisticians. I think you know the probability of getting different responses. I think it is pretty high.
DR. FRIEDMAN: As we are talking, my sense would be maybe two on the more general future oriented, and then maybe one on the more highly specific definitional.
MR. BARTLETT: Maybe pulling from those first two as kind of the yeast for a slightly different configuration for that third, more focused piece.
MR. HUNTER: Then you stage these -- I am thinking of that funnel, again, that I can't get rid of in my mind. You may have a more specific one that may use some pollen from that first event.
MR. BARTLETT: That is the February event?
MR. HUNTER: March event if the workshop is not until June.
MR. BARTLETT: If you do have that time, part of me really does think that you might well make use of that additional flexibility, due to these, and then stop and say, what do we get. It is probably going to generate some new questions on your part.
Who is good at those last two, and maybe we can impose upon them to come in and then sort of go that next step, certainly rather than trying to do it all in one fell swoop and find that we really needed to step back and we could really make use of these smart people's time.
DR. FRIEDMAN: We do have -- and Lisa and Karen -- have pulled together a list of names that have been mentioned.
What I would suggest is -- and I know we don't have time now -- but if folks could just go down that list and respond with any thoughts.
Obviously, a lot of it is going to flow depending on how we define the focus groups. It sounds to me like we are sort of lurching toward a definition of how we want to define the focus groups.
DR. SONDIK: It could be as few as two or three or four, but if you have a sense, if you stop and think, in a focus group it would be great to have this perspective, that perspective, whatever, and essentially try to establish your categories, and then you can start throwing people into that.
I would like to see some of the people as identifiable experts in these areas that will influence the future, not necessarily experts in health statistics systems.
DR. STARFIELD: You are not referring to the list right now when you say that?
DR. SONDIK: I am referring to people who would come to the focus group. Names don't jump out at me who are -- let's see, somebody in -- I am probably selling people short -- but in health care organization, or informatics, or biology. That is what I am getting at.
DR. BARTLETT: Nobody knows the time, but I guess at some point I was going to really encourage folks, people might have what they consider to be fairly specific questions that they would love to put on the table at these focus groups.
I think to the degree that people have the opportunity to do that, throw them into the hat. Then we can sequence them, we can shape them so that they have the right tone.
It will actually probably give us a better picture of what we want to do at these meetings.
DR. FRIEDMAN: Absolutely. I think a next step -- and Ed Hunter has taken that first step -- is basically re-drafting the earlier work plan, but focusing more on the objectives.
I just got it earlier this week. I think we need to circulate it and start doing that.
MS. WARD: Should we send names?
MS. BRIGHTMAN: Send the focus group list to me, to Lisa, at lcm9@CDC.gov.
DR. AMARO: Dan, what is the profile of the people that you are looking for, for this expert focus group, since these are the people who are supposed to see the future.
DR. FRIEDMAN: The profile has changed, and I think it has changed today.
DR. AMARO: To?
DR. SONDIK: I will give it a shot. First of all, I think we need some core there who do know something about the health statistics "system." They have got to have some familiarity with these.
I think someone who thinks future-like about health care organization and delivery, somebody like I am looking at Wendy Baldwin here, who is the first name on the list, certainly knows about health biomedical research and the direction that is going in.
She might actually fit my second category, which has to do with biology, DNA and that sort of thing.
Somebody who thinks about informatics -- I love this term, sort of like cybernetics used to be, but you know, information systems, using information, data structure, data analysis, data mining all of a sudden has really become -- at least the data mining has really become a field, and something that people used to speak about with disdain. My god, you don't know what you are after; why did you collect this information.
Now, this has really changed. This is a very legitimate idea. When you think about it, you go out on the internet, a lot of what you do is sort of data mining in a sense.
You don't know exactly what you are going to get, but you go out there and make a lot of connections of one sort or another. At any rate, those are the ones I wrote down.
I wrote down data analysis separately from data structure and data mining type things. It may be that thinking about the end user, the person who is actually going to analyze this, will tell us something about the structure of it in some sense.
I think privacy and confidentiality should be a piece of this, someone who thinks about this in the future and where that might be going, that should be a part of it.
I think as you think about the questions you might want asked, you might give some thought to who might give you provocative or interesting answers, not that any of the ones you are talking about would be the sole person to talk about X or Y or Z. They will probably make the other ones better by pushing them in their particular areas of expertise, I would think.
DR. AMARO: I would really like to stress that I think it is important, considering the limitations of the current data systems that we have, for a lot of the issues that we deal with in our subcommittee -- population-specific issues -- that we are thinking about the future and the increasing changes in population distributions and diversity, that we really try and get a group of people who will really be able to think broadly about that.
By that, I don't mean, just one African American, one Hispanic and one Asian person, but that everybody, or at least we try to get a group of people who really, regardless of what their own background is, thinks about the population shift.
In that vision of the future, we will be able to have that reflected and we won't end up with the same problems again.
DR. SONDIK: That says to me to try to get people who will think about how the data will actually be used. That I think is what you really put your finger on, when you are saying that.
I mean, what is it that we are going to need to know. What sorts of decisions are we going to be making, using this form in the future. I would almost put that under the devolvement, although that generally means decision making.
I see it as meaning we want more and more information on groups that are defined in more and more precise ways.
Change in the census may lead us to an even higher focus on that; not the census, but the collection of race and ethnicity information.
DR. AMARO: It is going to be a challenge to select these 15 or 30 people.
MR. BARTLETT: I was just going to say, if you are starting with two focus groups, you get a sense of how Noah said.
DR. SONDIK: I want to thank everybody for coming. I think this was not a start, but it is a great push forward in the middle of this process.
[Whereupon, at 3:30 p.m., the session was adjourned.]