[This Transcript is Unedited]
Room 425A
Hubert H. Humphrey Building
200 Independence Avenue, SW
Washington, DC 20201
DR. FRIEDMAN: So, let's meet. So, we are going to talk about where we are with the development of the final report for the 21st Century health statistics process. In front of you is -- do you have one, Barbara?
DR. STARFIELD: I have one --
[Multiple discussions.]
DR. FRIEDMAN: In front of you you have a prospectus that Rob Weinzimer and Ed Hunter and Gib Parrish and I have developed. Essentially, we are thinking in terms of a process that will take roughly a year. Although as I went over this time frame, Ed warned me that we shouldn't put anything down that had a time frame of more than a year and at the same time, when I looked at the time again this morning, I said oh, my gosh.
Essentially, what we are going to be doing is developing iteratively more detailed outlines of the final report over the next two, three, four months. We hope to have a first draft of the final report in July. The iterative outlines would be circulated among the workgroup; a first draft in July. This is the hope, the dream and the aspiration.
Obtain comments from the workgroup and the committee and NCHS. I don't know if it needs to go to the Data Council or not. And then the hope, the dream and the aspiration is that we would revise it at the second draft in August. Then at a period of two or three months were we send it out to a group of expert reviewers, people who have participated in the numerous hearings, local discussion groups and national groups, as well as obtain input from key organizations, some of who have already testified and then have the final draft ready next January.
One of the things -- and then we have an outline of the report. One of the things, which I will ask Ed to go over, but we have heard several themes come out very consistently in the hearings and in some ways the most recurrent, underlying theme that I think we have heard in all the hearings after the interim report was -- readers have been generally supportive of what we have done and at the same time, I think that there is a thirst and a demand for much greater specificity.
As Don Detmer said last June, the need for a road map, a designated driver and a budget. I think that at this point if we don't fulfill that need people will be very disappointed and I think we will not have done our job. So, the final report is really intended -- is to provide that.
Ed, do you want to go through the outline?
MR. HUNTER: Sure.
DR. FRIEDMAN: Everybody is wondering after that designated driver thing, does that imply that everybody else is planning on drinking and I wasn't sure what that meant.
DR. SONDIK: Can I make a suggestion? This thing makes more sense to me going backwards from the -- starting with the report outline and then going to the process and then going to the purpose because the purpose has all kinds of undefined words in it. Okay? Whereas, the report outline, I think, is really clear. Then the process, I think, makes sense in terms of the outline.
Just a suggestion.
MR. HUNTER: As far as I know, we don't have any particular audience for this document, other than this group and -- I guess the main point of this discussion is to go back and revisit the process thing, but also consider this the first iterative discussion of the outline.
I think the way in which we see proceeding is fleshing out in greater and greater detail the pieces of this until we wake up one day and you realize you are actually done if you just back space over the bullets and you have a paragraph.
We are at the point now I am just trying to organize the thoughts and then start filling in the blanks. So, it is totally open for comment.
In the first sections, the first is partly to deal with -- and many of these things are covered to some degree in the interim report and Rob has a -- sort of cuts and pastes the comments that we got from various hearings and sticks them into the internal report in the order in which they -- comments on the interim report.
For those of you that didn't see them, at the joint workgroup meeting several weeks or a month ago, there was a list of sort of common themes that came through the hearing. That is certainly useful background for this, too, when you sit down to actually really go through this.
The first part is really definitional, picking up on sort of really descriptive, sort of what is the -- what are we talking about. What do we mean by health statistics as opposed to other types of information, you know, the continuum of data statistics and the description of the organizations involved and trying to differentiate what we mean when we have this sort of known little thing when we say health statistics and we mean both the system and the product and sort of try to differentiate those.
And to talk a little bit about the elements of what we have -- what we sort of implicitly deal with in developing or using health statistics and talking about different aspects of the process of focusing a lot on content and what it is we are trying to develop statistics on but then what type of statistics do we mean, what kind of quality and characteristics, what are the different sources of data and trying to differentiate the system and the collection mechanism from either the purpose or later their use.
And then talking about how that is translated into a form that is accessible and usable and then finally the evaluation or the extent to which those are met. We talked about that as the cycle because the evaluation and forms serve as a repetitive -- that should help modify the data systems and one of the issues and recommendations later is how well those different elements of the cycles connect.
DR. FRIEDMAN: One of the things in this cycle of -- this development that we wanted to emphasize is moving away from the notion of health statistics as data collection. And that is why very purposely Ed in writing this put down identification of appropriate data sources rather than appropriate data collection.
MR. HUNTER: Then looking at it, it is more of a cycle. The issue is that you have identified a need, you identify a content need, you identify the kind of data you need to address that need and then you find where other data exists and you either collect it or you modify something that exists. And, hopefully, it gets us out of the discussion of the -- system equals health statistics equals use and the planning is done in a reverse mode rather than a forward mode.
We have in mind doing two different I guess templates for lack of a better word or schematics or something. One was one on a sort of model of determinants of health or -- the terminology here is used in different ways but we have sort of a map of health and the impact of different elements of the influence of this health on people that we would hope to lay out in order to help sort of visually portray the content area and then overlay on that at a later point what we actually have that bears on those content areas. That will then in turn help identify gaps.
DR. MAYS: Can you just give me an example? Because this is my first time hearing this through.
MR. HUNTER: Well, one of the leading experts on that thing is to my left. It is the Evans and Stoddard kind of model, with one example where you have -- trying to get away from the health care equals leads to health kind of thing, but different, you know, sort of -- portraying different influences from the environment, from individual behavior, characteristics, the health care system, sort of the policy and social background, all of which impact on individuals in different ways and all of which have interactions and relationships that you could pick up in a research point of view.
I think one of the things that helps you identify is that we do a lot better at measuring some specific elements of that model, like health care systems, you know, encounters and a lot worse of a job of looking at social determinants and environmental determinants and other things, that kind of -- carrying that to the next step kind of helped to identify, at least in the broader sense, the things that the health statistics system ought to be addressing and is or isn't doing very well at.
If you succeed all the way through this, you then have some kind of evaluation framework where you are always using that type of a model to evaluate how well you are doing at addressing content needs to find where that kind of a system --
DR. NEWACHECK: Ed, where does the model get described in this? Early on or --
MR. HUNTER: Yes. It is at the last bullet on the -- No. 1 is the development of these templates and we describe that. The second type of template, while we are there, is one that isn't about content. It is about the nature of the health statistics system itself. I am not sure how that actually lays out visually. We haven't even tried to do that, but there are kind of vague ruminations of things we could do to describe some things about the characteristics of data one might need to answer a content question and then taken together, those two things, you can use the second one to help identify places where the system is not strong.
For example, we don't feel at all that helping users with needs that are not experts in the field of statistics, actually identify, locate and get data in usable form, that sort of core group.
We don't do well at the feedback loop. That together with the template on content, I think, could be, you know, in some way that good diagnostics rule and in other ways a good way of describing places we think we need to focus recommendations in a later section.
DR. SONDIK: Would this have something in the front of this, someplace within the first few bullets that is a picture of interrelationships of various forces, institutions and so forth related to health. You know, I hesitate to say Evans and Stoddard, because that is one approach but since, you know, a lot of people know Evans and Stoddard, I mean, that is one thing that one could use.
I view that as sort of like, okay, that is, if you will, broadly speaking the health system. It is not just care. It is all aspects. So, what is health statistics? To me, health statistics is a probe to speak electronically. It goes in. You put it over here and you find out what is going on in the system over here. You put it over here and you find out what is going on over here. The probe may be on people. It is that kind of thing, but that shouldn't -- I mean, I would think that would go up front here, including some notion of the health information infrastructure.
But I think the model is really crucial to this; otherwise, it is focusing on the statistics model and that process.
MR. HUNTER: I think there are two very distinct models. One is we started with Evans and Stoddard and a lot of work that Barbara has done and sat around the table and tried -- how do you draw the picture of that health. I mean, it isn't strictly a research tool or an explanatory tool, but actually lays out the big block that you would want to describe as being important to understanding health. You know, we have boxes titled "Health Care," "Environment," "Genetic Biological Characteristics" and resources broken by community and personal ones; personal ones being behaviors and psychosocial aspects and type of policies.
That empty box up there, which I think is a stray box, but the idea being some simplified version of Evans and Stoddard that doesn't necessarily -- but it doesn't worry as much about drawing every relationship and every arrow, which I think is -- it is an application on the broad picture because you don't need to know going into wanting to develop health information which arrow leads to which thing because that is part of what the information helps you understand.
But laying that out in general terms, already you can look at it in the health statistics world. We don't have nearly as much on environment as we have on personal behaviors. We don't have much on genetic and biological characteristics and what we have on the health care system is probably not what we would describe as being important to helping understand health.
DR. SONDIK: I only bring this up because I don't see it in the outline.
MR. HUNTER: Development of template is to describe content. That is shorthand.
DR. SONDIK: That to me is very -- you know, sort of compulsive. I mean, I see templates, you know, and I immediately think of specifics on data collection; whereas, I think there is something more that is driving this. I think getting rid of the causality business, I think, is good because this should be a more general concept than that. Their causality, you know, if you want to look at it as theirs, you know, is one -- is a hypothesis.
[Multiple discussions.]
DR. FRIEDMAN: The one question, though, is when does the model get presented and if it is 15 pages into the report or something --
DR. SONDIK: I wrote down -- I think it should be right up very early in this, probably after some few words about the health information infrastructure that says, okay, we are honing in on --
MR. HUNTER: Well, clearly part of the introductory section. I mean, most of the rest of these things that -- I think we are talking about a 20 or 30 page portfolio. Most of the -- I don't know. It depends on how we want to write executive summaries. Most of these things -- these two templates are a page or two of general introductory stuff and then I think the point is to go back to those things and then use those frameworks in the following sections to -- as through the organizing framework for how we would describe current status, how we would describe staff, how we would describe -- enhancements and allow you to sort of put back into -- maybe even visually when you are done -- put back into places on the diagram the things that we are talking about doing.
It puts a lot of pressure on getting a good diagram, which, you know -- but that is --
[Multiple discussions.]
DR. FRIEDMAN: As Barbara and Ed know, I have been managing to procrastinate by reading every conceivable article ever written on the determinants of population. I walk around with a six page bibliography that I am constantly adding and subtracting from. Anything to prevent myself from sitting down again and specifying that, but I am going to do that in the next couple of weeks and then subject myself to the constructive but no doubt serious criticism that Barbara and Ed and Gib will provide me with.
MR. HUNTER: Well, we had a very energetic discussion a few months ago that got us this -- I think it probably still looked better in our minds and on the flip charts than it has yet on paper.
DR. FRIEDMAN: I do find this strangely overwhelming -- it just feels so --
MR. HUNTER: There is like 200 independent variables with -- but it is hard just to do it without judging what you think the relationships are because -- so, I mean, I think the first thing to do is take the errors out and that takes a lot of pressure off -- and then what you then worry about down the road is if you really want to look at relationships and you have the independent variables in one data system and then dependent variables in another data system, what do you do with that. That tells you you need to do something about linkage and --
DR. NEWACHECK: It also lets people read into it what they want to read into it without the arrows. That is, they can look at whatever piece they want to instead of --
DR. SONDIK: I think that is a great advance, taking out the arrows. Normally, I would -- it is hard to believe I would say that, but I think that in this context, I think that is very good.
[Multiple discussions.]
MR. HUNTER: The second part of this then -- why don't you take that one since it is in color and I will just have the --
[Multiple discussions.]
The second part -- the intent has been as I was saying to follow through on those templates and sort of use them as the evaluative framework. The second part would be to describe what we actually have and tell us what the -- sort of -- on the template as where we have strengths, where we have weaknesses and identify gaps both in content on that kind of a map and in system characteristics and data characteristics.
You could actually do multiple templates, one on, you know, the characteristics of data and we talked where we lacked data on geographic detail. It is not clear how many different templates make some sense here. So, we may have to play with that. But the detail and the quality of the data, the richness, the depth, is one of the elements of this second type of template and that would be how we would evaluate the current system.
Then probably coming from that methods and policy issues that cut across different elements of the current status. Many of those we identified as principles in the interim report.
The third area is actually starting to get to what we really want to do. In general, though, in the interim report, I think we got a lot of comment on and we have, I think, mostly reinforcement that we have identified appropriate principles and -- for us to finish writing Chapter 3 and get under way in Chapter 4, which is where we really get down to the road map.
We are going to try to fill in as much of this as we can and -- back and forth with you and others, but I think that is where the other parts of the process start to pick up as we want to get specific recommendations and next steps and sort of a road map, the level of details from some of these outside --
DR. FRIEDMAN: So, in terms of immediate activities, Ed is going to be taking a first crack at -- continuing the outline in greater detail for No. 1, I am going to be working on the determinants template and sourcing it and then Rob is going to be doing the first bullet under 3, which basically will be taking the existing principles and then taking a first crack at additional points that were made by the various people who have participated in our process.
DR. STARFIELD: By the way, you know, on the list of people that are working on this workgroup the three people that you mentioned on here. They actually should be on here.
MR. HUNTER: I sent Marjorie a note saying there were a couple who I had never seen.
MS. GREENBERG: It is a question as to whether they are working -- I mean, it is their call. If they are working as staff to the subcommittee --
[Multiple discussions.]
I have alerted Jim Scanlon, who is responsible for overall staffing of the committee. I don't know if they are down as staff to the subcommittee, though they really haven't participated in the subcommittee either.
DR. FRIEDMAN: So, basically, our plan is to try to develop a -- I would say first draft or very complete outline among, you know, Ed and myself and, you know, Gib and maybe some time with Rob and have everybody read it at various points. Then after we have either a very complete outline or a first draft, then sic Susan on it.
Our intent this time is to essentially do the writing ourselves and then have a really good editor/writer, like Susan, take a whack at it.
DR. LENGERICH: I am sort of new to this, but one thing that I think would be helpful also it there was some articulation of a vision for what -- is going to look like in the future. You have got principles and goals down here, but even -- what this is supposed to look like. I think the model is really helpful but it may be --
MR. HUNTER: Should be in that third bullet of the third item probably, although it doesn't actually say that because that is where we are actually saying -- I mean, that really should be exactly what you have said. It should be what the vision is because the next section goes on to how we actually --
DR. LENGERICH: But some kind of vignette of what it is going to look like.
MS. GREENBERG: I wondered if there were any vignettes in the NHII paper on the community part.
DR. STARFIELD: But, remember, we did away with those. We integrated --
MR. HUNTER: I think, I mean, whether they are there already, I think the idea of vignettes is really useful because it is not clear to a non-insider what it is we are talking about. We have had since conversations about the -- one are the missing pieces in the current system is that you don't have any particular place to go. That serves what for lack of a better term I would call it intelligent
-- you are a policy person or a researcher and you want answers. You don't necessarily have a data set or a data system at your disposal. What you really want is for there to be some transparent way to get the answers that you need.
It appears that the translator between the user and the people who actually manage the data system, whether it is an administrative byproduct or direct data collection -- and I mean, there are multiple parties that play that role in a more limited role. Anybody with a data collection system in some ways does that for their users, but it is not clear to a user that isn't familiar with that data set how you get to it or what you do.
So, trying to put in some vignette terms, you know, if you are a policy person and you have a question, what is it you do. I mean, how would this system that we are talking about facilitate you in a much more easy way getting an answer that you need?
DR. STARFIELD: We could go back to the original vignette because --
MS. GREENBERG: I think vignettes were effective and --
[Multiple discussions.]
DR. FRIEDMAN: John Lumpkin and I went to CDC to talk about NHAIMS and she was 90 at the time and -- and she said I hope the people at CDC will enjoy your hallucinations.
DR. NEWACHECK: [Comment off microphone.]
DR. FRIEDMAN: Ed and Gib and I have a contract, although I am not sure what the meaning of the word is since there is no money that exchanges -- you know, clings to us as government employees -- with Oxford and we have started soliciting authors. Actually, Barbara has agreed to write a chapter. Don Detmer has agreed to write a chapter.
Lisa Iezzoni and Mike Schwartz are going to be doing one on administrative data. I feel very positively about it. It has changed quite a bit since the original conception. The original conception was more as a collection of papers with a common theme. As it went through the review process -- yes, I mean, it basically --
PARTICIPANT: Are they still in it or --
DR. FRIEDMAN: Some of them are. Most of them are not. -- more and more consistently with, folks wanted something that could be used as a text with more structure and -- integrated, not a collection of papers.
PARTICIPANT: Will this be part of it in a way or a synopsis of this or --
DR. FRIEDMAN: There are two places probably. One is, I think, on the second issue side, there is a definitional chapter in the book as editors stand, myself -- and I think it draws some of the same thinking that we picked up through the vision process. At the end, there is the chapter that is really explicitly designed to reflect the -- that we come up with here and the book will come out after the -- it will be different in the book chapter, but I think it should be very --
MR. HUNTER: And then assuming that Sherman Black's paper, Sherman et al. paper and the one that was done by Jennifer Zelmer, the Canadian health information, would be there perhaps in some iteration.
MS. GREENBERG: But not the other papers?
DR. FRIEDMAN: What we have been discussing was the possibility of getting a third author who would do something that would integrate some of what the LaClinic(?) did and some of what Lorraine did into a more general chapter on health statistics policy, rather than having taken smaller case studies, reducing what they had to say, if that makes sense.
MR. HUNTER: -- oddly as case studies without -- they weren't representatives. They were more in depth.
DR. NEWACHECK: What about like a chapter on modeling?
DR. FRIEDMAN: There is going to be -- we will have models incorporated into at least one of the chapters and then hopefully we will be returning to it and then Barbara is going to be doing a chapter not so much on
models --
DR. STARFIELD: There will be models, but I think more important -- this is really complementary to the series of papers that are going to be in the HAPH that Dan and I are putting together. That will be specifically on that model, you know, starting with the Evans and Stoddard model ten years later, which is --
DR. FRIEDMAN: This is a separate enterprise that is going to be a HAPH special focus section on -- that will have -- Evans and Stoddard are going to be an article that is a ten year reconsideration. Then there will be an article by David Coburn and colleagues at the University of Toronto that has been critical of Evans and Stoddard. Then there would be one by Ilona Kickbusch(?) APAL(?), who was the driving force between -- trying to integrate 2010 health for all and population health and then there will be an historical article looking at models of population health and historical perspectives by Simon Schweider(?).
[Multiple discussions.]
Absolutely, yes. And John Milliner(?) and -- health policy networking kind of a -- on the impact of models of population health, maybe on health policy. A very interesting book, almost a 30 year history now on -- really how to use models, starting with Milan(?) and really before Milan, have really had an impact at the national and at the provincial level in Canada. It is really fascinating.
DR. STARFIELD: Dan and I will write our summary chapter, which I think probably will pull the model stuff together for that activity, not for the other one.
MS. GREENBERG: When is that supposed to be?
[Multiple discussions.]
DR. STARFIELD: No. Actually everybody agreed first shot -- we spent a long time figuring out to ask but everybody --
DR. SONDIK: To get back to this, I think the -- I would suggest that 3 not start out with the enhancements, but that it be more of what the goal is. What should this system be. It is kind of what we were saying was the vision and then enhancements, I think, is a subpoint under this, which is, you know, how could we change the current system here and there to achieve the goal.
I think it is very important that in the development process up above that in that review of the first draft that the -- that is No. 2, second bullet, that the three organizations be clear that they sign off on this, not as -- but that they say this is us. This report is coming from three organizations because it is going to contain more than we heard in the various meetings that we went to, that we have convened.
So, I wouldn't want to see it criticized on the basis of, you know, no one ever said that kind of stuff,but that it is the three organizations working together that is producing this report. I want to be able to use the report as this is the best thinking that has gone on about what health statistics should be and then use that as the framework, a lever for getting the resources to be able to build the system.
DR. FRIEDMAN: Ed, you raised a point a few months ago about how -- and subsequently we have had some discussions on about how far you thought the -- process could work and whether or not there might be a place in the report at which NCVHS might need to not quite go off on its own, but that in terms of more specific recommendations.
DR. SONDIK: Yes, I remember raising it, but I am using it in the specific context in which I did, I think it was to some degree NCHS and DHHS, the Data Council, are sort of -- you could argue that the disinterested observer, I guess, could say -- unbiased observer could argue that HHS is concerned about feathering its own nest with respect to this, okay. It has got a bias. It is NCVHS that really has no bias.
I mean, you know, this organization was called together to advise the Department. So, in that sense, its input is the most pure. My input is not the most pure for a variety of reasons.
DR. NEWACHECK: We are all employees of the Department.
DR. SONDIK: I have actually thought of that, but that is not a big deal because it is in an advisory basis. It is not an issue
PARTICIPANT: We don't get paid much anyway.
PARTICIPANT: Don't get paid anything.
[Multiple discussions.]
MR. HUNTER: That was my point. There is a secondary point, which is there may be -- any Department organization -- that we touch on that is becoming -- potentially delay getting something out because -- on the other hand, we don't know what those things are yet. So, we don't need to judge at what point we don't bother to go for clearance and at what point we think the controversy is.
DR. SONDIK: It is my experience that the reports that come from the purist organizations, so to speak, have their biggest impact, particularly if they are proposing some radical -- if it is within the Department, it is just
-- you know, it gets less -- it has less impact. That was the point.
DR. FRIEDMAN: Let me be specific. What we have discussed as a possibility and I completely agree with what Ed Hunter was saying. We don't need to cross that bridge until we come to it -- is that essentially what is now Nos. 1, 2 and 3 in the report outline, could be the joint product and then No. 4, while it would certainly be developed -- hopefully be essentially a seamless part of the report, could be the part that if NCVHS needed to, could be purely an NCVHS document.
DR. SONDIK: I agree with that. I think that part of 3, though, if it is as far reaching as I hope it is, mainly would come from NCVHS. The stuff that is a little more visionary, it kind of depends on how this fleshes out. We might get this tremendous buy-in from everybody at the Department.
MR. HUNTER: My only real concern is how to figure out how to get that buy-in with the administration changing.
DR. SONDIK: We did have -- and I don't go to every meeting, but we did have one proposal that came from Phil Lee, which was for a major shake up in the way the Department does business with respect to this. That would be ideal for that or however, you know, it is viewed by all these multiple organizations that this is going to go to. That process is non-trivial.
DR. FRIEDMAN: He, his son and his son's colleague wrote five pages. It was like a hundred pages.
[Multiple discussions.]
MR. HUNTER: Following up on that, I am just wondering if we would --
DR. FRIEDMAN: This has been a continuing source of discussion.
We would really appreciate any comments, really seriously. So, please send them along.
DR. SONDIK: You are figuring doing this in a year?
MR. HUNTER: Yes.
DR. SONDIK: I think it is entirely doable. We have got most of what is needed here, other than empiric, which, of course, is the critical piece of this, but I would think a few sessions getting some people together and -- some stuff can be hammered out.
[[Whereupon, at 5:40 p.m., the workgroup meeting was concluded.]