[This Transcript is Unedited]
Hubert H. Humphrey Building
Room 705A
200 Independence Avenue, S.W.
Washington, D.C. 20201
Proceedings By:
CASET Associates, Ltd.
10201 Lee Highway, Suite 160
Fairfax, Virginia 22030
(703) 352-0091
Committee Members:
Liaison Representatives:
Call to Order - Dr. John Lumpkin
Panel Presentation on Summary Health Measures - Dr. Edward J. Sondik, Moderator
Agenda Item: 50th Anniversary Symposium
Reports from Subcommittees and Workgroups
DR. LUMPKIN: Let's get started for today's session. We will start off by doing introductions around the room. I would just like to remind everyone that we are live on the Internet. So, when you speak, please speak into the microphone so that all those out in cyberland will be able to hear us clearly.
My name is John Lumpkin and I am director of the Illinois Department of Public Health and also the chairperson of the committee.
We will start off by going to my right.
MS. GREENBERG: I am Marjorie Greenberg, from National Center for Health Statistics.
DR. LUMPKIN: See, now, I start out by saying please speak into the microphone and --
MS. GREENBERG: Sorry. I am Marjorie Greenberg, National Center for Health Statistics, CDC. I am the executive secretary of the committee.
MS. COLTIN: I am Kathy Coltin from Harvard Pilgrim Health Care in Boston, a member of the committee.
DR. FRIEDMAN: Dan Friedman, Massachusetts Department of Public Health, member of the committee.
DR. NEWACHECK: Paul Newacheck with the University of California at San Francisco and a member of the committee.
DR. ZUBELDIA: Kepa Zubeldia with Envoy Corporation, a member of the committee.
DR. IEZZONI: Lisa Iezzoni, Harvard Medical School, committee member.
DR. SONDIK: Ed Sondik, director of the National Center for Health Statistics.
DR. HAYWARD: Mark Hayward. I am a panelist, at Penn State University.
MS. FRIEDMAN: Carol Friedman. I am also a panelist with the Epidemiology Program Office, the Center for Disease Control.
DR. MC KENNA: I am Matt McKenna. I am a panelist and I am from CDC Atlanta, as well, the National Center for Chronic Disease Prevention and Health Promotion.
DR. WAGENER: I am Diane Wagener. I am on the panel and I am from the National Center for Health Statistics, which is part of the CDC.
DR. WOLFSON: I am Michael Wolfson. I am on the panel and I am from Statistics Canada.
MR. BLAIR: I am Jeff Blair, member of the committee, vice president of the Medical Records Institute.
MS. FYFFE: Kathleen Fyffe. I am a member of the committee and I work for the Health Insurance Association of America.
MR. GELLMAN: I am Bob Gellman. I am a privacy and information policy consultant in Washington and a member of the committee.
DR. COHN: I am Simon Cohn from Kaiser Permanente and a member of the committee.
DR. STARFIELD: Barbara Starfield from Johns Hopkins University and a member of the committee.
MR. SCANLON: Jim Scanlon from the Department of Health and Human Services. I am the staff director for the committee.
DR. LUMPKIN: And then we will have introductions of the attendees, but we don't have microphones.
[Further introductions of the audience off microphone.]
DR. LUMPKIN: We are going to start off today. We have certainly a panel that we are looking forward to the presentation. I just need to make a brief apology. I have to do a radio interview with a radio station back in Illinois at noon, about the governor's budget. So, I will be slipping out slightly before the panel is finished.
I just wanted to give you my apologies. It is not anything that you say and given the nature of this interview, I would much rather stay here than having to go and do that.
Ed, I guess you --
DR. SONDIK: You are on record now as having said that.
DR. LUMPKIN: Over the Internet. I will probably start out the interview by saying that.
DR. SONDIK: You could have been talking about summary health measures instead of summary budgets. Summary health budgets.
Okay. Thank you very much. I want to thank the panelists for being here today. This is a -- I was asked to moderate and I will try to be moderate. We have asked each of the panelists to speak for up to 15 minutes, which would then allow us about 45 minutes for discussion.
I have had an interest in summary health measures for many, many years that goes back to when I studied decision analysis and how one values various states of health. At that point, it was more of -- I think, in the general field, it was more of one of a -- there was an intellectual interest in this topic with the idea being that perhaps someday there would be application for it.
Over the past 20, even 30 years, in fact, it has developed. It has developed to the point where now there is very great interest, particularly international interest. WHO has a strong interest in this and there is interest in how one develops these measures and, in fact, how to apply them. There is an excellent report from the Institute of Medicine of medicine that now goes back, I believe, two or is it three years, that summarizes a number of issues related to the development and the application of these measures and in general recommends a strong research program to help evaluate the candidate measures and how they can be used in a variety of different applications.
I think it is extremely important that the committee, the National Committee on Vital and Health Statistics, be abreast of the activities in this field. I say that because the development of our information resources clearly leads into and supports these measures and vice-versa. Having these measures clearly will influence the direction of our data collection.
Obviously, this is a critical point for the committee. The area, I think, and I think without having talked to our individual panelists ahead of time, I think you are going to get a sense of how dynamic this field is at the moment and the very large amount of activity.
I think it is important that the committee keep in mind this activity and the directions, multiple directions it is moving in, as you all consider the information, the health information resources for this country.
Well, enough of that as an overview. We have got five panelists here. We are going to cover a number of different topics related to this from an international perspective to differing perspectives in how these measures are developing and being applied within the United States.
It is a distinguished set of panelists. I am very pleased they have taken time to be here today. I am going to start with Dr. Wolfson and since there are people watching over cyberspace, let me just give you the order. I should say we are playing from left to right as they view the playing field.
Dr. Wolfson is going to begin and then Dr. McKenna, Dr. Friedman, Dr. Wagener and batting clean up will be Dr. Hayward.
Let me begin with our colleague from Stat Canada, who I am very pleased is here today. I should point out that he recently on this and related topics to the Data Council of the Department. This was one element in his discussion. I think it is important, particularly important for the U.S. that we collaborate and coordinate our activities with those in Canada because the comparison and contrast across our differing health care systems, I think, is extremely useful for us.
As we try to make these comparisons and contrasts, really there is nothing, I think, more important than having appropriate measures that we can use.
With that, let me ask Dr. Wolfson to begin.
DR. WOLFSON: Thank you.
Well, I am really pleased to have the opportunity to speak to you and I am going to try and speak and turn slides and not get in the way all at once. I don't know how easy that is going to be. You have had circulated to you a paper and in the short time available I am going to be rather haphazard in highlighting a few points about that.
By way of background, long background, over a decade ago, we started to have a serious look at the way our health statistics program was working in Canada and in Statistics Canada in particular. The general diagnosis was not great and one of the two key gaps is we didn't know enough about how healthy people were, what the outcomes of the health system were. We knew about what was going in and through relatively speaking.
So, that is part of the genesis of the work that we have been doing on summary measures of population health. The paper itself goes through a number of topics, basic concepts. It gives some illustrations, articulates a vision, raises some questions about causality and the empirical challenges.
I am going to touch on all of these, maybe, but only briefly.
To begin, we are all very used to a number --
DR. SONDIK: Would you like someone to turn the slides for you? Would that be easier?
DR. WOLFSON: As long as they can know when I say "skip" because I am not going to go through all of them, but sure.
The one summary measure that everybody is used to is life expectancy, but there is a different kind of summary measure that a lot of folks are used to, which is when you have done something like the National Health Interview Survey and you cross tabulate it, you get a number like the percentage of people who have X or don't have Y.
So that the crucial thing in that kind of summary measure of health status for a population is that there are two steps. There is a step of describing individual by individual, using some sort of classification system or questions. What is going on in that particular individual's life at that point. Then there is an aggregation process.
But all of the summary measures that -- well, the summary measures, like Dalley's(?) and others that I want to talk about actually involve three steps, rather than the two conventional ones. One is describing an individual's health state by some sort of a profile, aggregating what would be a set of classifications into a number between 0 and 1, which I will call micro-level aggregation because I want to distinguish it from the third step, which we can call macro-level aggregation, rolling up the results on health status for a bunch of individuals into some overall number.
It is breaking it into three rather than two steps, I think, that has been a major innovation over the last number of years with health-adjusted life expectancy and things like that. Let me emphasize that it is very useful to keep those three steps in mind as distinct and what one does for step 1 has some but no necessary relationship and there are options at each stage of the game here.
To give you a quick flavor, here are two micro-level descriptive systems, one that we have been using in our health interview surveys in Canada, called the Health Utility Index for McMaster University and others, that has been developed by a European group called EuroQual(?) and you can see that there are a number of dimensions for each system. In the case of the McMaster one, vision, hearing, speech, cognition, dexterity, mobility, pain and emotional status. In the case of the EuroQual, there are five -- there is a big question about how one ought to do this descriptive system, but just to show by way of illustration.
There are other descriptive systems. The most popular one by far is SF36. The main problem with SF36 is it was not designed to be rolled up into one of these 01 indices. But that is a crucial step if we want to make a summary.
This diagram isn't in the paper but I thought I would throw it in for fun. There is an important question about how one conceptualizes how one thinks about which domains are in or out of the descriptive system. There is a large literature on how people think of health. There is no question that it is very broad. If you ask lay people in various kinds of surveys, as Mildred Blackster(?) and others have done, but for purposes of the kinds of measures we are talking about, it seems that a useful distinction is intrinsic versus extrinsic, what is within the skin and what is not or without the skin, dare I say.
One of the problems we have in the various descriptive systems that are being developed is not all of them are clear in this kind of a dimension. And I would be happy to entertain a discussion if you want to get into that, but that is just a quick idea.
Let's skip that one.
For the descriptive system, the micro-level descriptive systems, it is, obviously, very useful if it has a conceptual framework. If it has some sort of a hierarchical classification system -- and you saw two simple examples with the EuroQual and the McMaster one -- there are more detailed elements. This is about -- the reason I put the phrase "short form standing" -- what does all that mean -- well, I had the privilege of attending yesterday afternoon a discussion in the Population Subgroup about ICIDH, which has hundreds and hundreds and hundreds of items and it would really be nice if the ICIDH could form an international foundation for the classification of health.
Let me open a little parentheses. We in Canada have no problem developing things, but descriptive systems -- in fact, we have charged ahead and put them on our health surveys, for example, but we would be -- it would be far more valuable to us and I think for the U.S. and other countries as well is if we could agree on a way of describing health in various countries so that we could make international comparisons so we could learn by what are the differences that make a difference, a friend of mine refers to.
A concern about the ICIDH is it doesn't have a natural short form version. It is this big long list but there is no -- it hasn't been designed in a way at present to have that. Another obvious concern or question or desideratum is that whatever classification structure you have ought to be operationalizable, operationalized, by an accompanying set of questions or ways of eliciting the information that everybody agrees on how to do it.
This picture is in the paper but it is black and white and it is hard to read. Let me just briefly tough on the bottom part. The other papers make the similar point, but the kind of index that we are talking about is saying let's go beyond the area under the survival curve, which is this envelope. The area under that survival curve is life expectancy. If you arbitrarily made a cut point and said anybody who has got less than a certain level of functioning is disabled, you could construct another index called disability free life expectancies. That is kind of dichotomous.
The idea here in that 01 business is to say if we can assign to each person year of life a number between 0 and 1 -- and here we have a metaphor -- and make the gradiate level correspond to that, then what we are talking about is the weighted area under the curve, the area that gives a 1 if it is white and .5, if it is halfway between black and white, whatever. That his health-adjusted life expectancy or functional equivalent life expectancy, something like that.
That is the core idea of these summary measures. This list is not unlike the previous one but I have added one more point. We need a standardized micro-level descriptive system if we are going to have one of these summary measures, like the McMaster index or something else. We need a consensus on how to aggregate that. I am not going to talk at all about it unless you want to ask questions, but there is a lot of controversy about how to map a description, a profile, an answer at five or six levels to each of eight questions into a number between 0 and 1.
We have the weighted life expectancy aggregation, but we can't stop there because the policy process wants to be able to ask questions like so what and what if we did X. So that we need to be able to do causal attribution. We need somehow to associate -- be able to associate, be relevant. In the case of Dalleys, this is done in a straightforward but in my view somewhat limiting way by building up from diseases. But we can come back to that again if you want because diseases aren't necessarily the most natural way to think about health problems.
One of the core ingredients, to go into a bit more depth for a second on this one, is this idea of how we measure health status and you saw the EuroQual and the McMaster indices a couple slides ago very fleetingly. That is a micro-level descriptive system and all that. We are putting it on our health survey, have been for some years.
But it is not enough. We want to be able to not only roll up the summary index, but be able to have a rich structured system of statistics underneath so that as people ask questions, which they inevitably will do, you can drill down and unpeel the onion. Say which group has how much of this or that. It will be important in a minute when I get -- later on, I will jump ahead, but let me just mention -- focus on the last couple of bullets.
It is one thing to put this on a survey. There is a discussion that a subcommittee of your committee is having about whether to put these kinds of questions routinely on discharge abstracts or health care encounter data, which I think is very important. But let me observe that right now when people do RCTs, they measure death as an endpoint. That is the only common endpoint. They may use some sort of health status or morbidity measures, but there is no standardization in that.
So, as a result, when one comes to do meta-analysis, one is extremely constrained in the kinds of insights or the wealth of information that one can pull together because we don't have a common story or way of assessing morbidity across these studies. Yet, it is exactly that part that we are driving toward. Again, the basic message here is death is all well and good as an indicator of health, but it is not nearly enough. Especially as we live long and we have chronic diseases, we need to do a much more systematic and careful job about capturing morbidity.
So, that means when you do RCTs, it would be nice if there was a standardized way of describing people's health status that we could put on those things, as well as on health encounters. Linking this to policy, I assume everybody in the room appreciates that the causal stories behind why a person is sick at a particular time or has a particular health problem are usually not simple. They are complex, multifactorial, often involve long lags. This is a real challenge to incorporate this, to include this in our thinking. And not all causal factors are policy levers. You can't just turn off cancer. You can do something about smoking as a trivial example.
So that one of the things we have been doing in Canada, which I won't go into anymore depth -- there is an example in the paper -- is imbedding this in a model. Ministries of finance, your Congressional Budget Office and Treasure have for 30 years now had detailed models so when they do tax changes, they can estimate not only the aggregate revenue impact, but the distributional impact. It is long past time that health ministries or health departments have the equivalent kind of analytical capacity. It is a challenge to build but it can be done.
I said there are a number of conceptual and ethical challenges here. Just to quickly mention a few of them that have to do with the first of those stages of the descriptive system, how do we capture the idea of the WHO, that health is more than the mere absence of disease, a non-trivial question. Social role functions versus intrinsic or within the skin and the usual concerns about psychometric properties.
Let me emphasize again that certainly from our point of view in Canada, the tremendous benefits that would be had if we had internationally comparable ways of doing this.
since the picture came out so badly in the photocopy I got, here is the color version so you can see the two bars. This shows, using the oldest method of constructing these kinds of statistics, the Sullivan method, what happens when we compute not only life expectancy but health-adjusted life expectancy in the Province of Manitoba. In this case we were able to do it because we were able exactly to match census data for a sample of the population to their health care and mortality records.
I think the NLMS here would allow you to do the similar statistics. Whether we are talking about education or income, there is clearly a difference in life expectancy by socioeconomic status and, if anything, the gradient looks a little steeper when you take account of morbidity not just mortality.
At the IOM meeting a couple of years ago, which was referred to, there were a couple of philosophers, who raised very significant concerns about the third of the three steps that I mentioned, this macro-level aggregation. Basically what that -- if you think about taking the weighted area under that curve with the gradiated levels, we are giving one person year one vote and there is no question that in any real policy discussion or resource allocation discussion, some people would gain more than others by a particular change in policy and resource allocation, whatever.
So, distributional concerns will come to the fore. They raise the question, well, we are not sure we are comfortable with the kinds of distributional implications that are embodied in or inherent or even occult in these summary measures. My reaction is to say that is fine at one level, but I think it is really -- there is a different way of getting around that. Instead of having a big debate about what are the ethical or distributional assumptions underlying the measure, is instead to think the way the Treasury does and says, okay, we are going to get the bottom line number about what the revenue impact of this tax change is, but we will also produce gobs of tables that show who wins and who loses by the change.
If the summary measure is imbedded in a statistical system that is designed to support this kind of distributional analysis, then just like tax policy and welfare policy and other kinds of policy changes that are routine in this town, one could imagine and one could plan on doing the same kinds of distributional analysis.
So, in a way, I am saying I don't think those concerns are nearly as severe as the philosophers have been saying.
So, by way of a few concluding comments, my view is there is a pressing need for us to work together to develop a summary measure of some sort, a GDP or a CPI of health if one wants, but this is not an area where there is a right answer that can be discovered if you just spend enough time and get enough researchers to work at it.
It will in the end require some process of judgment and consensus. Let me observe, however, and, again, perhaps it is a fault of my training as an economist and working in the statistical office next to the people who give you the national accounts in the CPI, exactly the same kinds of problems exist there. The ILO, many years ago, got together and got all the countries together and said, well, there is no right way to define unemployment. Here is the way we want you all to do it so we have comparable measures. Done.
Even earlier, people got together and said here is what we want to measure how the wheels of the economy are turning. It is called gross domestic product. Boom. There is an 800 page manual from the United Nations and others on how to do it. It is done. And it is widely accepted. Perhaps it is not questioned as much as it should be. But it is done.
I think that we should not let the perfection or the excellent be the enemy of the good here.
Thanks.
DR. SONDIK: Right on time. Thank you very much.
I think I neglected to mention that Dr. Wolfson is director-general of analysis and development for Stat Canada.
Our next speaker is Matt McKenna, who is chief of the Epidemiology and Health Services Research Branch, the National Center for Chronic Disease Prevention of the CDC.
Matt.
DR. MC KENNA: Thank you. Good morning.
I am here this morning to briefly discuss the United States Burden of Disease and Injury Project, which is a cooperative effort between our center at CDC in Atlanta and the Burden Disease Unit at the Harvard School of Public Health.
This project was initiated because it became increasingly clear at CDC that mortality or body count as a preeminent form of information for guiding public health priorities was inadequate. As Dr. Wolfson has alluded to, in a low mortality country like the United States, focusing on mortality information misses a huge chunk of the suffering associated with ill health.
From our perspective, being focused on health promotion and prevention, we felt that mortality promoted a high tech curative and rescue-oriented focus on health care and health policy, a perspective we wanted to try to reorient.
Another goal for the project was to produce a wide range of internally consistent, epidemiologic estimates for fatal, as well as non-fatal, health outcomes. We wanted to use the best data available to assure scientifically credible numbers because to be honest some of you may have experienced this, there seems to be an inevitable tendency for individual programs at CDC and perhaps some other health agencies, to attempt to maximize the burden of their respect conditions.
The burden of disease approach that is used by the World Health Organization requires a rigorous assessment of individual disease epidemiology that systematically discourages this type of inflation.
Finally, given the attention and influence of the global burden of disease results, it was felt that it would be useful to place the U.S. public health situation in the global context. This would highlight the commonalities of the U.S. with the international activities and priorities.
As many of you are aware, the index used in the global burden disease study is the disability adjusted life year. Now, this index is being updated regularly by the World Health Organization included in their annual report and there has been a lot written about it and how -- and I really don't have time to get into how this is better or inferior than other measures, such as quality adjusted life years, health adjusted life years, years of healthy life, disability free life years and others.
To be honest, I don't want to get into it because frequently the discussion seems to degenerate into a heated debate that seems to border more on the theological rather than the epidemiological. All I will say is that the fundamental concept underlying Dalley's holds that healthy live time can be lost in two ways. First, because of premature mortality and, second, because suffering, pain or loss of physical function.
Now, please note in this framework these general concepts of ill health, pain, suffering, physical dysfunction are what I mean by disability, not long term inability to work or other meanings that are frequently attached to this word.
Since premature mortality, known as years of life lost in the Dalley construct, is the first and most straight forward dimension, we developed a detailed assessment and mortality patterns in the United States over a 30 year period.
This revealed substantial differentials in life expectancies by county throughout the nation, as illustrated by this map. Here we are displaying the county's specific life expectancies at birth for males for the collapsed period 1985 through 1994. Now, this pattern is very consistent with almost all atlases that attempts to illustrate geographic patterns of health in the U.S. The upper Midwest, Utah and New England or the purple or dark areas here exhibit high life expectancies.
Low values are found in the Appalachian region, the Mississippi Delta, the I-95 corridor of the Southeast United States and some isolated areas in the Midwest, where there are large numbers of Native Americans on reservations, such as the Pine Ridge Reservation in South Dakota.
Not seen on this map are some areas that are small geographically and difficult to pick up, such as Baltimore City, Washington, D.C., Bronx and San Francisco.
Though the pattern of mortality was not expected, there were surprises from this analysis in terms of the absolute levels of life expectancy calculated. There were also astonishingly large disparities in these indices. For example, when the sex and county level life expectancies at birth were further stratified by race, we found values below 60 years for African American men in cities, such as Washington, D.C. and San Francisco.
To put these numbers in the context of the international picture, these are comparable to countries, such as Guatemala, India and Egypt. However, the infant mortality rate in these countries are about 5 to 7 times greater than those seen in countries with similar -- in the countries with similar life expectancies are 5 to 7 times greater than we are seeing in these counties.
In sharp contrast, the values for Asian women substantially exceeded 90 years in several counties, such as Bergen, New Jersey and Dade County, Florida. Now, this 30 to 45 year differential is about equivalent to the difference between Sierra Leone and Japan. However, there are occurring within the border of a single country and they are unparalleled in any other similar type of analysis that has been done in any other developed nation in the world.
Now, as mentioned earlier, the low life expectancy counties had infant mortality rates that were higher for the United States but not high enough to explain low life expectancy values we observed. To explore in more detail the major sources of premature mortality, we grouped the race, sex and county-specific life expectancies into eight categories and this slide would explain the years of life lost in values from the group constituting the 2.5 percent of the population with the lowest life expectancy as compared to the 2.5 percent of the population with the highest life expectancies.
Now, though there is a bit of a spike at the very young ages, as I mentioned before, this doesn't explain the differentials we saw in the life expectancies. In addition, when we grouped the causes of death into three main categories; namely, injury, which is orange in this graph, communicable or infectious perinatal causes, red and non-communicable or chronic diseases, which is white, the largest area under this bar appears to reside with chronic diseases, such as cancer and heart disease.
The largest amount of life years lost, the injury were observed not surprisingly in the adolescent and young adult life years and communicable disease was most prominent in young adulthood and almost all attributable to HIV.
The next phase in the analysis is to calculate non-fatal health outcome. Now, this requires a large range of epidemiologic estimations, as Dr. Wolfson alluded to earlier. We are building up from individual diseases. We have to come up incidence, prevalence, average duration of disease and these calculations are informed and constrained by information on cause specific and general mortality using a software program called DISMOD or disease modeling, which I don't have time to talk about, but it is available for downloading from the Harvard Web site.
We are performing these calculations in consultation with a wide range of experts from CDC, NIH, as well as voluntary organizations, such as the American Heart Association.
Now, the other link, the sort of third step here between combining the time lost to non-fatal health outcomes with the years lost to premature mortality or a set of what is called disability weights. These weights are valued between 0 for perfect health and 1, which is equivalent to death.
Time labeled specific conditions are then weighted by this to calculate the years lost due to disability. Now, in this graph we are showing that there is substantial agreement between the condition specific weights derived during a two-day exercise we conducted at CDC and those that have been found during a variety of such exercises conducted by the Harvard group around the world.
This made us reasonably comfortable. Now, there are a few exceptions because of dynamic changes, but that the weights in the global burden disease had general applicability to the United States. One set of diseases, for which we have very good descriptive epidemiologic data are cancers and because of this and the fact that I actually work in the Cancer Division of CDC, our estimates in this arena are further along than for any other non-fatal and fatal health outcome.
As you can see here, the Dalley's for cancer are dominated by years of life lost or the yellow parts of the bars. This is reflecting premature mortality. However, there is one exception to this top ten list, which is skin cancer. Now, I think the most useful dimension in this graph is not the exact ranking of these conditions, but that it does allow a reflection of the volume over one million cases of non-melanoma skin cancer that result in 500 to a thousand deaths every year in the U.S. and then when combined with melanoma highlights a preventable burden that is frequently ignored because few deaths are attributed to these conditions.
Now, based on the data from the global burden disease analysis and the work that we have done specifically in the United States so far, we have also developed a preliminary list, what I call on this slide, evolving list, conditions that are major sources of disability and premature mortality in the U.S. Two that occur in this top ten list that would never appear in any construct using mortality alone are osteoarthritis and depression.
During the execution of this project, the CDC and Harvard have received many requests for assistance in implementing this methodology from states, counties and even one Dow Jones Industrial Index company, General Motors, which my colleague, Dr. Carol Friedman, will tell you more about.
I have included the report from Los Angeles County in the packet of materials that you have and that was completed recently in cooperation with RAND and UCLA, as well as the L.A. County Health Department. We believe this interest on the part of a diverse variety of organizations reflects the need and demand for this type of approach to assessing health in populations.
This graph from an article that appeared in The New England Journal of Medicine a little over a year ago is further evidence of the interest in this type of measure. This graph used the Dalley measures from developed countries in the global burden disease study to assess the relationship between NIH research funding for conditions and the population burden of these diseases. It is my understanding that you are sort of familiar with this whole activity.
Now, in the articles the authors pointedly noted that specific estimates for the U.S. did not exist at the time this article was written. The National Institutes of Health -- and I see Jim Shutting(?) represented in the back -- subsequently held a conference to address the concordance between burden and their funding priorities. I am not proposing and we don't definitely think that burden should be the sole determinant of research priorities.
I show this to emphasize that there appears to be a great demand for summary data about individual conditions at the population level for a variety of purposes.
As I finish up, I would like to reiterate a few points and identify a few challenges in implementing and using disability adjusted to life years or other summary measures.
As I have said, the great value of these approaches is that they allow us to combine mortality and disability and, hence, broaden the scope of health conditions that receive attention in policy discussions. I think the epidemiologic approach used by the Dalley method also assists in getting us beyond the frequently used measures of burden, such as facility visits or dollars spent on condition because these are very sensitive to reimbursement systems.
Hence, Dalleys really try to identify what the major sources of burden to health are independently and what the health care system is going to pay for. Finally, this approach assists in helping data gap -- identify data gaps for diseases, which are inadequately addressed in current data collection efforts and what components of disease about which we do collect information are not properly addressed.
Finally, I think it really does encourage us, which is something academic epidemiologists don't like to do, to fully use the wealth of descriptive epidemiologic information that is already available and is honestly infrequently mined for, I think, the informational goal it contains.
Really finally, there are clearly a large number of challenges to executing this type of assessment and establishing this part of regular health information system. However, four major points I would like to mention is that though there are heated arguments surrounding the various dimensions of such analysis. In my experience, there is tremendous amount of robustness. In general, people really do think that being in intractable pain is worse than having facial acne. In large volume, medical problems do get reflected as having a large burden in the final results.
Unfortunately, there is a tremendous tendency to emphasize differences in critiques rather than constructive contributions. I would also like to emphasize that there is really no right way in all of this -- this is reiterating something Dr. Wolfson said -- these are perspectives and uses and I think indexes should be constructed in terms of how they are planned to be used, rather than claiming one approach is inherently superior or tapping in to God's vision of the way things are.
Also, I have emphasized that there is an enormous demand for these emphases. When they are produced, they are used. Therefore, it seems incumbent on CDC and groups such as this one to actively engage in contributing to them to assure quality and rigor in their construction.
Finally, we have had an agency-wide work group at CDC that has produced a recommendation to Dr. Coplan(?) that CDC actively explore the use of Dalleys and other summary measures. How this measure will be implemented, if at all, is still being deliberated.
I appreciate your attention and if we have a question period, I will be glad to participate.
DR. SONDIK: Thank you very much.
I would like to turn now to Dr. Carol Friedman, who is a medical epidemiologist from the Epidemiology Program Office at CDC, who is going to talk about that application, in fact, to General Motors. As they used to say, what is good for General Motors is good for the -- they don't say that anymore.
MS. FRIEDMAN: I used to be a junior high school teacher many years ago and I would never have envisioned myself sitting up here doing something like this with something so advanced technologically. I mean, I don't know how many of you remember the mimeograph machines and my students really liked it because they got to smell the ink and they thought it made them do better on the tests.
Good morning. My name is Carol Friedman. I am with the Epidemiology Program Office and I would like to spend the next few minutes telling you about a rather unique or I think a rather unique application of the summary measure disability adjusted life year and how it will be used to assess burden of disease among working Americans.
In fact, this will be applied to a major corporation in the United States, General Motors. You are probably wondering -- the title is "Cars, Strawberries and Dalleys: Assessing Burden of Disease Among Working Americans" and you are probably wondering what connection do cars, strawberries and Dalleys have and probably thinking to yourselves, not much. Well, let me tell you a little story and I think you will begin to see a connection.
About three years ago, General Motors executives came across an article in the newspaper describing how CDC was able to link the source of an outbreak of hepatitis A among school children to contaminated strawberries. Now, General Motors was under a lot of pressure to cut costs, particularly on the manufacturing side and they were so intrigued with the technique that CDC used, to how they track and monitor warranty claims for their cars and were able to save the company -- now, this is according to GM -- $1.6 billion.
Now, GM was so impressed with the help that CDC gave them and the fact that they were able to save so much money that they asked us to come back and explore other ways in which we could collaborate in other areas. So, one of the areas that we have started to collaborate with G.M. on is what is called Partnering with Employers for Prevention, or PEP. The idea is to increase the use of preventive services.
Now, when we went up -- and I wasn't part of this initial meeting -- when we went up to General Motors to meet with them, we had a -- we essentially gave them a menu of options in terms of, you know, which areas would you like us to help you with, keeping in mind that, you know, from CDC's perspective, we really wanted to help work in the area of health.
One of the options was what is called business response to chronic disease. That involved the corporation, General Motors, working within the community to promote health promotion related to chronic diseases. Another option was a staff exchange where we, CDC, would send somebody to General Motors to spend a year or two sort of getting immersed in the General Motors culture and learning about the health benefits program at General Motors.
Then in turn, General Motors would send somebody to CDC to spend time at CDC becoming immersed in public health culture for like an exchange of ideas.
The other one was this party, clinical preventive services and then this last one was burden of disease was brought to GM's attention, the possibility of, you know, well, maybe we could assist in doing a burden of disease assessment for your company.
And of all of these, the one that GM was the most intrigued with and the one they supported the most and wanted to do was the burden of disease, assessing that for their company. In fact, I think it had to do with the fact that Matt McKenna presented such a wonderful case when he went up there and presented this to General Motors.
Let me tell you a little bit about the company. I am sure most of you are aware that the company does produce a lot of cars and, in fact, General Motors is ranked No. 1 in the world, the biggest company in the world, based on revenues. It is also the largest provider of health coverage in the United States. It has got 1.5 million enrollees. If you think about that, that is about the size of the State of Nebraska.
It contracts with 120 different health maintenance organizations, ADC for service carriers, and it also contracts with many different preferred provider organizations. Now, General Motors also has an at risk work force. A considerable proportion of their population are overweight, sedentary and smoke. I don't have any hard figures for you. They did do a health risk assessment on a very small portion of their population. But I will tell you that the workers, who work in the plants on the assembly line, are allowed to smoke and I saw it myself. I was there watching them put together some trucks and they were there putting the car seats into the trucks with cigarettes hanging out of their mouths.
DR. STARFIELD: Excuse me. Could you tell us whether those numbers include just the employees or the families as well?
MS. FRIEDMAN: Which numbers?
DR. STARFIELD: Any of them.
MS. FRIEDMAN: The 1.5 million enrollees include active, retired and dependents.
The other thing about General Motors is that in 1999, they spent $3.5 billion on health care. I believe that is more than the operating budget for CDC. I may be wrong.
To put this in perspective for you, that comes out to about $1,200 per car, which is more than what they spend on the price of steel for a car. So, when you are looking at a GM car and you look at the sticker price, the extra $1,200 goes to pay for health care benefits for their employees and enrollees.
So, as you can see, GM was really interested in wanting to, you know, work with us. I don't want to make it sound like all they wanted us to do was to help them save money. They are very interested in worker productivity and improving the health status of their work force.
Now, what is that GM or what do we think GM can get from a burden of disease assessment? We think that such an assessment can provide General Motors with strategic health information, resulting in the establishment of projections necessary to identify areas of emphasis regarding safety in their plants, health promotion and health services.
GM has just in the United States over 200 plants. These projections could also identify health issues that GM could emphasize, track and promote through their health insurers. Remember, you know, GM is a major purchaser of health care and they contract with many, many different health insurers.
These projections could also provide GM employees with a clear picture of what they can do to maintain and/or improve their own health status. So, by utilizing this approach within the General Motors community, GM and its employees could establish a standard for examining health that may have far-reaching impacts on all levels of our society.
Now, how will we do this assessment, the burden of disease assessment? What we are going to be doing is we will be using the disability adjusted life year measure that has been developed by the global burden of disease and we will apply this to mortality data that has been provided by General Motors.
General Motors maintains a death registry of all of their active and retired employees. The registry goes back to 1981 and they have about 13,000 deaths a year that they collect for this registry. If there is any data that is unavailable for the calculations that we need, we will infer this from data developed by the global burden of disease project, for established market economies.
The reference population, particularly studied, will be persons, who are eligible for death benefits because of employment by General Motors at a site in the United States. These can be active or retired employees.
DR. STARFIELD: Is it just employees there, too, or --
MS. FRIEDMAN: It will be active or retired employees. Unfortunately, we don't have dependents. There won't be dependent data available.
In terms of actually calculating the -- and we haven't started any of the calculations yet. All we have gotten so far from General Motors, we have gotten the data from them and we are just about ready to start looking at the data and calculating this. But we will be calculating Dalleys. We have got two files from General Motors. Both of these files have been prepared by General Motors under consultation with myself and Matt and some other folks from CDC.
We have a mortality file and a population file. Years of life lost will be calculated directly from General Motors mortality file and years lived with disability will be calculated from known ratios of years of life lost to disability and years of life lost in established market economy.
Then we will also calculate death rates by the person year, death rates over person year by person time method and we would like to stratify by cause of death and disability.
Work site, we definitely want to look at it by individual plants to see if things are different at different plants, age, gender, employee status and year as well.
So, just some parting thoughts. This is what I consider a rather unique application of burden of disease methodology in a new setting. As far as I know, I don't think that this type of methodology has even been used by private industry. I may be wrong about that.
I think what can come out of this is the development of a model that can be used for other corporations. Of course, whether this will actually save money for GM, you know, that remains to be seen.
DR. SONDIK: Okay. Thank you very much.
You all have stayed right on time. I appreciate that. It will give us more time for discussion. Very interesting.
Our next speaker is Dr. Diane Wagener, who is director of the Division of Health Promotion Statistics for the National Center for Health Statistics. Diane, in particular, is going to talk about applications related to and activities within healthy people and indices.
Diane.
DR. WAGENER: Right. I wanted to talk about two activities that were intimately related at the National Center for Health Statistics and that is healthy people and the tracking of healthy people, which is work that my division is primarily responsible for and also some methodologic studies that we are conducting at NCHS regarding a number of issues around summary measures.
Then I will make a few general statistical issues points. Some have been made before. So, if I don't mention them, it doesn't mean that I don't think they are important, as previous speakers have brought these issues up.
So, first, healthy people. Such a deal. Each of you get your own copies of this two volume -- and each volume here is about an inch and a half -- of Healthy People 2010. Now, you will be pleased to know we are not putting it on your desk. You don't have to carry it home. We are going to mail it to your work and I know that some of you have already received copies. So, if you don't want a second copy, please let Marjorie know.
But this was just released in a massive snowstorm in Washington, D.C. a couple weeks ago.
This --
DR. SONDIK: You mean literally, not figuratively.
DR. WAGENER: Literally. I couldn't get home from the hotel. It was an impressive event.
But this volume, which is our newest decade in this activity of healthy people, which began in 1979 and every decade we have set new objectives for the nation, was just released and I gave you a small spread sheet, which summarizes what we have ended up with in this particular version.
There are actually 28 focus areas in this document and a total of 467 objectives, which is up from the 319 that we had before. But this group of workgroups are really cagey. I mean, federal employees tend to be cagey at times. I kept imploring them to keep the number of objectives down and the number of measures down. You can see I lost out at 467, but worse than that, they have crowded several measures in some of these objectives.
So, in fact, the actual number of measures in this document are 790 at a minimum, because there are several objectives in here that are called developmental. We don't have data for them yet.
There are objectives like the number of states, territories and bribes that do thus and such. Well, that is not one measure. That is at least three measures; the number of states, the number of territories, the number of tribes. So, 790 is an underestimate of the number of measures we have here.
Approximately 74 percent of those measures are currently measurable. So, given the enormity of this, one of the issues that the surgeon general is particularly concerned about is having a small set of indicators that he could use and encourage, state, local individuals to -- or communities to track. He calls these the leading health indicators. So, in fact, there are ten leading health indicators that are meant to capture a broad range of all the issues addressed in this healthy people document.
The ten leading health indicators are physical activity, which I think you can isometrics with the book, if you want to. I don't know if that counts -- overweight and obesity, tobacco use, substance abuse, responsible sexual behavior, mental health, injuries and violence, environmental quality, immunization and access to care. These are listed in the first chapters of the book here.
So, these ten areas are particular indicators that he would like to use to follow. They are the actual measures for the indicators. They are all modifiable behaviors or services that can be provided. So, these are not health status indicators like we had in the previous decade. These are more determinants of health indicators.
Healthy People 2010 has two overarching goals. One of them is to increase the years and quality of healthy life. Hence, our interest in summary measures of health. We need to have ways of tracking this particular overarching goal. The second goal is to eliminate disparities of health. By that we are focused not only on racial and ethnic. We are focused on gender. We are focused on income, education, urban, rural. We are focused on a number of dimensions that one might want to be concerned about the inequalities of the health experience of the American public.
So, we are particularly interested in how one might use summary measures of health to address these two overarching goals. In the previous decade, a particular summary measure of health called "Years of Healthy Life" was developed to track one of the goals in that decade. We are continuing that, but we are also doing a number of methodologic studies to try to understand what the assumptions were that were built into that. Are there are additional summary measures that might be used.
So, let me talk briefly about the different methodologic studies we are conducting right now at NCHS. Our concern, of course, is that we have annual estimates with our summary measures because we want to track it annually over time.
So, one of the methodologic issues we are looking at are the use of various health variables in our health expectancy calculations. One can look at function. One can look at self-assessed health. One can look at disease. One can look at disability. There are a variety of different measures that are collected on a regular basis through our surveys and we were looking at these in terms of trends over the past decade in comparisons of various sociodemographic subgroups.
We are concerned about the comparability of data and the assumptions that are built into the data that are used for these summary measures. For instance, in many of the health expectancy measures that we are using where you have some measure of morbidity and some measure of mortality, the data on the mortality come from census -- the population data are census estimates. The data on the morbidity tend to come from our National Health Interview Survey, which is the civilian non-institutionalized population.
So, if you are going to start combining those two populations and those two estimates in a single figure, you are mixing apples and oranges. So, one of the things we have been doing is trying to identify other sources of data for the institutionalized populations in terms of their health and try to combine that in a meaningful way into summary measures.
The largest institutionalized populations, of course, are the nursing home populations and the incarcerated populations. Given the different age structures of those two populations, you get very different effects on the health expectancy measures and you get very different relative changes when you are looking at subgroups of the population because the different subgroups are in these institutionalized populations in different ways.
We are looking at the impact of population -- on the sociodemographic subgroups -- the continuity of data on activity limitation, which is one of the morbidity measures that we are particularly interested in is problematic for us because in 1997, the Health Interview Survey changed its questionnaire and we have been doing quite a bit of work on trying to understand how we would look at trends over time for activity limitations.
Then, finally, as I mentioned briefly, we are trying to identify new data sources. The health status of the incarcerated population, the Bureau of Justice statistics does have some data that we are -- it is limited, but we are trying to incorporate it and we are also looking into how we might combine some of the longitudinal data we have so that we have information on transition into the different measures.
Now, right now the most active longitudinal study we have deals only with the elderly, which is problematic for doing a general population summary measure.
So, now let me just mention a few of the broad statistical issues. I think it is important to recognize that the first question when you look at summary measures is what is the question. I mean, this is something that most of the previous speakers have alluded to. Based on what your question is, you would be dealing with either incidence or prevalence data. If you want to describe the state of the population right now, prevalence data may be adequate. If you are trying to assess whether or not you are going to have an impact or whether or not there will be a change in the future, incidence data becomes much more important.
Are you interested in specific diseases or general health? These are all very valid questions and in a particular use, the data would be different, that you would be driven by this. If you are looking at resource allocation, are you allocating your resources by burden or by impact in terms of programmatic changes.
If you are evaluating -- if you are using value judgments in how you would weight the different states of health, are you interested in intrinsic or the extrinsic, which was the VEN(?) diagram that Michael showed you earlier. So that there a whole range of uses from the monitoring that I am doing to surveillance, to resource allocation, to performance, to clinical studies, to looking at interventions. I think that from your point of view the answer is what is the question, well, it is yes, yes and yes, because from your point of view of looking at the entire range of the activities in the department, what we probably need is a montage of measures that would be used so that there is no one system that will be the ultimate answer.
Given that, if you have several measures, trying to encourage a coherence between them would be important; not only a coherence between the different measures, but also between the different sources of data. I am particularly concerned in tracking healthy people to be able to disaggregate it so there are a number of dimensions of the disaggregation that will impact on the data requirements that we have and trying to pull together data from different systems.
Then in terms of data needs, this is an area we are also looking at. Some of the measures deal with transition probabilities. Some of them deal with severity. Some of them are cohort-based data. Others are period-based information. So, the kinds of valuations we can use from you are issues like how can we accommodate this vast difference in terms of the data needs. What new data systems do we need? Do we need special studies to augment the existing studies that we have to provide us, at least on an occasional basis, the kinds of data that are needed to do some of these epidemiologic evaluations that were talked about.
How do we get there from here, given that we are in a world of where people won't wait and they are not eager to pay for it? So, how do we get there with these many different data systems?
Finally, ,what I would encourage and would hope that you would give us some advice on are research areas. A number of areas that we are not currently working on but one might find very important have to do with the usability. I mentioned that we have transition data for the elderly. We don't have it for the younger populations or at least recent data.
Also, what you do for a youth population in terms of evaluating health would be very different from what you might consider in your elderly population. So, across different ages, across different cultural groups, we know there are problems when you ask somebody what do you think of your health, problems in the sense that there is a heavily culturally bound answer there.
Maybe that is important. Maybe that is what we want to include in our numbers. Maybe not. That is something that ought to be looked at. The scope of health assessment and the psychometric properties were mentioned earlier. Then research into the evaluations of the different states, either self versus group evaluations and the effect of wording.
These are all important areas. In some sense with summary measures, we are pulling so many different kinds of data together at the same time, we have a vast array of research needs, all at the same time. But I think as somebody has said, it is an area that is active and it is international and we are going to continue to look at it both in the context of healthy people and other contexts at NCHS.
DR. SONDIK: Thank you very much, Diane.
Our last speaker before our discussion period is Dr. Mark Hayward, professor of sociology and demography. He is also director of the Population Research Institute at Penn State.
As I said earlier, he is sort of in effect here batting cleanup.
DR. HAYWARD: I am batting cleanup probably because I am academic. I am a demographer. Most of you know what a demographer is, but we are really -- I am going to put in a pitch for gene therapy for my occupation. We are really looking for a cure to give us a personality so that we could move up to being an accountant. So, gene therapy seems to be the only way that we can move forward on this.
I emphasize that what I am doing here I am giving you one academic's perspective. Actually, I am giving you two academics' perspectives. My co-conspirator -- I didn't want to put her name up here -- her name is Eileen Crimmons(?) and she is at the University of Southern California and Eileen and I have a very long working relationship with each other and we are pretty much in sync about how to think about how we do things in thinking about population health.
I am going to talk to you today about how a demographer might go about thinking about gauging population health. We are really looking at, as many people have talked about here, a multiplicity of ways of gauging population health. What we are really focused on is the individual life cycle experience. Many people say, well, life expectancy is a measure of a population. It is really a measure of the average individual's aging experience in a population.
So, we are really looking at a life cycle phenomena like life expectancy and we are very much interested in major chronic diseases, in particular, fatal chronic diseases. We are also interested in issues of impairment, functioning -- you notice, this is something I agree very much with what Michael said earlier, this issue of within the skin is a real important notion, functioning, which is limitation. We are defining it here in performance at the level of the person and disability.
And disability really is this incorporation of a social role that intersects with an individual's ability to perform at the level of within the skin. Now, I am going to throw up a little -- a crude graphic here of about the way that we have been thinking and this is a very common framework. This is not our framework. This is not anyone's framework. This is a framework that is shared across the world actually. If you are familiar with the ICIDH folks, this is just -- this is basically the same framework as they have, but we have called it a different name. Thanks to two American colleagues, Lois Fabruga(?) and Alan Jetty(?), we have a different name for it, but it is basically the same phenomena, which is that when you are looking at health in the older ages, you are really beginning to look roughly at the onset of chronic diseases, which lead approximately and roughly to impairments, which can lead roughly and approximately to functioning and which can then ultimately result in disability. It looks very clean. It looks very neat. It looks very hierarchical. It never works that way in the reality.
And, of course, guess what, you can die from any one of those positions. So, in fact, within any box, you can change your functioning status. As an example, you can get better. You know, it is not an evolutionary process as one moves down the spiral.
One thing I should note is as you move -- you are really moving from the organ system, from a disease, like a pathology down to impairment, which is like a blindness, which is functioning, which could be -- you could partition into upper or lower body if you wanted to do the NAGI(?) items or you could do it kind of within the skin as a whole person and down to disability, which usually refers to does your health limit the amount or kind of work you do, does it limit the amount of activity you can do in a social role.
Death, we seem to have a shared meaning on. For our purposes, we are really looking at all of these kinds of things -- actually we have skipped the impairment component and we are looking at diseases and functioning and disability and we are very much interested in the years of life that one spends with and without a particular disease or set of diseases, such as disease-free life expectancy.
We are also interested in the functioning, which is the years of life with and without functional limitations and also disability. One of the things that we should get used to is that how we gauge these different domains of health and what goes on in the population is often times dictated by the kinds of data that we have. So, we live within the constraints of data. I can give you a retrospective biography of how we have done things.
In our work, it looks very neat and very organized. I can tell you in reality it was never that way. Let me just say that we -- in our sense as academics, we are really interested in all three of those things because we think that the usefulness of all of those measures is it allows us to provide -- have information that can monitor population health in different ways.
For example, if you are interested in changes in disability, you really want to understand how changes in functioning and diseases are related to these kinds of phenomena, rather than just leaving them out there hanging.
DR. STARFIELD: Can I just ask you about disease? You mean, individual diseases or you mean people without diseases?
DR. HAYWARD: I mean individual diseases and I will show you that in a minute.
One of the things I was just going to make a notion here for the internationalization of the issues. Michael has talked about this and I really want to push this as well. We are beginning to see that most surveys of the health of national populations collect information on long term disability. The United States has made a pitch for this at the level of the census. We really concur with that.
What we are doing often times is moving towards a set of measures, which really are measures of functional limitations, which tend to be left not completely free, but less sensitive to cultural boundaries and those kinds of things. There is, of course, much less focus on things like self-report information on diseases, which absolutely don't make any sense whatsoever in some international contexts.
So, I would never ever encourage that. One of the things, of course is that when you ask a self-report on a health survey such as that that we do in the United States, a lot of times the self-report items frequently ask a respondent has a doctor ever told you, which, of course, is contingent on health service use and an accurate diagnosis. We rarely have indirect self-report measures or screens available to us, except on some things like cognitive impairment.
I am going to skip some things. Let me just talk a little bit about functional ability, which is where we have done quite a bit of our work. One of the nice things about functional ability is that it seems to -- it is made up of your old ADLs and IADLs and there is an infinite -- a growing literature on how to measure all those things. Let us emphasize that in many cases there is no gold standard as far as we are concerned.
One of the things that we think is happening and I think because all of you are familiar with how to measure them, one of the things that is useful in the IADLs and ADL world is that there is a growing consensus that the ADLs and the IADLs can be used to develop an ordinal scale of health, where you can look at severe and moderate levels of health problems. No functioning problems, for example, some functioning problems. Unable to live independently and unable to provide for self care of an individual.
I am going to show you an example that Eileen and I have done. There is actually an article in your packet that gives you an example of the work that we have done using the kind of ordinal scaling of these things. It is kind of parallel to the work on the severity weights, except that we have shied away from the weights. Our reason is not political. It is not policy. It is that we are nervous that the weights are totally endogenous with respect to the health process itself.
We can't be confident, unlike economics, that preferences are exogenous. So, because we are really wanting to explain things at the academic level, we tend to shy away from weights. There is absolutely no reason not to use weights. There are many reasons for using them. So, I don't want to -- I agree with the kind of catholic perspective that has been endorsed up here about you have to decide what the question is before you can figure out what the measure is.
The point is, yes, we think we can kind of develop an ordinal scale, based on the ADLs and IADLs. These levels of the ordinal categories are strongly associated with mortality and the nice thing is that this gives us an idea about the kind of changes in health that occur with AIDS. How many downward -- how many declines in functioning and of what type tend to occur and how many people actually regain function over some period of the life cycle.
DR. STARFIELD: Could you make it clear as to whether you are dealing just with the elderly in your comments?
DR. HAYWARD: In this one we are dealing just with the elderly because IADLs and ADLs have very little application to a younger population.
I am going to talk a little bit about some work that we are doing on disease-free life expectancy approach to modeling population health. I have given you a crude outline here. The first box that we are assuming that people start out in is a box called "no" and I should have put "fatal chronic disease" in here because we can have arthritis, which is clearly something that is characteristic out there.
There are many ways to escape this box. One way is that we can go directly to heart disease, do not pass go, right into heart disease. Another way in terms of our data is that you can move into some other fatal chronic disease, like diabetes as an example. You can go from no disease into a comorbid condition that we have created called other and heart disease and mostly because, I mean, in reality it is probably not true that people get two fatal chronic diseases at the same point in time, but given our data, that is what we observe.
Of course, I designated death as in terms of black with the black arrows coming out heart with the red arrows. I don't know if the red shows up there. It is kind of blood red. So, what we are looking at is -- what this kind of rough schematic allows us to do is to begin to apportion life in the various states of being. How long one would live without any kind of fatal chronic disease, how long one would live with heart disease, how long one might live with heart disease and other fatal conditions, just other fatal conditions alone.
One of the things that are not nice and neat, but we actually can get expected trajectories of health from this kind of model, where do you go first? In particular, when you are looking at different population groups, where you go first might be relevant.
I can get into specific definitions later. I hesitate to show these numbers because they are clearly preliminary and my estimation routine is off and we are nervous and the whole bit. So, let me just borrow a piece of paper here so I don't really mess things up.
I will just focus on females. Okay? What I have here is a set of life expectancies and these life expectancies for females are slightly off. They are slightly lower than they should be and that is a matter of our estimation routine.
But what we have here is females -- the first group is females, age 50; the next group is females age 60 and the next group is females age 70 -- whoops -- 50, 70, 90. Okay. Refuse to look at the males. Males are irrelevant here and I want to focus on females because it is one of those cases where we -- if we think heart disease is not a female problem, well, I think we have, you know, pretty good evidence that it is and all of you know that anyway.
We have estimated and this is an estimate and it is a rough approximation that females will live at age 50 another 29 years and of those 29 years, roughly ten of those years will be spent in a category that we call inactive, which means that they will be unable to carry out one of the ADL measures that we have used to benchmark activity in this analysis, that on the average, women age 50 will spend 13.7 years in the no-fatal disease category, but they will have 2.2 years of life in the no-fatal disease category when they are functionally impaired. So, that is the kind of function problem that is coming from non-fatal conditions.
They can expect to live approximately seven years with other fatal chronic conditions 2.5 of those years and with being functionally limited. They can expect to spend 2.6 years of life with heart disease only. Now, notice the other disease and heart disease. This is the comorbid condition. This is the one we are often times interested in. Here we have -- they live roughly 5.8 years with other disease and heart disease and 4.1 of those years is spent with a functional limitation.
Here you can see the kind of link between -- and it is an associated link. It is not a deterministic link. It is a link between comorbidity in terms of disease confluence and the implications for functional limitations. That is where it really increases in terms of the proportion of life that one spends in this particular disease category.
Now, I am going to move on a little bit and see where we go, but just to make a -- I want to make a quick point here and if you just look at the upper panel, people can spend a long, long time of life living with a disease, a fatal chronic disease. I mean, yes, we call them killers, but, my goodness, they can last a long time in a person's life. That is because we are pretty good and we are successful at keeping people alive and moderating the negative consequences of many, many, fatal, chronic diseases.
You have a long period of life, a much longer period of life that is spent with a disease, compared to the period of life that one spends functionally impaired. The period of life that one spends functionally impaired is quite small relative to the period of life that one might spend with a chronic disease.
This graph is a -- this slide here is an example of where we are looking at the functioning problems that occur with age. This is, again, in the elderly population because we are looking at ADLs and IADLs and one of the things that we should notice is that we observe in our data, in the longitudinal data set, that people can jump around, believe it or not, in all these different directions. Of course, for those of you that are clinicians, you do observe people that jump around in all these different directions.
One of the things that was nice is we were actually able to track these kinds of changes that occurred that allowed us to use a model, as Diane was saying. There are models out there that are based on prevalence. There are models out there that are based on incidence. This is a model that allows us to use incidence rates to track changes in the health biographies of individuals.
Again, one of the things I note from this kind of model is it really doesn't give you a way of tracking the expected trajectory of health decline in the population.
Oh, one of the things I got out of order was one of my slides. Let me just go back and show you from that little model that we were generating before on heart disease, this is the prevalence of heart disease for females -- well, I shouldn't say for females. I just say females. Forget the 50 years of age. Oh, yes, that is 50 years of age and older.
So, we start out here in the population. Here is 50 years -- here are women 50 years of age and basically the majority of that population has no fatal chronic diseases. We see some other fatal chronic disease, other than heart and heart. Very small. Now, look as -- with age and once you get into the upper ages, the comorbid conditions are huge and, of course, comorbid conditions with heart disease, not surprising, given the prevalence of heart disease anyway are really amazingly broad and these are females.
So, I mean, anytime you want to think about campaigning for heart disease as a female problem, I think there is a lot of evidence.
How am I doing on time? Am I done?
DR. SONDIK: Yes, you are really done.
DR. HAYWARD: All right. I am toast. All right. Let me just take one last chance and show you -- I won't show you the stuff that we have been doing on race/ethnicity, but I am going to make a pitch for race ethnicity because when Matthew was up here talking about the huge disparity that exists in race/ethnic groups, it is something that we are working very, very hard on in the work that we are also doing in terms of benchmarking racial disparities in the burden of disease.
I think our work will pretty well compliment the kind of work that Matt is doing because there is enormous disparities that are masked by the average expectancies that we pick up in our statistics, enormous disparities.
So, where are we going? Where we are really going is trying to demographic model that integrates all of the major fatal chronic disease conditions. We are pushing that because -- and it is a lot like the POET(?) model that has been developed in Canada -- we are pushing that because there are lots of counterintuitive ways that diseases combine and we don't really understand all those things that go on. We want to pay particular attention to comorbid conditions, especially as they arise among the race/ethnic groups and across different social groups. We really have no good way of doing that at the current time in terms of benchmarking how one experiences health over the life cycle.
The other thing that we are doing is we are really pushing this model towards a microsimulation approach and, again, I am not sure whether this is similar to the POET or not, but our goal is not to get an expectancy. Our goal is to get the health biography of every individual in the population. The reason is is that there are pockets out there that we need to begin to identify in terms of the frequency of particular kinds of combinations of health experiences that we want to begin to play out.
Again, it goes into Michael idea. We do have at the government level simulation models that do very sophisticated work at predicting what is going to happen at -- you know, who knows what we could do.
So, the bottom line, where are we going, we are really interested in the racial burden of disease. We are really interested in social inequality of disease and we are really interested in the kind of counterintuitive and counterfactual things that can come out of especially the microsimulation models.
What happens when you change smoking? Well, smoking doesn't just affect one disease. It affects other diseases and it may actually increase other diseases in ways that you didn't anticipate before you ran these kinds of models.
So, end of story. Thank you.
[Applause.]
DR. SONDIK: Thank you very much.
I would like to open the floor for discussion.
Questions, comments? I guess we don't have any. Yes, over here. Paul.
DR. NEWACHECK: I want to thank the panelists for really terrific presentations. I have a question for Dr. Friedman about the General Motors experience.
I was intrigued by your description of the interest in conducting that work. I wonder if you could talk a little bit more about G.M. intends to use the disease burden data and what you think about the prospects of other private sector companies, using this kind of information.
MS. FRIEDMAN: Well, I really think what G.M. wants to do with this information is they really want to look at it at this point plant by plant because they have over 200 plants and the plants are actually different in terms of what they do and let's say, you know, it is a Saturn plant. It is set up differently than say the light truck assembly plant in Warren, Michigan. They could identify, you know, what are the leading causes of, you know, disability or injury in this plant and what interventions can we put into play to reduce, you know, these injuries or have an impact on the burden of disease.
I think -- I mean, in the back of their mind -- we are dealing with big business and, of course, in the back of their mind is always to have some way to reduce health care costs, but they are also are really interested in increasing worker productivity and I think that General Motors is unique among a lot of the manufacturing companies in that they take a real interest in improving of health care services, improving work place situations so they can increase the health status of their workers and as a result increase worker productivity.
DR. NEWACHECK: Does it seem like there is the additional interest by other corporations?
MS. FRIEDMAN: Well, that is an interesting question. You know, I had mentioned the strawberries and cars and everything and that appeared in The Wall Street Journal about a year ago and shortly after that appeared in The Wall Street Journal, Ford Motor Company called us up and said, oh, well, we would like to save some money, too.
So, there has been interest. We have gotten some in -- Caterpillar has expressed an interest, as I mentioned, Ford. There has been a little interest expressed in the air -- by Delta.
MS. SALINSKY: I wanted to ask a couple of follow-on questions to that.
How comfortable are you, particularly moving to a plant by plant analysis giving that you are using that you are using a lot of inferred from a pretty aggregate data source that is actually going to reflect the disease burdens within those individual plants?
My second question was has the UAW been involved in any of these discussions?
MS. FRIEDMAN: Are you with the UAW?
MS. SALINSKY: I should disclose that when I was with the Lewin Group, General Motors and the UAW were a former client and we were working on some community level interventions. And I was curious as to how they have been integrated in practice.
MS. FRIEDMAN: Let me answer your second question first, which is an excellent question.
This is a really new collaboration that we have going with General Motors in the areas of health, particularly where we are looking at health claims data. I mean, NIOSH has had a very long tradition of working with General Motors, but, you know, primarily in the area of occupational health.
So, early on we did express a lot of interest in pulling the UAW into this and we have been reiterating over and over again, we really want to get the UAW in on this. We feel like they should be if not active partners or full partners in this collaboration, at least be well informed.
G.M. sees the importance of that and we are going to be pulling the UAW in as a result of that. Haven't quite figured out -- see, I am leaving it up to G.M. I think it should G.M. to go to the UAW and say, oh, you know, we are working with CDC. We are doing this -- you know, we are doing all these neat studies as opposed to CDC, us, circumventing G.M. and going directly to the UAW.
One other thing that has come out of this is when we first started working on this, we didn't have any kind of a formal memorandum of understanding in place. So, we just recently drafted memorandum of understanding and in it it, you know, specifically lays out, you know, we want the United Auto Workers to be involved in this and, you know, other partners as well.
To answer your first question, I don't know. I mean, I really don't. All we have gotten so far has been the two data files I mentioned to you and we have had a few problems with the data files and we have had to go back and forth, you know, to G.M. to get the data files corrected. I think this is going to be sort of an evolutionary process. We are going to see what we can get out of the data files that they have sent us so far.
I know that that is G.M.'s interest is to try and look at it, you know, not maybe every plant, but, you know, take a subsection of their plants and take a look at that.
DR. SONDIK: I would appreciate it if you could give us a reference actually on the inference of the morbidity data from the mortality data, which is in essence, I think, what you said you are going to do. You are going to take years of life lost and then estimate the disability or the morbidity data from that.
DR. MC KENNA: Ed, if you read the global burden of disease, they have projections in there and that is what we are basically talking about is instead of projecting into the future, you are projecting onto different populations and the methodology that was used for future projections use that and it is explained in the textbook, where they use the ratio of disability to premature mortality, based on demographic and other socioeconomic changes that are anticipated in populations, made predictions about mortality and then computed disability based on those current ratios.
So, it is basically that methodology. I would just say that we have been as straightforward as possible with them about our concerns around these inferences and in fact, in some ways -- excuse me?
PARTICIPANT: Who is "them"?
DR. MC KENNA: General Motors -- about the difficulties and the potential pitfalls, just inferring the way we would be.
I don't think they see this as something that will be used directly or I think they understand the limitations, but what they are looking for are alternative ways that have epidemiologic bases to look at the impact of health -- negative health effects and health conditions on their population, other than their administrative data because they are acutely aware, based on their own analysis of how sensitive their administrative data is, one, to their health benefits package that they put together and, two, the practices of the local provider community.
They want some kind of reference point or some kind of gauge based on good epidemiologic information about what the best guess is of what the real problems are in those populations. They may change some of their benefit plans to adjust those.
MS. FRIEDMAN: In fact, I had done an analysis not related to burden of disease, where I did look at the use of a couple of clinical preventive services and looking at it by the type of health plan that the enrollees -- and it turned out that there was a much lower use of, say, mammography for those enrolled in fee-for-service and that it turned out that fee-for-service enrollees did not get a covered health maintenance office visit. So, as a result, I mean, G.M. is just considering this, but they are actually thinking about adding a covered office visit for their fee-for-service enrollees.
DR. STARFIELD: Given the original name of CDC, Centers for Disease Control, it is not surprising you pick a disease-oriented approach, but you have got this other name added on of health promotion, which is person oriented, not disease oriented. I mean, how do you deal with the very real phenomenon of comorbidity, which Mark Hayward showed us very graphically and now, you know -- there is not much literature on it, but it is coming out now that people are thinking about it? But how in the world do you deal with that?
You know, there are other methods, other than Dalleys.
DR. MC KENNA: I am not propose that Dalleys are the solution to that problem. As a matter of fact, I think they are abysmal in dealing with that. We see this very much as a first step. You are right. It is very biomedical disease oriented kind of approach.
I come from a center that has a large cadre of people who are very interested in health promotion sorts of activities, cross cutting, not disease oriented, looking at life style and those sorts of things.
However, from a policy standpoint and in discussions with individuals, there has not been in advocacy groups or in governmental agencies there is -- it is sort of like when our former director had a priority of infrastructure. Nobody's wife ever died of infrastructure. Nobody has got a lot of interest in that. So, there has to be some sort of framework for linking this or making the type of attribution that Michael was talking about.
I personally think you are right, that there is a whole area of comorbidity that needs to be extensively explored, but in our discussions with people who are further along with this than us, the Dutch and so forth, we have been warned that this is an enormous area. It requires a lot of resources and just off the cuff -- this is totally anecdotal -- that they are not really sure in the context of delivering the type of information policy makers are interested in; that is, the juice is worth the squeeze.
So, we are moving forward and hoping that others, a little more sophisticated, can lend to that discussion. We are just trying to get things started, using this particular approach.
MS. FRAWLEY: Carol, one of the things this committee has been concerned about is privacy and confidentiality and I was curious from the data that you received from G.M. whether it was non-identifiable data?
MS. FRIEDMAN: Yes. We made that very, very clear to General Motors that the data had to be stripped of all unique identifiers and that is how it is. And we also -- at CDC, we did go through an IRB related to this project.
MR. SCANLON: Do any of the panelists know of any work or you are doing it yourself, where you are -- where the work is bringing the previous work on the economic burden of disease and the national health accounts, along with this focus on population health status summary data focus? Is there a correlation, is there a way of integrating sort of what has been learned on the economic burden side, on the national health accounts with this population measure of health status?
Michael, you have probably thought of it.
DR. WOLFSON: Yes. We are in the thick of producing such estimates jointly with our laboratory, Center for Disease Control and the Canadian Institute for Health Information. As a number of my fellow panelists have said, this is not easy.
If I could maybe take the opportunity to jump back to Barbara's question, I wonder if a way to think about it -- the title of the session is summary measures, but you raised the question about comorbidity. Maybe it is useful -- certainly, in my own mind, I distinguished the indicators from the analytical tools that one builds around them and the best analogy for better or worse is an economic one. It is one thing to produce the national accounts and it is another to produce the macroeconometric models that do all the forecasting and stuff. You want to make sure that the accounting structure is designed to support the kind of analysis you want to do.
In that regard if one is concerned about comorbidity and the health of the whole person, then one really ought to start with summary measures that deal with the health of the whole person, for example, defined in terms of functioning and then say, well, in order to understand why a person is functioning one way or another, there may be one or, as Mark's analysis showed, there may be several diseases that are contributing to that and one needs to develop a set of longitudinal surveys or administrative data that allow you to put the pieces together and work on it.
This is not going to be an easy project. We are certainly, I am very happy to say, reasonably well advanced in having the kind of data systems that will allow us increasingly over time to disentangle these things, but it is a big project. The indicators of one part, the developing the data systems to allow you to put the pieces together is another.
DR. COHN: First of all, I want to thank our speakers. This has been fascinating and I think potentially a very useful set of discussions.
I did have -- I think from my view, I am sort of listening to all of your different methodologies and there are lots of different methodologies that we are hearing here. The one thing that I couldn't figure out as I listened or read any of your documents was how much of the work you are doing is based on what I would describe as sort of standard data sources versus survey instruments versus, et cetera, et cetera, et cetera, additional pieces that you have to go out and get additional information from the populations around versus, as I said, what is readily available.
Can each of you comment only because I think at the end of the day we would like to be able to have data available to meet sort of legitimate needs as opposed to having everything being surveyed independently.
DR. WOLFSON: I thought there was a neat little point tucked away in one of Matt's slides about data gaps and analysis gaps. When we started our work in this, we started not from the point of view of what data exists in Canada because then we wouldn't have been able to answer our question about data gaps, but let's take an important question. How healthy is the population? How much of their less than good health can be attributed to different things and then said what are the implications of the kinds of data we would need to do that.
We are proceeding along many, many or a number of routes, not all of which may be appropriate here, but let me just mention one thing that came up and I sat in on the data, the Population Committee, yesterday afternoon and there was a question of whether or not to put functional status on the administrative records here on a routine basis.
We have been conducting a national population health survey, which we are about to enlarge to the analogue of the scale of the behavioral risk factor survey here and 95 percent of Canadians have basically agreed with us linking their health survey responses to their health records. So, we don't need to think about putting functional status on the health records because we ask it along with a bunch of other things, whether it is ADLs and mental health type questions on the survey and then we can link it to their various episodes of visiting doctors, visiting the hospital and create these person-oriented longitudinal files.
Now, these are going to be useful for a whole lot of things. They will be useful for us constructing the summary measures of population health, like health-adjusted life expectancy that I was describing. They will also be useful for us to disentangle the upstream causal pathways, not only to diseases but further upstream to smoking or cholesterol or other kinds of risk factors, including socioeconomic things, which lots of folks are keen on.
DR. WAGENER: Let me briefly answer with regards to the healthy people. As I mentioned, the summary measures we are dealing with right now are largely focused on the overarching goals. I would imagine, I would envision in the Healthy People 2020, the next decade, they are going to want to have each of these focus areas and that will be the kind of drill down that you are talking about in your model. Right now we can't do that.
We don't have the data to have summary measures for each of the focus areas. Right now with the overarching goals, certainly for the national data, we are focusing on data that already exist and are regularly collected because we are in the field now. We need to be tracking right now, but even so we don't have the data for all the different subpopulations I characterized. I can't give you years of healthy life with the Asian Pacific Islanders right now, for instance. So, there are a lot of data needs even at that global macro national level.
DR. MC KENNA: The only comment I will make is that we are fairly ad hoc at this point. The data I presented you is what I think is probably the best quality data we have in the nation right now and even some people think that is inadequate, which is the SEER registry information, which is a way to get a cancer burden in the country that doesn't depend on self-report. It depends on what is our well-accepted, top flight, world class registry information with good follow-up of patients.
We feel pretty comfortable with that information that we can calculate disability adjusted life years, with given the assumptions and the weighting with that. We engage in an activity -- when you are ad hoc like this and using whatever you can find -- stealing once again from Michael Wolfson -- data confrontation, you have a survey that says one thing and you have got a variety of other data sources that say another. You confront them and try to resolve them the best you can.
Some of that includes expert opinion, but that is what we are working with.
DR. WAGENER: Excuse me. May I grab the mike back just briefly.
I was asked to tell the people on the Internet how they could get these two large volumes if they really wanted them. There is an Internet site called WWW.Health.Gov/HealthyPeople, all as one word and that should be a source for you to get the documents.
DR. HAYWARD: In the three studies that we mentioned -- that I mentioned earlier, the functioning study is self-contained within a very large national representative survey. So, we were able to track individuals. It is a longitudinal study on aging. So, we were tracking individuals for quite a long period of their lives and estimated both mortality and functioning changes that occurred.
In the disease model, we are also using a very large, nationally representative survey to track individuals with diseases and functioning changes. With the study I didn't really go into -- thank you, God -- was the race and ethnic disparities on disability. We used many data sources and it is one in which is fraught with nervousness because to -- we came up with the best estimates that we could at the time on Hispanics, lots of controversy surrounding Hispanic health in the United States. We established estimates for Native Americans.
Let me tell you we probably want to place a confidence interval a mile wide around that. We have Asian and Pacific Islanders. The estimates that we are most comfortable with are the ones on black Americans and white Americans, but when we get into the other race ethnic groups, it is very -- you know, you are treading carefully.
MS. FRIEDMAN: I was just going to mention, well, we have got the mortality file, that we have got that information from the death registry that General Motors maintains and also, General Motors has the ability to link -- they can link records from their death file with administrative or personnel files. So, that would be a way to get additional information.
If we have to go to a survey, that is going to create some problems. We have mentioned survey to General Motors and it is like, whoa, well, we are definitely going to have to get the UAW involved. So, we are not even talking about that at this point.
Then as Matt mentioned, we will be making some inferences.
DR. SONDIK: Actually, I think Dr. Cohn's question is very important for the National Center because it relates to our long term strategy for the information that we would be collecting. So, we would be -- speaking for the Center, we would be very interested in your guidance on the implications of what you have heard here today for the direction of the resources that we have at our disposal.
I think there are a number of changes that we could make, modifications that we could make, but I think it is kind of interesting that we are in a sort of chicken and egg situation here. Almost every member of the panel here said that the measure should be derived from, if you will, the uses that it will be put to.
On the other hand, that is an ideal and I don't think one can do that without experimentation and research and so that in some sense, we need the data in order to be able to help us identify which measure would be useful for which types of purpose.
So, this, I see, as a process that we need to move forward on. But it is quite clear from the set of examples that you have heard here, I think, of the importance of this and to me it is part of the nature of the devolvement in health, that local areas and/or general motors or other companies are clearly interested more specifically in the health of their charges, whether they are employees or populations and what can be done to improve that.
It is such a difficult area that there is no way out, I think, other than to use these indices as stepping stones toward understanding what types of actions to take.
MS. COLTIN: We had two other committee members with questions. Kepa.
DR. ZUBELDIA: Yes. My question is on a different line and it is following up on the question from Kathleen to Carol. But I would like to get the reaction maybe from other members, too.
You mentioned that G.M. gives you the data deidentified. If we roll the clock forward two plus years and HIPAA is then in effect, the current HIPAA privacy proposed rule would require that you also remove the date of birth and that you remove the city, county and zip code. With that kind of cleaning of the data, would you be able to do this type of work? Would you be able to conduct a study, for instance, plant by plant or to identify morbidity and mortality by age groups and things like that?
What is going to be the impact in your work if this happens?
MS. FRIEDMAN: I think it would make it more difficult.
DR. WOLFSON: If I can comment from a Canadian point of view -- I am not familiar with the details of HIPAA, but let me be clear in Canada that there has been a lot of agonizing, as well, about privacy and confidentiality. There is a clear and widespread appreciation of the difference between the use of health data for caring for individual people and health data for epidemiological and statistical purposes.
There is a widespread consensus that one needs to carve out for bona fide statistical uses the ability to maintain the kinds of items you just mentioned on the records. Deidentified, but non-nominal, but still potentially identifiable because otherwise it would not be useful.
MR. SCANLON: If I could talk to the same point, Kepa. There were -- the proposed privacy rule, first of all, has an authorized disclosure for research that actually would allow the disclosure of identifiable information, along with other conditions, such as an IRB or an IRB-like review. So, it is -- which is not that much different than what occurs today in at least publicly-funded research, but in addition and what you are alluding to, there was an approach to deidentified information you will recall in the proposed rule. If the information is deidentified, then, of course, it is not a disclosure anymore.
The deidentification could take place along two routes. There is one route that applies to what are termed statistically sophisticated entities, such as large research organizations where there is a fair amount of leeway given to how the deidentification would take place. It wouldn't necessarily preclude any of those items.
Then there is another level, which literally lists about 18 specific items that Kepa is alluding to, which sort of would apply to those that are theoretically not so sophisticated and it was meant to give guidance to, say, the individual doctor's office where there is not -- each doctor's office doesn't have an epidemiologist or a statistician, obviously -- to give some guidance to the field generally where there is not a lot of participation necessarily in research and where some guidance at that level was given.
They literally list, how many -- count them, maybe 9, 10, 11, 18 even items that would have to be stripped. There are ways to get around that as well, obviously. You don't have to put date of births. There are ways of aggregating that information. You could have age at large birthday. There are ways to get around this, but at any rate, in many respects, the research situation has not changed, at least for publicly-funded research wouldn't be changed.
For the whole area of deidentification and anonymous information, there is this new -- there is an approach that is sort of -- would accommodate a statistically sophisticated unit or research unit to be able to deidentify a tape with a fair amount of flexibility, as long as the end result was minimal.
For those less sophisticated -- and that was really intended as guidance generally -- there is literally a list where you would say if you have taken these off, most likely you have deidentified.
DR. FRIEDMAN: I wanted to follow up on the plant level measurement and on Simon's question about ongoing data sources and it is essentially a question for Diane Wagener.
The year 2000, the consensus health indicators, have the disadvantage of being largely mortality based, but at the same time they had the advantage of really being scalable from census tracks to -- or from block groups to census tracks to counties, to states, to the nation, et cetera, et cetera. And it seems that the 2010 leading health indicators have the advantage of not being mortality based and at the same time they seem to have really lost almost all of that scalability.
I am wondering as NCHS is investigating summary measures and tracking 2010, the extent to which in those summary measures scalability is something that you are using as a criterion.
DR. WAGENER: Let me address a couple of things.
The last question about whether or not we are currently addressing the scalability, that has not been on our active radar screen, but it is certainly something in the back of our minds. The BRFSS at the state level did collect some data in the early 1990s that could be used in this sense for many of our summary measures and they do right now actively include some measures, like self-assessed health that can be used as a summary measure at the state level.
With regards to the leading health indicators, as opposed to the health status indicators, health status being the mortality and the disease incidence of the previous decade, the focus there, the determinants of health model of the leading health indicators, I would argue are scalable but the data don't exist. So, there is a distinction there between what already exists, which the health status indicators did have as a focus and what doesn't exist because these limited questions that one can get through a telephone interview on fiscal activity could be done at a local level and are being done in the BRFSS. Many of the statistics in the leading health indicators can be derived from the BRFSS right now.
There are -- and I won't stand up and defend a couple of the measures, which really are going to be very difficult to assess because they are low frequency values. So, at a local level, you are just not going to have the outcome that will give you a reliable statistic.
Like the mental health measure is whether or not those with depression are being treated. I think that is going to be very difficult to assess at a local level, but we did try to get as many as possible that were scalable to a local level. But the data don't exist. I mean, I am willing to say we don't have questionnaires right now that are in the field that provide that data.
MS. COLTIN: Is this on this point?
DR. SONDIK: It is a little different.
MS. COLTIN: All right then. I am going to take the chair's prerogative. I have a question also for the panel, because we are running low on time.
As you may or may not know, HCFA has made a decision to use the SF36 to develop a performance measure to evaluate care provided to Medicare enrollees and HMOs in this country and to look at two year change scores in the SF36 to do that.
The instrument that they have developed to collect that not only collects data ont he SF36, but also includes ADLs and IADLs and self-reported disease conditions. There is the possibility also clearly to link the results from these surveys with other data that HCFA has available, as well as information on mortality in that population.
Is the CDC or any of your agencies thinking about working with HCFA to look at comparing different methods, some of those you have discussed today, for doing the same sort of thing because clearly with the data they are collecting, you could use the Dalleys' approach. You could use the IADL, ADL approach and compare and contrast it with the approach that they are taking looking at changes in the SF36.
Given what some of you have said about sort of the developmental nature of this science and these methodologies and their decision to go ahead with a particular methodology that really isn't that tested and to put that information out into the public domain, it would seem to me a good policy to at least evaluate their methodology and look at some of these other possibilities.
Is this on the radar screen or something that one of your groups might be interested in looking at?
DR. SONDIK: Why don't you start and then I will finish.
DR. WAGENER: In the Office of Analysis and Epidemiology at NCHS, we actually have a person on detail, so to speak, from HCFA. We are very interested in trying to work with them to evaluate these data. We are still very early, as you said. This is just something that is -- the data collection certainly hasn't been completed. So, we are still very early in the process, but we work with HCFA on a variety of data requests and also on a variety of data analyses. So, we are certainly looking at that for the future.
It is not on the horizon right now because we are still so early.
MS. COLTIN: The first round of two year change scores, those data are in the field now.
DR. WAGENER: Yes, right. But it will be a while before they come out and they are cleaned up.
DR. SONDIK: It only recently came to my radar screen, actually, and I think it is very important and I look forward to working with HCFA on this. I think it is -- one of the things we don't know n this is we don't know sort of, I guess you might say, the Type 2 error in this. We don't know how far off we are in applying a particular measure to a problem and then taking actions as a function of what we find with that measure because we just don't know whether those measures really accurately reflect the problems that we are dealing with, either accurately or how much they are leaving out from that. It may be that we are looking at one problem with it, but, in fact, the issues are really in another area.
It is really critical that we do the research to understand the implications of all of these measures and how to apply them.
I just wanted to make one point for something that the committee is heavily involved in and that is the vision for health statistics.
I see this as very much related to that. In particular, I go back to Michael Wolfson's first point about a coherent system and the implications of that and the implications of being able to do drill down or scale up or however one wants to look at that. I think that is a very important part of a vision and one of the things we have to assess, we in general have to assess, is what is feasible, what isn't feasible, what is useful in all of this.
This activity -- and I think you have seen really five really excellent examples of it, I think you can see how it relates to a current vision and a future -- current reality, if you will, and a future vision. So, I encourage all of us to keep this in mind as we think about the broader vision and think about the implications for the information that we should be collecting routinely, as well as the information that we collect on non-routine basis, which can be used to inform many different applications.
Not all the data that we collect has to be collected on a constant continual routine basis and special research studies, I think, are very much of -- in my view, anyway, are very much a part of a vision for health statistics. All of that needs to be combined.
MS. COLTIN: Okay. Well, it is time for us to take our break for lunch and we are due back here at 1 o'clock.
I would like to thank all of the panel for wonderful presentations.
[Whereupon, at 12:15 p.m., the meeting was recessed, to reconvene at 1:07 p.m., the same afternoon.]
DR. LUMPKIN: Why don't we get started.
First of all, I have a 5:30 flight out of Reagan and that is because the 6 o'clock flight was completely booked by the time we set our arrangements. So, that means we have to be done by 4:30, unless there are some objections to that.
MS. FYFFE: We need to be done by 4:00.
DR. LUMPKIN: We do have a panel --
PARTICIPANT: From 3:00 to 4:00.
DR. LUMPKIN: Oh, is it from 3:00 to 4:00?
PARTICIPANT: Well, it may say 4:30, but I think it is --
DR. LUMPKIN: Okay. So, we are going to shoot for when our panel is over. We hope to get all of our business done before the panel arrives. Okay.
The first item that we have is the 50th anniversary symposium. You have at your places a rough agenda for that. We will start the day off with the regular business meeting of the committee. We will adjourn around noon to go to the Academy and then we will start at 1 o'clock with the welcoming with a representative from the National Academy of Science Committee on Health Statistics, who are the hosts for our meeting.
After that, we will have reflections and challenges on the past of the role of the committee by three past chairs, Don Detmer, Judith Miller-Jones and Ron Blankenbaker.
Followed by that, we are looking to have comments from notable individuals, such as John Eisenberg from the Data Council and others. We will follow up then with a presentation of the interim NHII report. We discussed that at the NHII work group and we are looking for a report, roughly about 20 minutes of a presentation, ,followed by two reactor speakers and we will be seeking two individuals that we can sort of get different perspectives as a response and then open discussion for the rest of the hour and a half.
So, that would be roughly 50 minutes of discussion. Following that will be a presentation from the Health Statistics for 21st Century Workgroup of their interim report and then a reactor speaker panel. Haven't thought about that yet.
Okay. Either a speaker or a panel, yes, followed by closing remarks and then a reception in the Great Hall of the Academy. Then we will spend the following day with business of the committee.
DR. STARFIELD: John, could you clarify that? I have got in my book that we have got the following two days for committee meeting, but not the morning of the symposium, which is --
MS. GREENBERG: Oh, you didn't have the morning of the symposium?
DR. STARFIELD: Well, it is just not in my book. We have the following two days or one day?
DR. LUMPKIN: Well, and then the third day would be available for subcommittees if they choose to meet.
DR. STARFIELD: Okay. So, in fact, we have to set aside three days.
DR. COHN: The days of the actual meeting are the morning of the 20th and 21st and 22nd. No? Yes?
DR. LUMPKIN: And the 21st and then the 22nd will be available for subcommittees to meet.
MS. GREENBERG: Unless there is -- you are thinking about the CPR --
DR. COHN: I think my only concern is is that depending on what the business of the committee is -- I mean, as long as the CPR workgroup report has adequate time to be fleshed out, hopefully voted upon and all that, I am comfortable with that. I just don't have a clue what else is on the agenda and all that at this point.
DR. LUMPKIN: Well, I would think that that would be the priority on that agenda for us to through. I think that given that schedule, you know, we will have the symposium so we probably can cut back on presentations. We may not be able to have a panel at the full meeting. But I think we can work around that.
MS. GREENBERG: I guess if people have the three days on their calendar, we would ask them to at least hold them for awhile until we can sort of form the agenda and see what subcommittees want to meet.
DR. LUMPKIN: Okay. Any questions about the symposium?
DR. FRIEDMAN: It might be nice to build in a break somewhere in the afternoon. I think you have it non-stop for four and a half hours.
MS. GREENBERG: We are planning to invite, you know, former members, current and former -- obviously, current members and staff, former members and staff, members of the Data Council. We talked with John about, you know, various -- representatives of various organizations that we have worked with through everything from, you know, public health issues to HIPAA issues, et cetera, but if you have thoughts about particular people or organizations that you would like us to send an invitation to, if you could e-mail that to Debby Jackson, which is -- or to me if you have my e-mail and I will -- we are right next to each other, but then that would be helpful because we are starting to work on the invitation, which will be joint signed by someone from the National Academy and John.
MS. FYFFE: Is Donna Shalala going to be speaking or --
DR. LUMPKIN: We will see what we can do.
MS. GREENBERG: We talked about trying to get her.
DR. LUMPKIN: We might try to get her. With the Secretary it is normally just a scheduling problem, if she is available.
There obviously will be limited spots available. So, what we will do is send out invitations and then also there will be availability for the public to attend. So, we are going to have to have people secure slots and invitations well in advance of the meeting. We think this might be the hottest ticket in town.
Okay. Anything else that I forgot?
Okay. The report is in the process of -- to be handed out at the meeting.
Okay. NHII Workgroup, taking to heart many of the comments. We dramatically changed the outline and we followed a sketch that Dan has suggested. So, what we are going to be working on and trying to move this into is opening up with three or four future scenarios, kind of how things happen today versus how they happen in the future, that the second part -- there are five parts to the outline -- the second part would be the discussion of impediments, why we can't do this now.
The third part would be then what we need to do to overcome these barriers or these impediments. The fourth would be an introduction of the role of the NCVHS and NHII and then the fifth part being a discussion of the NHII. We are looking for about 20 pages. In addition, we will then have a series of appendices that would go into some depth on some of the issues.
For instance, we would include the paper that was prepared on privacy by the NCVHS. We would include some of the documents we use for HIPAA, such as there is a WEDI(?) document on what EDI is. We would use the paper that is being prepared by a CPR workgroup. All of these would be considered as appendices and would be added to the final document.
One of the issues that we had some discussion on and we felt that it was very important that we included discussions on privacy throughout the document, including as part of the scenarios, ,as well as other discussions, rather than looking to have a particular section at the end on privacy.
That is what we have. There are a couple of things for which those who are not on the workgroup should feel certainly welcome to participate. We are going to be collecting some potential scenarios. So, if you have any potential scenarios that may come to mind -- and this is not writing out a whole scenario. This may be two or three sentences of a scenario that you think may illustrate. We are taking it from the community, the personal health and the caregiver dimensions, but recognizing that they are going to illustrate the integrated nation of the system.
If you have any of those, please send them. If you think that there are any impediments that you would like to contribute to our list of impediments that we are going to be creating, also feel free to e-mail those to me anytime over the next two weeks.
If there any individuals, any members of the committee, who would like to -- can't make it to the committee meetings, but would like to continue to be involved, there is a list serve that is set up and please let Marjorie know and we can make sure that you get on the list serve.
Dan, Jeff, Barbara, any additions?
DR. STARFIELD: No. I think we are not only inviting, we are encouraging scenarios from anybody. If you could contribute something, it would be very helpful.
DR. LUMPKIN: Standards and Security Committee.
DR. COHN: Good afternoon.
I think, as you remember, we spent awhile yesterday morning talking about the annual report on HIPAA implementation, referencing -- this is Tab 7 in your book -- based on, I think, all of your input plus the discussions by the subcommittee in the afternoon. We have drafted a paragraph for your consideration to add to the report.
We would see this paragraph being in the executive summary, probably under -- it says "Progress in HIPAA Health Data and Security Standards." It will likely be a second paragraph and then in the body of the text, the body of the report, it would actually be the first paragraph of that same section.
Would you like me to read the paragraph?
DR. LUMPKIN: Please do.
DR. COHN: It says, "To date, the HIPAA process has generated five NPRMs with the first issued in 1998. The NCVHS is concerned that almost two years have passed without the issuance of a final rule. Our concern reflects a growing industry concern about a loss of momentum. Further delay in issuing final rules for standards raises the possibility that industry resources committed to implementing the HIPAA standards may be diverted to other priorities.
"Moreover, delays in implementation translate into the loss of economic benefits from administrative simplification. Consequently, the NCVHS urges the Secretary to commit adequate resources to ensure the timely publication of these final rules."
Anyway, this is basically what we thought was a reasonable paragraph, but we want to bring it --
DR. LUMPKIN: Can I just sort of maybe add
in --
DR. COHN: Please.
DR. LUMPKIN: The NCVHS urges the Secretary to continue to commit --
MR. GELLMAN: The question that this raises is are the delays the result of a lack of resources. Jim, is that a problem? Do you have the resources you need to do this?
DR. LUMPKIN: It sounds like me on the radio.
MR. SCANLON: Let me answer what any agency head or director would say. We can always use additional resources, but I don't think resources are the main issue for at least the first of the standards.
MR. GELLMAN: It seems to me that the tone of this thing is that the tone of this thing is that the Department has delayed this. That is what I said yesterday -- I don't see that. I think the Department has gone at this in a fair and honorable way in trying to get it through and there are delays. The amount of time that the statute contemplated for this was not enough. There are plenty of problems in getting these things cleared outside the Department as well. OMB, I don't know how much time you killed trying to get these things cleared through other agencies and OMB.
I mean, I don't have any problem, you know, saying -- you know, the point about there is growing industry concern. I certainly accept Kepa's comment on that and, you know, I don't see any reason to say anything critical of the Department's delays here. I think there are delays. They are natural and I think that we may want to -- this committee and the Secretary may want to reaffirm the importance of HIPAA and the need for this to progress, but I don't see the need for any criticism.
DR. COHN: So, you are really talking about the last sentence? I think what you are saying is that everything is basically -- that we are urging, obviously, it moving forward and I think the issue is is exactly how we finished the paragraph.
MR. GELLMAN: Well, the second sentence, you know, concerned the two. There is a pass without the issuance of a final rule. I mean, that suggests that there has been some willful failure to issue a final rule and that is what hasn't happened, at least in my perception, that we haven't had a final rule, but it is not anybody's fault that we haven't had a final rule.
DR. LUMPKIN: The complexities of the process
-- the NCS is concerned that the complexities of the process have resulted in two years passing without the issue of filing a rule. Consequently, and then you go to the last sentence, "The NCVHS reconfirms its commitment to the process and urges the Secretary to communicate her continued commitment to the process.
MR. GELLMAN: Yes, something like that.
DR. COHN: Did you get that? I didn't get that. Say the last sentence again, John. I am sort
of --
DR. LUMPKIN: Well, consequently -- well, I think we could just say the NCVHS reconfirms its commitment to the HIPAA process and urges the Secretary to communicate her continued commitment.
MR. SCANLON: Well, why don't we just say to ensure that -- I mean, her commitment, I think, again, she is committed, obviously, as well. I think the focus is that you want to have the final issue promptly and urges the Secretary to --
DR. LUMPKIN: Well, see, but I think what I heard Bob saying, what I agree on, you know, you look at the process as we have seen it and it is hard -- you are hard-pressed to say you should have done this, you should have done that. But the issue that Kepa raises is there is not much that is heard, you know, because all of this process is internal.
What really is needed is some sort of statement saying, yes, we continue to support the process and we will get it done as soon as we can.
MR. GELLMAN: How about if we send Bill Braithwaite to more conferences.
DR. ZUBELDIA: I think Bill has had a tremendously positive impact with all the conference he goes to, but he is losing credibility, too, because every date he says this will be ready by such and such a date. Every date has gone past.
DR. COHN: Let's not point fingers anywhere that we can avoid doing that. Maybe with this last sentence -- I mean, once again, we are -- actually we are beyond wordsmithing. We are actually trying to figure out the right conceptualization here.
But what I have been talking about is basically the NCVHS urges the Secretary to take steps to ensure the timely publication of these final rules as opposed to commit resources or anything else. Is that a -- and she can use whatever discretion --
DR. ZUBELDIA: I would use the word "to expedite."
DR. COHN: What?
DR. ZUBELDIA: I would use the word "to expedite the publication."
DR. COHN: Take steps to expedite?
DR. ZUBELDIA: Expedite the publication of final rules.
Let's say that they are timely. Let's see if she can do something to make them quicker than that.
DR. COHN: So, to expedite -- now, do we say expedite the timely publication --
[Multiple discussions]
Okay. Expedite the publication. Okay. I think what we are also saying is in the second sentence then that we are putting that sort of parenthetical about because of the complexities of the process, that there is obviously, a reason why all this has happened and we don't think that it is bad will on anyone's part. But on the other hand, we do want to see this happen. I think we are concerned that it does need to move forward in an expeditious fashion.
DR. STARFIELD: I don't think we need that second sentence actually. I don't think it adds anything.
DR. COHN: Well, it seems a little odd if we only have the third sentence.
MR. BLAIR: Maybe it is just simply a paragraph that other folks within the Department need to get a little bit more attention, get a little bit more support, to wind up saying NCVHS is urging -- you know, is saying that there is a concern and saying we really need to give attention to this so that when some other choices of the division of resources between one thing and another come into place, there is a little bit of ammunition here to keep the resources in place to get those rules out.
MS. FYFFE: It needs to be a very high priority.
DR. COHN: Yes. I think that is what we are saying here.
MS. FYFFE: Can't we just say that, though?
DR. COHN: That would be placed in the last sentence that says, basically, the NCVHS urges the Secretary to continue to --
DR. STARFIELD: To take steps to --
DR. COHN: To take steps --
MR. SCANLON: To expedite the process.
DR. COHN: That is right. Take steps to expedite the publication of these final rules and we could say and make it a high priority. That is a little redundant somehow.
MR. BLAIR: I think the word "expedite" tends to get people's attention.
DR. COHN: Yes. Anyway, I think with those changes, at least I am -- it sounds like that we made an improvement so far in what we were talking about. Actually,, as I look at Jim Scanlon, did you catch the changes in the second sentence?
MR. SCANLON: In the last sentence I did. I think I have it all. Should I try to read what it sounds like --
DR. COHN: Please, including the second sentence, yes.
MR. SCANLON: I was actually adding a little bit more. The first sentence would read: To date, the HIPAA process has been extremely open, collaborative and productive and has generated five NPRMs, including privacy standards, with the first issued in 1998. Because of the understandable complexities of the process, almost two years have passed without the issuance of a final rule.
Then our concern reflects the growing industry concern about a loss of momentum, further delay in issuing final rules for standards raises the possibility that industry resources committed to implementing the HIPAA may be diverted to other priorities. Moreover, delays in implementation translated to the loss of economic benefits for administrative simplification.
Consequently, the NCVHS reaffirms the importance of the administrative simplification initiative, something to that effect, and urges the Secretary to expedite, to take all steps to expedite the publication of these final rules.
DR. COHN: I think the second sentence needs to start out with the NCVHS's concerns. So, I think we would continue that. Is that -- everyone comfortable with this? Now, having said that, Jim sounded like he was sort of making it up as he was going along.
[Multiple discussions]
I don't mind anybody making it up. What I do mind is that we won't remember it after we vote on it.
DR. LUMPKIN: We have a motion from Simon to adopt the annual report with the addition of the previously read paragraph in the section on "Progress of HIPAA Health Data and Security Standards."
PARTICIPANT: I second the motion.
DR. LUMPKIN: The motion has been seconded.
DR. COHN: I would ask the leave of the committee also that we are able to make some wordsmithing changes to the executive summary to tighten it up a little bit. I presume you are all comfortable with that.
Okay.
DR. LUMPKIN: It has been moved and seconded that this report be adopted with leave to the committee chair and the subcommittee chair to review any wordsmithing changes and that it be submitted to Congress. Oh, and what is a reasonable deadline for additional comments? Two weeks? Okay.
If any members choose to submit additional comments, as we have now established as part of all of our reports, they will be included if they are submitted within two weeks.
DR. STARFIELD: Is it understood that as a courtesy we will send this to the Secretary? We talked about this yesterday, about not sending something to Congress without --
DR. LUMPKIN: Yes. I think this paragraph is benign enough that we can pre-send it to the Secretary.
Jeff, you had your hand up.
MR. BLAIR: Oh, my hand was to vote in the affirmative. I thought we were voting.
MR. SCANLON: I might add that as part of the subcommittee's deliberation yesterday, this would not be the only communication about the urgency of issuing final standards. I think we would ask John to raise the issue of urgency at the March meeting of the Data Council and then in addition we have asked Bill and Karen Trudell, who co-chaired the overall Health Data Standards Committee to convey this. There is this concern about urgency and diversion of resources. So, this won't be the only means of conveying the impression.
MS. FRAWLEY: John, I was just going to suggest in our letter to the Secretary that maybe we ought to take some of these points and put them in the cover letter so that it is not buried in the report, you know, that we are concerned about the --
DR. LUMPKIN: You mean, the transmittal letter. We are, obviously, going to send a transmittal letter, which will be fairly -- which will be almost contentless to Congress, just kind of saying here it is and you are suggesting then that we would have a cover letter from the Secretary emphasizing this particular point, to the Secretary.
MS. FRAWLEY: Correct.
DR. LUMPKIN: Okay? Everybody understand what we are voting on?
All those in favor signify by saying "aye."
[There was a chorus of "ayes."]
Those opposed say "nay."
[There was no response.]
Abstentions?
[There was no response.]
It carries.
Anything else, Simon?
DR. COHN: Well, yes, I actually do want to just -- we have one other possible work item that may occur between now and the next full meeting of the NCVHS. Now, while I am hesitant to make any comments of the possibility that we may see anymore NPRMs in the next -- between now and the next meeting, I think that there is a hope that we will see a claims attachment, notice of proposed rule, in that time frame. I just want to -- once again, since we are -- obviously, this is something we know -- we believe is likely to come out, the expectation will be is that the subcommittee in its meeting either at the end of March or in early June will develop comments and a draft letter to the -- in response to the NPRM.
Depending on the timing of that, it may require us to have a conference call for discussion and acceptance of that. So, I just want to alert people that that is something that may be coming up between now and then.
DR. LUMPKIN: So, you are moving that we would follow -- establish a process whereby if the NPRM is issued, the subcommittee would meet by phone and would then make a recommendation to the full committee on a phone conference call to approve the NPRM?
DR. COHN: That sounds as good as anything. I guess my hope was is the subcommittee might actually be able to meet in person, but the full committee would have to meet by phone.
DR. LUMPKIN: Okay. In person.
DR. COHN: Is everyone comfortable with that as a process for this?
DR. LUMPKIN: Okay. So, we have a motion to adopt a process to approve response to the NPRM should it be issued between now and the next meeting. Is there a second?
MS. FYFFE: Second.
DR. LUMPKIN: Thank you. It has been moved and seconded. Is there further discussion on the motion?
All those in favor signify by saying "aye."
[There was a chorus of "ayes."]
Those opposed, say "nay."
[There was no response.]
Abstentions?
[There was no response.]
Okay. Carries.
Anything else, Simon?
DR. COHN: I think that is it. Actually, Jeff, do you have a --
DR. LUMPKIN: Anything from the workgroup?
MR. BLAIR: No. I think we covered everything yesterday in the meeting.
DR. LUMPKIN: Thank you, Simon.
Privacy and confidentiality. I would like to just take this opportunity to thank you and the committee. I know it was a lot of work over a short period of time and I think the final result was certainly -- reflects positively upon the full committee. So, thank you.
MS. FRAWLEY: Thank you.
The Subcommittee on Privacy and Confidentiality is meeting all day tomorrow and we will be having a hearing on the Internet use for health information and privacy implications. We are very fortunate, Gail Worlick(?), the staff to the subcommittee, put together some very fine panels. There is also an entire briefing book of privacy policies for Internet sites.
Many of you may not be aware of the fact that the Federal Trade Commission is currently investigating some of these sites for some of their privacy practices and there has been a study done by the California Health Care Foundation that has also revealed some concerns about privacy policies.
So, tomorrow, we will be hearing from Jan Laurie(?) Goldman, who actually did the study for the California Health Care Foundation and she is from the Institute of Health Care Research at Georgetown University. We have Calvin Weiss, who is the CEO of Health Magic coming. Sam Shugar(?), who is the president of ForHealthyLife.com, Thomas Booth, who is from PersonalMD.com, Tanya Glazenburke(?), who is from the Medic Alert Foundation. That is our morning panel.
Our afternoon panel will be Richard Smith, who is a consultant, Sam Carp(?), who is from the California Health Care Foundation, Robert Musochio(?) from the American Medical Association, Christine Varney, who is a partner at Hogan & Hartson, formerly from the Federal Trade Commission and John Mack, who is president of the Internet Health Care Coalition.
So, we expect to have a very interesting day tomorrow and certainly are pleased that we have a number of very interesting speakers.
DR. FRIEDMAN: Kathleen, would it be possible for members of the committee to get a copy of the briefing book and any other materials that are presented tomorrow?
MS. FRAWLEY: Yes.
DR. FRIEDMAN: I can't speak for anybody else. I certainly would really like to get --
MS. FRAWLEY: If you let -- anyone let Gracie know that they want a copy, she will be happy to get it to you.
DR. LUMPKIN: Why don't we just send it out to the full committee.
MS. FRAWLEY: Gracie, full committee.
DR. LUMPKIN: If you are not interested, raise your hand. Obviously, this is a very hot topic.
MR. BLAIR: Incidentally, that briefing book was just done beautifully. It may have taken awhile to read through, but it was very informative.
DR. LUMPKIN: Is this going to be on the Internet?
MS. FRAWLEY: Yes.
DR. LUMPKIN: So, for those of you -- I don't know if you have ever tried to -- no, it is private. So, it is not going to be on the Internet. For those of you haven't checked it out, if you can't listen to it when it is being simulcast, it is on the VA Web site. So, it is available to be listened to at your leisure for those of you who may want to get access to it and aren't available tomorrow during the day.
MS. FRAWLEY: That is it.
DR. LUMPKIN: Thank you.
Subcommittee on Populations, Liza is off at her other meeting, trying to be at two places at once.
DR. NEWACHECK: We are going to do -- we have a presentation on functional status, which is a new initiative that we are going to do and then I had asked Kathy to give us an update on quality and Dan to give us an update on the 21st Century vision.
DR. LUMPKIN: Okay.
The new project on functional status is just really underway and, basically, what it is about is assessing the feasibility and utility of adding a measure of functional status to administrative records, whether they are claims forms or claims attachments or encounter records of some kind or another.
The idea that having a measure of functional status on administrative records could be useful for a variety of purposes, including, for example, setting payment levels. We are doing risk adjustment in managed care, could be useful for quality improvement efforts as a measure of outcome. It could also be useful for epidemiological and statistical purposes in terms of providing a population-based measure of health status.
There are a whole variety of questions, though, that we are thinking about and trying to address them at the subcommittee. These include how would functional status be measured. There are a variety of different measures available of functional status, that all of them have limitations of one kind or another.
Who would actually make the functional status assessment? Would it be done by a provider or caretaker? Would it be done by a patient themselves filling out a form by self-report? Another question concerns how much would it cost to do this? Even if it only took a couple of minutes to do, multiply that out by a billion physician visits a year or whatever and you start to get into some real money.
There are also concerns about how accurate it would be. Functional status is inherently a very subjective concept. As we heard this morning from Mark Hayward, it is also a very changing concept. It is not constant and people's functional status changes from month to month, year to year, even day to day. So, there are some issues there, too.
Then finally the question of how would the information actually be used and what could we do to prevent abuse of that information. So, those are all questions that the subcommittee is thinking about and we have only had one hearing on this and it was last month and it was cut short. So, we didn't even get through our first whole panel of presentations and such.
So, our plans now really are to focus on this over the next six to nine months. We have at least two subcommittee meetings scheduled to talk about this and we are intending to pursue two different approaches. The first one is to just think broadly and to try to get broad input on the issue of feasibility and utility of having a functional status measure on the administrative record. We would like to hear more from providers, from consumers, from policy makers, from researchers about their perspectives on this issue. We started that process last month, but we certainly haven't completed it.
The second is to conduct a more in-depth assessment of one particular mechanism for categorizing or classifying functional status. That is the ICIDH. And we talked a little bit about that this morning. That used to be at the international classification of disability and handicaps. Now, it is called the International Classification of Functioning and Disability, even though they have kept the old name, the ICIDH.
This is basically measured as being pushed pretty hard now by the World Health Organization as a means or a companion to the traditional ICD or International Classification of Diseases and moving beyond just disease classification to classifying function and disability. There are about, I think 40 countries around the world including the U.S. that are participating in the development of the ICIDH.
Later this summer, the National Center for Health Statistics is going to be involved in a beta test of the measure and we are hoping that Gerry Hendershot, who is staff to the committee, and Paul Placek(?) will be able to give us some early results and early results on that after that beta test is completed this summer.
So, basically what we are doing is pursuing these two strategies, one to get more information from concerned parties about the utility and feasibility of adding this measure or a measure of functional status to administrative records. The second is to learn more about and to explore more about the idea of using the ICIDH for this purpose.
So, that is where we stand now.
Any questions?
DR. COHN: Paul, I will apologize because I actually had initially intended to be at the first set of hearings.
I guess just for my own curiosity, I am -- I think this is an important area, but I am struggling to still figure out sort of what it is. So, I am curious (a) if you have come up with a definition that you are all using about all of this and (b) the question about -- since we just listened for two hours to summary health measures, both -- I mean, is this another piece that is being investigated at the same time because they seem very, very close in concept?
DR. NEWACHECK: There certainly is some overlap here. We didn't think about it in that context, though, but there certainly is some overlap. We don't really have a defined approach to thinking about emotional status at this point. I think we are willing to keep an open mind and try to think about various different ways of defining the concept of measuring it.
That is really one of the purposes of having these hearings is to get ideas from people in the field and to bounce ideas off from the field. It is kind of an open book at this point and it is very -- I think this is a very embryonic point in this project. So, if there are ideas that people have or if people want to come to the upcoming presentation hearings and that, we would be delighted to have you all come.
MR. SCANLON: Paul, is the -- I remember one of the earlier thoughts of the subcommittee was just to support a commissioned paper or literature review on the current state of measurement, the art in functional status. Is that still on the agenda of the subcommittee or --
DR. NEWACHECK: I think it is still on the agenda. We have not -- we are thinking about working with Les Ted(?) as a contractor on that, but it hasn't been finalized by any means.
So, shall we move on to Kathy.
MS. COLTIN: Well, as you may remember, the Workgroup on Quality has been trying as much as possible to accomplish its goals within the meetings and activities of a number of the other subcommittees and the full committee. So, we have been pursuing our goal of trying to identify the data needs and gaps for measuring and improving quality of care.
In doing that, we have tried to incorporate and be sensitive to the interests of the full committee by structuring a lot of our panels and presentations to occur within the full committee meetings. So, we have had two of those so far and you are going to have another one later today where we are going to talk about the kinds of data that are needed to measure and improve patient safety concerns.
So, that has been our main approach. The other has been to work with the Subcommittee on Populations on this functional status initiative, primarily from the standpoint of saying what kinds of functional status information would be valuable for assessing and improving quality of care because as Paul pointed out, there are a number of different purposes to which measures of functional status could be put; risk adjustment, being related to both payment and to quality measurement, but other purposes as well and our purpose is really saying, well, how would it be helpful for advancing improvements in quality.
So, we have been trying as much as possible not to hold separate meetings. However, we did hold a meeting in November after the last full committee meeting, where we heard about some of the perspectives of those who are heavily involved in trying to do performance measurement, particularly performance measurement in the managed care arena, but also comparing performance in the fee for service and managed care systems as HCFA is trying to do and hearing from those who are engaged in that process about their perceptions of the limitations of existing data in trying to do that.
So, we will probably be trying to continue to work through the full committee through these presentations to accomplish our goals and through the subcommittee initiative on functional assessment.
DR. LUMPKIN: Any questions?
Okay. Dan.
DR. FRIEDMAN: Since the November committee meeting, the Workgroup on 21st Century Vision for Health Statistics has been focusing on five or six different things and before I do that, let me just emphasize yet again that this is very much a joint process with the committee, namely, you know, Barbara and Paul, who have been very active with the Data Council and Jim has been active certainly with NCHS, where the staff involvement, everywhere from Ed to Marjorie, Ed Hunter, Lisa Bregman(?), Rob Winezimmer(?), Gerry Hendershot have really spent a lot of time and continue to spend a lot of time on it. So, it is sort of drawing the lines between what is the committee, what is NCHS, what is the Data Council, is really impossible on it.
Having said that, there are five or six things that we have been engaged in. First of all, immediately after the last committee meeting, we had the National Academy of Sciences Committee on National Statistics Workshop and several of you were there. I thought it was
very good and the speakers were very good. The discussion was very good, very stimulating and very much looking forward to receiving a draft of the workshop report since that.
Second, in mid-November, we had our initial and our pilot local discussion group. These are essentially parallel discussion groups to the national ones that we had last spring, focused on -- it occurred in Harrisburg, Pennsylvania and it focused on local and state perceptions of health trends, health information needs associated with those trends, health information priorities and then appropriate roles for state and local and the Federal Government and private versus public.
That was, again, I thought it was an extremely productive session. I thought we heard some new things, some issues very well articulated and in March we have a couple more of those local discussion groups, one in Albuquerque and the other one in New Orleans. We tried to schedule those, tried to select communities for those, and this was at John's suggestion, where in states that in turning point grants, so they had already done a lot of thinking about health population and health information needs.
Third, a fuller Web site, including the work group materials and the commissioned papers and so forth should be ready any day. NCHS has really pulled that together. We have contracted with Susan Canann(?) to develop either an interim draft report or a draft interim report or an interim report, depending upon how we feel at the moment.
The governance on that, considering it is a joint process will be interesting, but we have essentially a rough outline for that report, which is going to follow the structure that we have used for the discussion groups and that report is very much going to draw up and reflect all of the discussion group summaries that have been developed and that will be on the Web site and will also draw upon the commission papers, the Sinstat(?) workshop and so forth.
The 2010 conference in late January, Ed Sondik chaired a panel that included Myron Straf from the National Academy, who presented a summary of a discussion at the Sinstat workshop. Barbara Starfield gave a presentation on -- I don't want to characterization, but gave a presentation and I gave a presentation and our presentations reflected less the process and more our own personal senses of what is important in developing a 21st Century vision.
Then in addition to that, we submitted to a publisher either a draft interim or interim draft prospectus for the edited volume that we have been talking about. One of the things that has not been launched yet, but we are talking about starting in July with a series of NCVHS public hearings held not in Washington, emphasize not in Washington or only one of them in Washington, perhaps starting in Chicago in July, which would seek feedback on the interim report and try to start developing some of the recommendations.
The interim report is basically going to be here is what people have said about health information needs, health information priorities, not going to contain anything reminiscent of recommendations but we would like the public hearings just to both focus on reactions to the interim report, but more important than that help us move towards the vision itself and help us move towards recommendations. We haven't fixed on the number of public hearings yet.
We have talked about three or four with one of them being around here and the other three being in different parts of the country.
DR. LUMPKIN: The only concern I have about the meeting in Chicago is do you really want to kill this project. The last time we started out in Chicago --
[Laughter.]
DR. FRIEDMAN: We have discussed that. That was definitely one of the first things that --
PARTICIPANT: [Comment off microphone]
DR. FRIEDMAN: Actually I did want to mention that we might want to do a tag team process because I think that that would be a good opportunity for us to have many of the same people who would be interested in the 21st Century and NHII. So, I think that is an excellent idea.
Actually, perhaps we should discuss further where we start.
MS. GREENBERG: Are you superstitious?
DR. FRIEDMAN: No, but perhaps really -- we may want to discuss further where we start.
PARTICIPANT: [Comment off microphone]
DR. LUMPKIN: I think we are done with that.
Okay. Anything else on the 21st Century?
Future agendas? Maybe if we can --
DR. COHN: Do we get to fill in some items for June now?
DR. LUMPKIN: Yes. This is where you get to put in your placeholders for June.
DR. COHN: Can I put in two suggestions then?
One is is that we had been thinking about the possibility of a panel at the June meeting relating to coding systems and the issues surrounding possible future changes to, for example, ICD moving from one version to the next and what sort of industry needs are related to that, what timing should be occurring, et cetera.
I don't if we acquire hearings on that, but I think a panel to begin to explore the issues might be very appropriate in a June time frame. Anyway, that was one thought.
The other piece that came out of the standards and security session had to do with the idea of a briefing related to issues related to the public key infrastructure from a national basis to get a view from government, recognizing that the digital signature standard is being delayed because there appears not to be sort of ready for prime time, but to understand the issues and opportunities from a national view and what ought to be done in this area.
Am I articulating that well enough for subcommittee people? Anyway, that was the other piece. We didn't necessarily think there was a panel for that but maybe a briefing.
Kepa, you had a --
DR. ZUBELDIA: We also talked about having a panel from early implementers.
DR. COHN: Oh, that is a June hearing -- July hearings. Those are completely separate.
DR. LUMPKIN: Okay. So, that is a briefing on electronic signature?
DR. COHN: Yes. Digital signature.
DR. LUMPKIN: Digital signature. Okay. And a short discussion on CPR?
DR. COHN: About a half an hour and we will go from there.
PARTICIPANT: Do we have those dates in July?
DR. COHN: What are you talking about, for the hearing?
PARTICIPANT: Yes.
DR. COHN: Okay. Now, we have been talking about the meeting dates for the next NCVHS meeting. For July we have hearings for the Subcommittee on Standards and Security, which are currently scheduled for --
[Multiple discussions]
MS. GREENBERG: Simon, what did you say was the main focus of those?
DR. COHN: Those have to do with the early implementers, issues related to HIPAA.
MR. BLAIR: Let me just clarify. Did I get -- when you were talking about the hearing for codes and hearing for electronic signatures, were you saying for July or for June?
DR. COHN: No. I was actually referencing a panel at the next NCVHS meeting on codes and any briefing at the next NCVHS meeting both in June on digital signature. The digital signature was more of a briefing.
DR. LUMPKIN: Okay. Anything from population, quality, anything for the June agenda?
PARTICIPANT: Not yet.
DR. LUMPKIN: Dan and the workgroup -- both workgroups are scheduled for the symposium. Anything on privacy for June meeting? No new NPRMs?
MS. FYFFE: Are we concerned about the privacy provision section, I think it is 501 of the Graham-Leach-Bliley Act, because that covers financial institutions -- yes. The proposed rules for that just came out two days ago and comments are due March 31st. The issue is --
DR. LUMPKIN: I am not concerned because I know nothing about it. Maybe I should be, but --
MS. FYFFE: The issue is -- well, there are a few issues. You have got the Department of Treasury, the FDIC, the Federal Reserve and other financially oriented agencies of the Federal Government promulgating regulations related to non-public personal information for financial institutions and some of those financial institutions are, in fact, organizations that indemnify against medical illness, which are insurance companies.
There might be a concern of how the proposed rules for privacy of medical information that came out from HHS may or may not overlap with the privacy provisions of Graham-Leach-Bliley. Does that bother anybody?
DR. COHN: Let me just ask for my own edification. Is this something that we need to potentially write a letter than then needs -- from the committee and should be discussed tomorrow at the Subcommittee on Privacy and Confidentiality or is this
a --
DR. LUMPKIN: Well, we can add that to the agenda, I think, of the privacy committee. I think they need to discuss that before we --
Barbara.
DR. STARFIELD: At the end of the panel this morning, Ed Sondik urged this committee to undertake a discussion of what we heard today, in particular the fact that the three major agencies in HHS are going completely separate ways and we have some input into that.
DR. LUMPKIN: It sounds like that is something that needs to be discussed, but perhaps we could get that preprocessed by the Population Subcommittee or at
least --
DR. STARFIELD: I think that the committee itself probably ought to offer having heard the whole thing, but that is sort of my feeling. You might want to get a reaction from that.
DR. LUMPKIN: Okay. Is there reactions -- are there -- how do we want to structure that discussion? Do we want to bring in additional panelists? Do we just want to have -- I am looking for how we can develop a proposal or something to act on.
DR. STARFIELD: Yes, well, I am think that, you know, that everybody around the table -- that is a question -- I mean, I think everybody was involved in it. So, it seemed to me that what we need to do is not bring in more people but discuss our reactions to what we heard and then on the basis of that make some discussions about what we want to do as a committee or subcommittees.
DR. LUMPKIN: Okay.
DR. FRIEDMAN: I think that is a very good idea. I agree with Barbara. I do think that there are things that the committee could do that could be potentially useful.
MS. COLTIN: I would support that but I also would recommend that we invite a representative from each of the agencies, who is involved in what they are doing. I mean, some of the people were here today but there was no one from HCFA, for instance. And have them available to participate in this discussion because I think we will have questions and not to have them come and make formal presentations, but rather to be here as resources for us as we engage in the discussion.
DR. NEWACHECK: Are we talking about the notion of having a set of leading indicators and what those indicators should be or are we talking more about the specific methodologies that were discussed today, like the Dalleys and things like that? In your mind, Barbara, what are you thinking about?
DR. STARFIELD: I think they are related. I think we could do both. I mean, it is not only the methodology but it is the uses.
MR. SCANLON: Other ideas for the June meeting or --
DR. LUMPKIN: No. Actually, I would like to follow-up on this a little bit more because I am still having some trouble. I identify the problem. I am just trying to figure out how that session of the meeting would go. Would you see just throwing it open to the floor for discussion? Do we want to take a few minutes now to spend a little bit of time trying to define the issues that we want to address? Because I think we function better in a full committee when we have a proposal to adopt or motions to make and we haven't had a lot of just, okay, now we are going to discuss this issue.
DR. STARFIELD: I think it would be very good for us to discuss it now, but I think we should finish our business and then use the whole remaining time to 3 o'clock to --
DR. LUMPKIN: Okay. Well, let's --
DR. STARFIELD: Are there other things we have to do?
DR. LUMPKIN: No, no, no. That was really what I was going to get to is that we do have time between now and 3 o'clock. We can finish up the additional items and then move into that.
Jim.
MR. SCANLON: A couple of other items. We probably may want to have a briefing at the June meeting of the National Research Council report on the -- the report that was released yesterday, probably a full committee briefing from Ted Shortlif(?) or other panel members at the Academy to come over and brief us on what they -- much of what they recommended and looked at in that report fits in very well with the overall HII that the committee is talking about.
Secondly, in the privacy research area, we with AHRQ have sponsored a study at the Institute of Medicine on -- I am sorry -- HHS is sponsoring an expert panel study at the Institute of Medicine to look at best practices or just practices in IRBs behavior in looking at records-based research. John, I can't remember now whether we would expect findings by June, but I don't think the full committee has been briefed on the overall report. Maybe by June, the end of June, the Academy would have some preliminary findings to report.
MR. MANNING: Their conference is in March, March 13th and 14th. I can't give an estimate of whether there will be anything by June.
DR. LUMPKIN: Okay. A couple other items and then we will go into our discussion of this morning's panel.
The first item really has to do with our meeting on the 20th and 21st and which we all have also are holding the 22nd. If we try to do all that we have listed on here, we will need to do the full three -- two and a half days plus the symposium. If that is agreeable, then we will just plan on it being a three day block. Okay?
The last item before we do this -- and I have had just a brief discussion with Barbara and Dan on this one, just toss it out, I was going to save this for the Executive Committee but I think we have a few minutes to discuss this. This has to do with our inter-relationship with the NCHS.
The committee has gone through a transformation and the transformation that most of us who are the class of 1996 are not really aware of but there was a major transformation as the committee was re-geared up to take of HIPAA responsibilities. Prior to that, the committee had a very full agenda and came up with some important documents, such as the core data set and other documents, but they also had adequate time and expertise to play a key advisory to the National Center for Health Statistics.
The new reconstituted committee, I don't think we can adequately meet that charge and it is just a discussion that I have had with Ed Sondik. To that end, we have had some discussions -- I have had some discussions and I suggested to him that perhaps they should explore having an advisory committee that could advise them on issues related to survey consolidation, survey coordination, developing state level data.
I would also like to recommend to him that we would propose to the Department that the chair of that committee would be automatically a member of NCVHS, which -- you know, since we have open slots that could naturally occur. I just wanted to toss that to the full committee because if that is agreeable then should he require any support, then it would be, I would believe, my role as the chair to write an appropriate letter supporting that that is something that we don't believe that we can adequately do, but it is important for the two committees to interact so that there is not -- so that there is consistency between the advice that is being given.
Discussion?
MS. FYFFE: What do you mean by open slots?
DR. LUMPKIN: Well, we have turnover in the committee.
MS. FYFFE: Are you talking about the slots for the full time Federal Government employees or are you talking about the slots that are the appointed positions?
DR. LUMPKIN: Yes, the appointed positions.
So, rather than -- well, for instance, we always -- there are people who are on the committee when someone leaves, for instance, when -- well, a position like David's position or a position -- generally, we have an epidemiologist on. Those kind of positions would be -- we would look to fill with the chair of that committee.
MS. FYFFE: I guess my more fundamental question is this committee has 18 appointed folks. Would one of those now go to an appointment to someone with the Federal Government?
DR. LUMPKIN: No, no, no. This would be an external advisory committee to the NCHS that would have a chair who would be appointed by either the Secretary or the director of NCHS or the director of CDC. So, it would be an advisory committee person.
MR. SCANLON: Maybe I can clarify. And one and a member of this committee, one of our 18, would also be asked to serve as the chair of that committee. So, they would be occupying two. We don't lose any seats here, but we do ask one of the members to play a double.
MS. FYFFE: Can we start over with your -- give me some constants here. What does this have to do with NCVHS?
MR. SCANLON: That was our job. Technically, it is still our job but we are incapable of performing it.
DR. COHN: I guess I have a more fundamental question about that because I think maybe I am -- can I try a question or two, John? I mean, to me the question is not the -- if we are going to add another member or subtract a member or whatever to the committee. It is more the fundamental issue of are we completely incapable of advising NCHS, which is I think the assertion you started out with.
And I don't think you quite mean that.
[Multiple discussions]
I guess what I find as I reflected on your comment, I said, well, gee, I personally do not feel capable nor is my skill set such that I can do some of the things that NCHS needs. For example, your comment about survey data and getting down to the nits of how you consolidate some of their survey pieces.
On the other hand, there are major areas that NCHS is involved with, including code sets that I think is very much part and parcel, a major concern of this committee and I would hate to see that swept away along with the nits that you are describing. I don't really mean nits, but I mean very specific issues. So, I think it gets down to really a question of what a charter might look like of a group handling what areas you feel that are no longer our areas of expertise.
DR. NEWACHECK: John, I just wanted to add I agree with Simon. I think we ought to very carefully consider whether or not the committee can provide this advisory role. We haven't done it but that is partly because we -- we have to do the HIPAA projects, obviously, but we have a certain degree of flexibility in choosing other projects and other activities and maybe this should be given priority as a committee activity.
DR. STARFIELD: I guess there are two of us here who were here before HIPAA, Kathy Coltin and I, and I can assure you that this is a -- what we do with regard to NCHS is very different now than what we did then. I mean, it basically was the whole committee activity there and dealing with the surveys and we are not dealing with it at all now actually and it is hard to see how we could, given the numbers that we have got and the capability that we have got.
I mean, we have got other capabilities but we just don't have those in enough numbers anyway.
DR. FRIEDMAN: My own sense of this or my own whatever, fantasy about this, is not that NCVHS would stop being involved in anything we were involved in. Having said that, I agree with, I guess, John and the conclusion that I guess John and Ed Sondik have come to that I think NCHS, like NIOSH, like the rest of CDC, you know, needs some institution specific advice and institution specific ongoing advice and some of that would undoubtedly continue to come from the committee as it now does. But having said that, I don't think that the committee is -- has enough people -- I don't think that the committee has enough people in any individual category to spend the amount of time to devote to that ongoing advisory role and in addition to that, our advisory role, as Don Detmer used to point out to me when the discussion came up, is to the Secretary more than to NCHS specifically. There may be some advantages in keeping those two roles distinct.
MS. COLTIN: You know, John, the only concerns I have -- and I agree with everything that has been said in terms of the shortage of time, of people who do have expertise to advise NCHS among the committee members here and that most of us are committed to working on other activities of our various subcommittees, but I would hate to see the Subcommittee on Populations, which is the main locus of discussion for these kinds of issues, discussion of surveys, discussion of existing data sources and how they may need to be improved and so forth, marginalized really as a result of creating a separate body that would take that on.
I would, you know, want to see it structured in such a way that the NCVHS Subcommittee on Populations actually was able to play an important interactive and partnership role with that advisory committee, that to the extent that there were opportunities to have overlapping agendas, that that should be encouraged and that when there are activities going on in this new advisory committee that relate to areas that have been identified as priorities for our subcommittee, that, in fact, that collaboration be actively pursued wherever possible.
So, that would be my main --
DR. NEWACHECK: I just wonder about how much time would be required to do this, that is, to provide this advisory role. I mean, so far, at least in my membership on this committee, we have spent zero time in a formal advisory kind of role. I am just wondering if we were to spend, say, a portion of, say, subcommittee time or full committee time doing this activity, would a couple of hours per meeting, would that be satisfactory or are we talking a lot more than that?
Are we talking about full day meetings once a quarter or something or do you have a sense about that?
DR. STARFIELD: No. It wouldn't be enough. You know, it consumed the whole committee before. That is an overstatement, but you stop to think about all the surveys and the NCHS is changing its surveys. it is going into new and innovative kinds of things with linkages and longitudinal stuff, none of which we dealt with before.
So, I think it is a really considerable effort, at least as much as it was before.
DR. LUMPKIN: But let me add that I don't think that would change the role as NCVHS or NCHS is -- gets to the point where they are ready to say this is what we are thinking about doing a survey and doing this survey integration and I believe they ought to come here because conceptually I see that the role of this committee has changed from where it used to be in that its role is advising the Secretary and its higher level data issues concerning the whole Department and that our point of input tends to be at the point of the Data Council or to the Secretary.
So, I would see that they would -- well, first of all, whatever process it takes to put this committee together, it is going to take years.
DR. STARFIELD: Are there any alternatives, John, to just setting up a separate committee? I mean, is there any way we can enlarge this -- I mean, I think most of us would prefer to keep it in this committee if there is any way we can do it.
DR. LUMPKIN: I am open to alternatives and, you know, enlarging the committee takes an act of Congress.
MR. BLAIR: Could I get one clarification on this, if I may?
DR. LUMPKIN: Yes.
MR. BLAIR: As the discussion proceeded, it sounded as if we are not divorcing ourselves of the role of being an -- in an advisory capacity to the Secretary with respect to population statistics or vital and health statistics. That is continuing. What I think you were proposing was very sharply focused and that was that an ad hoc role that had evolved over time was that our committee because we worked so closely with NCHS had in many ways been considered to be an advisory board to NCHS.
I don't know if that was ever a formal responsibility that we have and I thought that that was the piece that you were surgically winding up saying let's not -- let's separate that piece out. We will still have the communications by having somebody, whoever is the leader of that new emerging advisory board still be on NCVHS. We would still be on NCVHS. We would still retain all our responsibility in advising on these same issues to the Secretary.
Is that a correct characterization of your proposal?
DR. LUMPKIN: That is conceptually how I think of it, yes.
Simon.
DR. COHN: Actually, I will pass because I think some other people have some comments.
DR. LUMPKIN: Jim.
MR. SCANLON: Let me also talk about the evolution of the NCVHS as well. The committee is 50 years old this June. So, obviously, it goes back even before there was an HHS. It initially started out literally as this country's advisory group to the Office of Vital and Health Statistics, which was a predecessor of NCHS, but there was a time when everyone realized that data issues were broader than the statistics agency. For quite awhile, probably for at least the last 15 years -- Marjorie, correct me if I am wrong -- the Department has viewed it as a broader source of external advice to the Department in general and all data issues.
Within that time, I think the committee still had the same struggle of how much time to devote to NCHS on certain issues. Clearly, it was not the full time occupation of the -- even the pre-HIPAA committee here, to devote full time to NCHS. There was a lot of interest in uniform health data sets. There was interest in research confidentiality and on the surveys there was sort of a mix depending on the interest of the members.
There was a great interest in long term care issues for awhile. So, I don't think even the pre-HIPAA NCVHS was not devoted solely to NCHS advice. Clearly, when HIPAA requirements came along, HHS did not want to have two committees advising us in the data area. We went through a lot of internal discussion about how we could make the NCVHS serve as our broad health information policy committee for HIPAA and broader interests, as well as incorporate the prior population-based focus that it had previously.
I think everyone was fairly pleased with that and I think in general most people still are. Population-based data is more than NCHS. Certainly, HHS looks to this committee for broad policy input on population-based data.
But Dr. Sondik, I think, feels that he needs some advice specifically for NCHS and I think that is sort of what prompted him approaching John. I think they did try to look at were there ways of using a subgroup of this committee or designated members of this committee to kind of serve as an NCHS director's advisory committee. I guess that is what prompted this area of, well, maybe I need my own advisory committee.
DR. LUMPKIN: And my response, which was similar to Don's and I think it is really when recognizing the challenges, we have got a lot of stuff that is on the agenda for the Population Subcommittee that we just don't seem to be able to get to. The issues related to race and ethnicity, I think, is a very issue, far transcends anything available at the NCHS. We just haven't had time to get that on the agenda.
So, looking at the kinds of work that would need to be done, I thought, you know -- and this is why I am bringing it up, is that it would be better to look at that committee that may be formed if it comes up that we would not be opposed to it and much as when CDC formed the -- HISS board, what does it stand for? Health Information and Surveillance System Board, which is an advisory committee on the CDC on surveillance systems and developing standards for those kind of systems.
But my concern was is that there is a -- that is an internal committee, advisory committee to the CDC, just as this would be an internal advisory committee --
MS. GREENBERG: No. That is made up of CDC -- that is basically an internal group, not an external --
DR. LUMPKIN: There is an external advisory group that is similar, but just as that committee was formed, I would see that this committee would be concerned that an advisory committee to the NCHS could go in their own direction and my thought is is that by trying to concretize a relationship by having, you know, an overlapping key person would be the way that we could assure where that committee is going and where we are going are not going in opposite directions.
That is kind of my thinking.
MR. GELLMAN: You think an exchange of hostages will accomplish that?
[Multiple discussions]
DR. LUMPKIN: Thank you, Bob, for putting it in that perspective.
DR. FRIEDMAN: John, another very similar possibility to what you and Ed had discussed might be having several more hostages so that, in fact, there would be not just a point -- a single point person, but certainly the work of the Population Subcommittee, the work of all the subcommittees, because they are all relevant, could be more fully represented than just one person. In other words, having a larger subset of NCVHS members who serve on the other --
DR. LUMPKIN: Paul.
DR. NEWACHECK: I think if we are going to have a separate committee, I would be much more in favor of that approach than having a single person. I think if we had a single person representing the NCVHS and the directors' advisory group, it would basically be giving up our responsibilities and absolving ourselves of that role that we originally had and I think that would be a big mistake.
I think we have to keep in mind that the HIPAA stuff won't be there forever hopefully. I mean, hopefully this will all be resolved someday and we will all be doing different things and we don't want to have
-- the National Center for Health Statistics will still be there presumably and we don't want to have our role relegated to the sort of non-existent --
DR. COHN: You know, I am struck -- as being a member of the Executive Committee, I think we are even talking this coming summer about doing what I describe as strategic planning and this is sort of a key element of strategic planning, like what is the committee going to do in the future and what are relative priorities.
It is sort of suddenly being dropped on the full committee. First of all, I guess, it would be useful for me to understand from the Population Subcommittee and, of course, the chair isn't here right now, about if this were one of the priorities that she had to deal with, where this would fit into her sense of relative priorities because I don't have a sense from the subcommittee that this just hadn't been something that was really on the radar screen at a high level and, gee, as they look at this versus some of the other activities whether or not this takes a higher priority. I don't know, but it does sort of have that feel of sort of strategic planning, strategic decision-making.
Now, having said all of that, I am still a little confused about this one because the only thing that you have commented about so far -- and, once again, my view of NCHS is that they do a lot of things of which surveys happen to be a piece of it and, yet, the only example you have given so far about them needing advice that we are not giving is surveys.
Now, certainly, I personally -- if, indeed, the issue is is that NCHS needs a special survey advisory group to advise them specifically on surveys, I mean, that is one set of issues. If, indeed, NCHS thinks that they need an advisory body to advise them on a broad range of issues, some of which I really think is mostly this committee's responsibility, then I guess, you know, I feel differently. So, it gets also down to being an issue of charter.
DR. LUMPKIN: Let me just sort of -- if I can sort of see what I can take away from this, we are kind of not opposed to the idea. We are kind of not in favor of the idea. We need to process it a little bit more, get a little bit more information, recognizing this as an issue that needs to be addressed.
Obviously, HHS can do as they choose, but I think that we would like to explore options to consider the interaction. I will just take this from a very selfish viewpoint, as chair of this committee and knowing the amount of dedication of its members -- and I really have not been on an advisory committee nor do we have any in the State of Illinois that has as much of a commitment from volunteers essentially as this committee does.
Taking on additional responsibility just doesn't seem to fit and knowing the commitment. So, you know, the idea of having more hostages would be wonderful. I also know that most of the members are already hostage to this committee and if you get any of the responses that I get from my employer about your activity on this committee, the thought of doing more stuff is a little bit of a problem.
I believe that people on this committee should primarily be doing our committee's business, which is a lot. So, given that, let's let this sort of sit for now. I don't think we are in a position for making a decision. There is going to be an Executive Committee meeting in the summer where we will do some long range planning and we will come back with additional suggestions to the committee at a subsequent meeting.
MS. COLTIN: Have we set a date for that Executive Committee meeting?
DR. LUMPKIN: It is in Boston and it is probably like the 27th or 28th.
MS. GREENBERG: First of all, we are having on May 9th. Dan tells me that could be a problem for the two of you, is the next Executive Subcommittee meeting, a one-day meeting in Washington, currently scheduled for May 9th. Then this retreat is the 17th and 18th of August in Boston.
DR. LUMPKIN: It will be hot and humid in Boston, but --
MS. GREENBERG: We may kill Boston forever in August, too.
DR. LUMPKIN: Okay. So, now we are going to -- this is going to go on the back burner for now. Let's take the remaining time and do some follow-up discussion of the session this morning.
DR. STARFIELD: Does everybody know what we are talking about? You are talking about the panel.
DR. LUMPKIN: The panel.
DR. STARFIELD: Well, let me just throw out something that seems clear. We have got three major agencies in HHS, each going separate directions with regard to measurement of health of the population, at least with regard to summary measure of health of populations.
Kathy Coltin elicited the fact that HCFA is doing one thing, CDC is clearly doing another thing with the disease oriented stuff and NCHS is doing something quite different in its surveys. And there ought to be, you know, some way of putting them altogether. That is really what the integration of data -- what the Data Council is supposed to be thinking about and it hasn't happened yet.
What can we do to help it? I mean, I think that is the issue.
MS. BLANKENBAKER: Linda Blankenbaker from NCHS. Ed Sondik is the chair apparent of what will be called Interagency Working Group on Summary Health Measures that will include people from -- at higher levels from all of the Department's agencies, who are doing something like this. So, I am in the midst now of putting this committee together.
DR. STARFIELD: So, we just joined the last two issues we were discussing.
MS. BLANKENBAKER: And we could probably have something to say in June because I am hoping we will have our first meeting at the end of March, first part of April. This is really being convened as a response to an IOM report that was requested by Dr. Varmus at NIH, really to look at the burden of disease in terms of NIH priorities and so forth.
One of the recommendations that came out of the IOM report was to convene an interagency committee to do this and it was thought that -- actually I thought maybe the Data Council could actually take this up, but it was thought that that was probably not the best venue that NIH really wanted to go forth and have this particular committee. So, it is going to have an NIH representation of a number of the institutes, who are also doing a number of things in this regard, plus HCFA and CDC and NCHS and SAMHSA if we can con Sam Corper(?) into joining us and any of the other departmental entities that are doing something in this regard.
So, we should have something to say that will make this look a little more coherent perhaps in June.
Jim.
MR. SCANLON: I guess I would also caution that I think all of our panelists this morning discussed the developmental nature of these measures. There are others and none of them is necessarily the standard. I think we are not looking for -- to declare one of these to be the industry standard. I think we want to encourage a lot of work, but we would probably like to have one agency or two looking at these in a critical evaluative sort of way. I don't think we are trying to stop work on all of these and I think that is -- it is not yet a standards issue. Let me put it that way. There is still a lot of developmental work.
HCFA is not here to defend itself but HCFA is the biggest health insurance group in the United States. They have to do things of a regulatory nature and a lot of forces impact on them that require them to take action rather than a research agency. So, I think to bring these perspectives together -- the focus is to look at all of this, not to declare a winner.
DR. FRIEDMAN: Jim or Linda, do you think it would be helpful -- having said that, do you think it would be helpful, for example, a set of criteria to be developed, which could be used in saying, you know, here are the elements that are needed by MCOs, by state governments, by local governments, you know, et cetera, et cetera, in culling through the different through the different measures and, hopefully, at some point moving towards the measure -- hoping at some point moving towards a single measure or a set of measures that could be used at various levels and various levels in the country by various types of organization in the country, at the plant level, et cetera, et cetera, as well as internationally.
Diane Wagener isn't here, but, I mean, for example, I am already being asked -- people are already demanding of me and want to know why we can't provide data at -- for communities of 20, 30, 40, 50 thousand on the 2010 leading health indicators because that is the expectation. I am sure I am not unique in that.
MR. SCANLON: Well, in fact, I think we were just thinking maybe we could -- we will have to look at this, but maybe we could, as we did for the HIPAA data standards, think about a liaison on the committee here. I don't think that would be a legal problem, but that might be a possibility.
DR. COHN: I am just going to comment, I continue to be struck that this work and the work of the functional status activity in the Population Subcommittee feels to me very, very close and I am -- obviously, when I ask the question about what sort of data needs -- what data are you using to do all these things, I didn't hear anybody saying, gee, industry standard data. One worries that NCVHS will come up with recommendations on functional status.
Another group will come out with recommendations on summary health measures. They won't require all the same sort of data elements, data requirements and we will be left with once again another non-standard data situation.
So, I guess my comment would be it is obviously that that subcommittee needs to work very closely in its work on these areas with that activity to make sure that they converge.
DR. LUMPKIN: Marjorie.
MS. GREENBERG: I guess when I first saw the report from IOM on the summary health measures, I think it actually mentions the NCVHS as a forum for looking at these issues and it struck me that Michael Wolfson's point particularly was how beneficial it would be to have some international consensus, which of course -- it is sort of like in standards -- I am always amused when we can't agree on a national consensus, we decided to go and try to work on an international consensus as if it is easier or something.
But if you think about the origin of this committee, it actually did come out of a WHO recommendation and one of the areas that the committee used to work in and I totally concur with what Jim said about -- certainly in all the time I have worked with the committee since the early eighties, it has not been primarily a committee to NCHS, but it did have more time to address the NCHS issues.
My first subcommittee assignment was on the Subcommittee on Statistical Aspects of Physician Payment Systems and I can assure that has nothing with NCHS, but in any event, they did -- this committee did used to and was intended initially to be -- to work with international data issues and that is the longstanding interest in the international classification of diseases really goes back to the origins of the committee. This is a very important issue right now in international health statistics. I think other than some of the work with the NHII has brought in some of that international perspective, but like everything else, there hasn't been as much time to address to that.
So, this does tie in a number of different really historical roles of the committee, as well as current. I know that Liza was -- certainly saw the same connection you mentioned.
MS. COLTIN: I would say another area of overlap is that NIH isn't the only constituency for information that is useful for setting priorities. Clearly, I know that the process that is being used at NCQA, at the Performance Measures Coordinating Council and I would expect eventually by the Framework Board of the National Forum is, you know, how do you pick the areas to develop performance measures if you can't look at priorities according to some logical scheme that includes disease burden or any other type of burden. I think economic burden is also a factor that we looked at.
So, I think that there is a lot of overlap between the work of various members of this committee and various subcommittees in this area and interest, I think, in participating.
DR. STARFIELD: I don't think we are anywhere near recommending standards or a gold standard measure, but what I heard Ed say, which was very persuasive to me is that at least the different agencies should be trying each other's methods, you know, for the purpose of comparison and what they each accomplish.
MR. SCANLON: Actually, Barbara, I think one agency maybe should be -- one agency could be the focal point. NCHS, historically, had very much a lead role in health status measurement. I think that is probably what Ed is ultimately potentially thinking, that perhaps -- I wouldn't let him recommendations, you know, before they are cooked, but that would be a possibility that rather than everybody trying everybody else's measures, that there be a focal point for analytic and evaluative work and comparative work in this area.
Obviously, the general purpose health statistics agency is a fairly -- you know, it would make sense to start there.
MS. COLTIN: I want to clarify the intent of my question earlier about HCFA's approach and role in all this. It clearly was not meant to be critical of HCFA because they are under a lot of pressure to develop measures, not all of their own making. I mean, Congress has imposed upon them the requirement to do this and they have had to grab onto whatever was available and that may not always have been the result that would have occurred had there been a more deliberative process and evaluation.
But that is not to say that those choices can't evolve over time as new knowledge is generated about what might be preferable alternatives. So, my suggestion was merely to take advantage of an opportunity that appears to exist to look at what they are doing and to compare the method that they are using with some of these other methods that the other agencies are exploring, given that the data will exist to compare those methods and to then use that information to perhaps make a more deliberative choice as to what may be the best way for them to proceed into the future.
I think that that is not at all out of line or atypical of the approach that the predecessor to the Subcommittee on Populations took in looking at some of HCFA's choices around, you know, measures of post acute care or, you know, the comments that we made about OASIS(?), the choice of OASIS or the MDS as different options for looking at home care and so forth.
I think that is very much in line with the kinds of evaluations and advice that we have given and we should continue to do that.
DR. LUMPKIN: So, we need to at least add something to our agenda in June to get a report back, hopefully, from the first meeting of the group and perhaps if it is possible, we will have a liaison with that group from the committee.
Okay. Let's take a five minute break.
[Brief recess.]
DR. LUMPKIN: I would like to welcome our panel members to update the committee on a committee report, which has been quite startling. I suspect that it might be fair to say that never has a report by the Institute of Medicine garnered as much attention, including a response by the White House immediately after it was released. Of course, this committee is kind of familiar with that.
We have had the opportunity to have our work prohibited by the Vice President. So, we know what that is like. But it certainly is a pleasure to have both of you here to present on the Institute of Medicine report and the response and the QIC.
We have Dr. Janet Corrigan and Dr. Gregg Meyer. As I understand it, the IOM goes first.
DR. CORRIGAN: Do you want a little bit of an overview on the report itself or basically what has happened since the report out? I don't know how much you have -- if you have all read the report, I don't want to spend a lot of time -- you want the overview of the report. Okay. Great.
DR. LUMPKIN: Please, both introduce yourselves and speak into the microphone for us, please.
DR. CORRIGAN: Okay. I am Janet Corrigan. I am director of the Health Care Services Division at the Institute of Medicine.
DR. MEYER: I am Gregg Meyer, director of the Center for Quality Measurement and Improvement at the Agency for Health Care Research and Quality.
DR. CORRIGAN: Let me give you just a little bit of an overview of the report. This is the first in a series of several reports that the Institute of Medicine will be releasing under the Quality of Health Care in America project. The medical errors topic was selected as sort of a starting point for several reasons.
Our committee felt very strongly that there was a very good evidence base to support that this is a serious problem and to call for very strong action, but the second reason was also because we felt that this is an issue that can be readily understood by both the American public, by providers, by administrators, by policy makers and by others. So, unlike some other aspects of quality of care, which are a little bit harder to kind of get your arms around and to reach an agreement about, the whole area of medical errors is one where many different constituencies can generally reach a consensus once they have looked at the evidence.
For the most part, these are incidents, they are events that really just should not happen. The Institute of Medicine committee estimated that deaths due to medical errors, deaths that occur in hospital settings, range, if you look nationwide overall, we extrapolated from two large scale studies, the Harvard Medical Practice Study and then the most recent study in Colorado and Utah that Eric Thomas and his colleagues have conducted, and then developed point estimates, extrapolations for the nation overall from those two studies.
That is where you get sort of what is commonly referred to in the press as a range and it really isn't a range of 44,000 to 98,000. But at any rate, that is where the figures came from. If you take even the more conservative estimates, you have something that is comparable to roughly the eighth leading cause of death.
So, it clearly ranks right up there in terms of an important policy issue that needs to dealt with by the health care system.
The IOM committee came up with a strategy that is fairly comprehensive and calls for changes at both the national level and the state level and the local level and in both the private sector and the public sector. I think sort of at the heart of the set of recommendations is a recommendation that a center for patient safety be created within AHRQ and that the resources for that center be significant.
Our committee recommends that appropriations somewhere in the range of 30 to 35 million during the very first years, that is per year, be appropriated and then that move upwards of a hundred million dollars a year for patient safety and that that center do several things, that it provide national leadership and a focus on safety for the country overall, that it identify some key measures of safety that can be tracked on a national -- at a national level and a state level and that it produce an annual report to members of Congress and to the President and to leaders in the private sector, letting them know whether we are making progress at meeting the goal of decreasing avoidable and preventable errors.
Our committee also felt that based on the knowledge and technology that is currently available to improve patient safety, that it is reasonable to expect a 50 percent reduction in medical errors over a three year period. That is interesting. Some people think that that is a very ambitious goal and other people think that it is not enough. But what it does reflect is a strong sense on the part of the committee members that this is something we can tackle, that it really can be done.
Other industries have tackled safety problems and done it very successfully and that there are many systems out there that can be put in fairly quickly in hospital settings in particular and that is where we recommend starting this effort, in the hospitals, that really will have a significant impact within a fairly short time frame.
In addition to the center for patient safety, it is recommended that there be both mandatory and voluntary reporting systems. Let me start first with the voluntary. The purpose of the voluntary systems -- and there are some that already exist, the Institute for Safe Medication Practices, the JCHO has a reporting system of sorts. They aren't heavily used, but they do exist and have had some degree of success.
The voluntary systems that we reference are ones that would primarily collect information on near misses or errors that resulted in very minimal harm. These are not the serious events. These are the kinds of things you could learn from. So, you want to report them and carefully study them and track them and look at trends and inform the industry overall when you learn about particular conditions that are creating an unsafe environment. They are the early warning signs.
The committee recommends that legislation be passed to extend peer review protections to the data and information that is reported into voluntary systems. We want these voluntary error reporting systems to thrive and we want there to be no impediments to reporting or fear of reporting. So, we really want these systems to thrive and for there to be many more of them and much greater participation on the part of the health care industry in such voluntary reporting systems.
That part of the recommendation on reporting is less controversial than the other part, which has to do with the mandatory reporting. The mandatory reporting, over the long term we do expect that it would improve safety because it would provide an incentive for all hospitals in particular to put in internal patient safety systems and to improve their environments, but the primary purpose of the mandatory reporting is really for accountability.
The events that go into the mandatory reporting system are those that hare very, very serious, that resulted in death or serious permanent injury for the patient. It is a very narrow band of events of incidents, not intended to be a long list, intended to be a very short list of those really serious events, the kinds of things that, as I said, most agree they just shouldn't happen.
They are thought to be ones that indicate possible breakdowns in management within health care organizations. They are also ones that have felt that they are so serious that the community has a right to know about them. So, our committee recommends that information on these events, these very serious ones, be released to the public, not information specific to physicians or nurses or to the patients involved, but institution organization level data.
The focus here is on organization reporting because medical errors are very rarely, only a small fraction are the result of impaired or incompetent providers. The vast majority are systems problems. They result from environment that we have in health care that make it very easy to make mistakes, rather than having environments where it is hard to make mistakes.
So, the reporting system serve somewhat different, really very different purposes. The mandatory reporting that our committee has recommended would be nationwide and the reporting would be standardized with the new quality forum that Ken Kiser(?) heads being given responsibility for developing the standardized list of events to be reported. What are the most serious ones, as well as the information on those events that would come in descriptive of them. That would be standardized. However, our committee was recommending that the states be responsible for administering the mandatory reporting system. We think that there are -- we know that there are at least 20, now some estimate 23 states. We don't have the exact number, but it is somewhere in the low twenties is the number of states that currently have mandatory reporting systems of some sort.
They are highly variable in their performance and we don't frankly know a lot about their performance. We don't know a lot about the voluntary systems either and they, too, the ones that we know about, are variable. So, this is an area where we are breaking new ground to a great extent, both in terms of voluntary reporting and mandatory reporting.
So, the idea here basically is nationwide standardized reporting but administered by states with disclosure of information at the institutional level to the community. It is that latter point that I think has perhaps been most controversial. Our committee felt very strongly that a community has a right to know what is going on in the health care organization that is serving them.
The third part of the approach or a third set of recommendations really relates to strengthening accreditation and licensure processes, all of the various oversight processes that might be strengthened to focus a bit more attention on patient safety issues. So, it is basically a call to professional groups, to licensure, to accreditation organizations, to peer review organizations, to all of those involved in oversight and in the training and retraining and certification of providers and accreditation of institutions to focus greater attention on patient safety.
Then the fourth part, which is I think really the most exciting is a strong call to health care organizations, that they begin to implement proper systems. Our committee thinks that a really good area to begin with, where there is sort of low hanging fruit has to do with medication errors because we have methods available to reduce medication errors, whether you choose to go with automated order entry systems or you choose to go with the low tech interventions, like having a pharmacist as a part of your team that does rounds in the hospital, both of which have been shown in studies to be very effective interventions.
That is one area that we would recommend starting with. Since the report was released, as John indicated, it received a lot of attention. It was covered in just virtually all of the national, as well as a great deal of the local news media, both print and broadcast media. And we got a couple days of coverage because different events related to this sort of kept coming, so it kind of went on a roll. We got overwhelmed by it. All of the Academy's phone systems went down on the day that we released it and we had a heck of a time trying to respond to it.
But we also, I think -- one of the things we did try this time actually for the first time in the release of a report is we did a satellite media tour and we had Lucian Leape and Don Berwick and Molly Joel Coye, they went to a studio where they were for a couple of hours in different rooms and they can't see who is on the other end, but there is different media from communities from all across the country that are fielding questions at them and they are on film.
So, that allowed us, I think, to really get the message out in a better fashion, more broadly, both not only within the national media but within the local newspapers and the stations. That has really continued right onto this day. It is still getting considerable report.
There have now been a lot of hearings on the Hill, quite a few which Gregg probably is following even closer than I am at this point, both at the Senate and the House, with very keen interest and we have seen some really good actions taken on the part of the business community. The Business Round Table has announced an initiative. They have established a committee to start to work mostly with the Fortune 500 companies. The National Business Coalition on Health has identified it as their top priority and they are beginning to take actions and programs they are putting in place over the next couple of years.
So, those are very promising as well and we have heard a lot from states that are interested, some that are looking at other reporting systems and trying to figure out how best to structure theirs. There is a meeting at the end of March that AHRQ is -- a liaison program, convening a lot of folks from states. So, we will have a little bit better idea then how quickly some of the ideas are going to spread.
DR. LUMPKIN: Thank you.
Gregg.
DR. MEYER: I have handed out my slides to you and let me first say that we did not learn from the systems error of the IOM in terms of producing their report, which apparently ran out of copies very quickly. We followed their lead. We ran out of copies of ours equally as quickly.
With that said, this is available to you at the Web site of the Quality Interagency Coordinating Task Force, WWW.QIC.GOV, and the full text is downloadable from there.
I am talking to you this afternoon from my viewpoint as a member of the Quality Interagency Coordinating Task Force, as opposed to someone from AHRQ, although I will be happy to enter into discussion into various questions relevant to either of those roles.
I am going to skip through the first few slides because those were covered by Janet Corrigan in terms of the IOM report. What I would like to do is to start telling you a little bit about some of the administration actions on quality. It will take a minute for the projector to get on line here.
But really the first of these actions was the President's creation of the Advisory Commission on Quality or the Quality Commission in early 1997. In response to the release of the quality report, there were two specific actions. One of those actions was the establishment of the Quality Interagency Coordinating Task Force or the QIC. The QIC was the body that created the report that I am discussing this afternoon.
To give you a bit of background on the QIC, the Quality Interagency Coordinating Task Force is the group of all federal partners. So, this is the Federal Government pulling together around the table all departments and agencies that could have something to do with health care quality or could have an interest in health care quality.
For example, it includes the usual suspects that you would suspect; CDC, Agency for Health Care Research and Quality, the NIH, HCFA, Department of Veterans Affairs, but there are others that you may not suspect, groups like the Department of Justice because the Bureau of Prisons has their own health system; groups like the Department of Transportation because they regulate ambulances. So, they, too, are part of the health care system.
There is a very broad net of federal players. The other action that the Federal Government took was the launch of the National Quality Forum or the National Forum for Health Care Quality Measurement and Reporting. That is actually a very important body in terms of discussion, I will give you in our response to the errors report.
One important thing, it is recognized is that there has been some federal action on errors. For example, there has been research that is funded by the Agency for Health Care Policy and Research and before that the National Center for Health Services Research. That is basically the funding of these studies that really formed the corpus of evidence that made up much of the IOM report. In addition to that, we have done some other studies, for example, testing the use of computerized systems in terms of pharmacy ordering, some of the work that has been done out in Latter Day Saints, looking at the availability and the applicability of informatics to improving patient safety.
There has also been action elsewhere in the Federal Government, the VA Centers for Patient Safety. The CDC has a National Nosocomial Infection Survey. The FDA does their data collection in terms of drugs, devices and biological products, such as blood. The VA has a reporting system on medical errors.
We also collect administrative data from hospitals as part of the Health Care Cost and Utilization Project, the HCCUP project, which is being -- right now, some additional work is done on the quality indicators there to make that more applicable to looking at medical errors.
I am going to skip over some of the specific things that we have done in the past to really focus on our federal response. What the President charged us with doing on December 7th, was within 60 days to, number one, to evaluate the recommendations from the IOM in their report, "To Err is Human." The second is to identify prevalent threats to patient safety and the third was to identify ways to reduce medical errors throughout the nation's health care system.
So, the President asked us to look at what the IOM said that we should be doing, determine which of those are appropriate roles for the Federal Government and then look beyond what the IOM recommended and ask ourselves what else can and should the Federal Government be doing to address the patient safety issue.
Let me go through our response relatively stepwise over the next couple of minutes. The first of their calls was for leadership and specifically what they asked for was the establishment of the center for quality improvement in patient safety or a patient safety center at AHRQ. We have taken -- decided to take an approach in our recommendations back to the President that, in fact, is a mirror image of what the IOM has done.
The IOM considered patient safety and medical errors within the broader quality framework, within their works on overuse, underuse and errors. We have decided to take a similar approach in that we are going to integrate these patient safety activities in a much enlarged and much broader focused quality measurement improvement activity at the agency.
We are going to support research into medical errors reduction and we have begun some of that initiative already. As you may have noticed, in the middle of December, about a week after the IOM report was released, we released an RFA looking for best practices in improving patient safety. One could be cynical and say that that was a response to the IOM report, but for those of you that are familiar with the way the government works and the budgeting processes of research agencies, you recognize that the first draft of that was written six months earlier, before we ever knew this was going to be a significant issue.
We are also working on converting findings into improved practices. So, it is not just building new knowledge, but it is also trying to develop some of the tools that can be used to improve patient safety and then trying to translate that research into practice, following the AHRQ pipeline; fundamental research, developing tools, translating research into practice.
We are also going to work quite quickly in trying to conduct some national summits on patient safety within the next year. Particularly what we are looking for there is some broad agenda setting to get us to make sure that we pull in all the stakeholders so that we use whatever appropriations the Congress provides us to work on this issue in the most wise fashion.
The FDA will be pulling together similar summits looking at drug and device safety and the VA will be looking at patient safety practices.
In terms of model programs that currently exist, HCFA is going to publish regulations requiring in their conditions of participation to have ongoing medical error reduction programs in place. The definition of those is going to be garnered with the help of the Quality Forum.
The OPM is going to require all federal employee health benefit plans to seek accreditation that includes evaluation of patient safety programs to reduce errors. So, we are using the Federal Government as a very broad system of leverage here. The VA and DOD, Department of Defense and Veterans Affairs, are both going to continue to lead there by example. They both have a number of patient safety programs in existence. They are going to be evaluating them and generalizing them over the coming years.
The most controversial point in our recommendations, as Janet pointed out to you, is clearly those dealing with reporting and the difficulty here is balancing two competing needs or two competing purposes and those are the purpose of learning, to collect data for the purpose of learning and to collect data for the purpose of accountability.
I will tell you that this was truly a struggle among the participants in this process and still to this day, obviously, if you are following the latest series of articles in The New York Times and you continue to look at it over the coming days, you will see that it continues to remain probably the most controversial part of our recommendations.
What we did was we again followed the recommendation of the IOM, saying that there is an important need to know and that need to know is best exemplified by a certain number of limited circumstances and those were the most egregious, preventable errors, things that we all would agree should never occur. What the QIC is going to do is we are going to urge public disclosure of provider use of a forum identified of proven patient safety practices and we are also going to be asking the forum to help us define what those egregious errors are.
So, that is not coming from within the Federal Government, but it is coming from the private sector, with some public sector input of the forum to help us come up with that list of what those egregious errors, in fact, are.
It is important, though -- I want to point out to you the difference between collecting that data on errors that occur and the other point here on the bottom of the slide is the other thing that we are asking the forum to do is we are asking the forum to say what are things that health systems should have in place to improve patient safety. For example, should they have a physician automated order entry for medications? I haven't written a prescription in seven years and, yet, I have filled prescriptions every week in my work as a general internist. I do that through an automated system through the Department of Defense.
It catches interactions. It tells me when there are allergies. We are looking to the forum to identify a series of evidence-based systems level processes that will improve patient safety and then we will be asking for health systems to report to us whether or not those systems are in place. So that patients have the ability to look and say I am going to be getting my care through this system or through this institution and, yes, they have these things in place that should, in fact, safeguard my care throughout my hospitalization.
In terms of state reporting systems, one of the key issues that is clearly highlighted in the IOM report, particularly if you go to the back of the report, in the appendices, is that although these state systems are in place, we actually know relatively little about them. One of our first and probably most important steps in terms of engaging the reporting issue is going to be evaluating those systems that already exist.
We particularly want to look to see what is working, what isn't working, how much of an issue is it that mandatory reporting will drive reporting underground and have consumers, have patients found this reporting to be useful. I think those are all open questions and we intend to move with due haste to try to start some evaluations to look into those.
In terms of voluntary component of reporting, again, the purpose here is for errors and for those of you that are familiar with aviation safety, what we are talking about here is something similar to the aviation safety reporting system. That is the system where a pilot of a crew member notices that there was some error, some mistake, some problem during their flight or on the ground, which did not lead to a problem, didn't lead to a crash, yet is worrisome to them. They call into an 800 number. They report it. They do it anonymously. The data is stored anonymously.
It is collected in narrative reports. They are collated and there is feedback given back to pilots around the nation on a regular basis in a small newsletter that every commercial pilot looks at. We try to use that as our model of what we would like to move towards over a period of time in terms of voluntary reporting system.
So, this is not just including errors where there is actual patient harm, but it also includes close calls, those instances where patient harm was averted, but it wasn't because we had a system in place to prevent it. It was because we got lucky. The idea is to make sure that we learn from each of those instances.
Again, this is to be confidential. There will be peer review protections here. Those will protect the reporter, the person who notifies the collection body about the error or the near miss. That will include the patients involved. That will include the providers involved. That will include the institutions involved. This is a confidential reporting system. Again, think of the ASRS, the aviation safety reporting system.
Our goal is to develop a national data set for research on patient safety that we would derive for this over time. That is a longer term goal. That is something that we hope to develop with time and with appropriate funding.
Other issues in terms of reporting systems, the Health Care Financing Administration is planning to do a pilot where they will ask 100 hospitals to volunteer to participate in a program of mandatory reporting. That will be evaluated and the lessons learned from that will help us decide in terms of what federal actions are important in the future.
Before I leave on some of these reporting issues, I do want to emphasize one important fact that is somewhat obscured in some of the discussion, after the release of the Federal Government's recommendations and that is that we intend to revisit this issue of mandatory reporting in three years. It may very well happen that we find some best practices that can be generalized to all 50 states and we will push forward with that effort. However, we may also find that there was limited value and we would have to reconsider our recommendations in terms of mandatory report. We have left that door open and we have done that with great thought.
The VA is going to augment their new voluntary data collection. The Department of Defense is going to institute a model very much like that that the Veterans Affairs currently uses and that is going to be piloted in some of the DOD centers around the world.
Again, OPM is going to use the leverage of the Federal Employees Health Benefits Plan to move some of this forward.
In terms of building on reporting systems, AHRQ, my home, is a research agency and what we hope to do is we hope to lead the effort to evaluate those existing reporting systems, again, evaluating those state system with due haste.
We also are going to work to include information on errors and patient safety in the National Quality Report that I spoke to you about a few months ago. That has been requested specifically by the Congress. It is also recommended in the IOM that we produce an annual report on errors. We are going to work on integrating data from different sources. This is a long term database agenda to provide information and data for the best and brightest researchers to go ahead and give us some good solutions to patient safety challenges we all face.
The VA is going to establish a voluntary system to supplement their existing mandatory system. They have a mandatory system in place at the present time. They will be developing a voluntary system.
I already told you about the Department of Defense in terms of building their system. The FDA is going to expand mandatory reporting for blood banks to all the 3,000 registered blood bank and blood establishments. They are also going to move to the next phase of an error reporting system.
In terms of what we are going to do beyond the IOM recommendations, first and foremost, we recognize within the QIC, within the Quality Interagency Coordinating Task Force, within the Federal Government that the errors agenda and challenges are large, that they are difficult and that we will be unable to make significant impacts without significant partnerships with the private sector, working with providers, working with purchasers and, frankly, working with patients and the public to try to really augment our efforts in terms of medical errors.
The Federal Government can lead, but Federal Government action will absolutely not be sufficient. It is necessary but it won't be sufficient and we haven't deluded ourselves to thinking that would be the case.
We are also beginning some small projects. For example, we have a collaborative project with Dr. Berwick's Institute for Health Care Improvement to reduce errors in high hazard environments throughout the Federal Government. So, we have the VA, the Department of Defense, the Coast Guard, people from the NIH, AHRQCC, others, participating in a breakthrough-like process that is going to be looking at high hazard environments.
There are some additional actions here in terms of the FDA and the Department of Labor in terms of educating the public through the Federal Employees Health Benefits Plan and also through some work on drug naming and packaging to avoid the Celexis(?) Celebrics(?) problems that we have encountered recently.
In terms of investments in improving patient safety, the President's request for the fiscal year 2001 budget includes $20 million for research and to causes and remedies for errors at the Agency for Health Care Research and Quality. We envision that this would include research into fundamental issues related to errors, the fundamental science of errors, things like looking into human factors, psychology, the ability to translate other lessons learned from other industries into health care, and also some applied research, to look for best practices out there, evaluate what is going on so that we cannot only develop the science base that we need to improve patient safety, but can work immediately to start to take some measures and do some research that will enhance patient safety as well.
In terms of the FDA, some of their programs will be augmented under the President's budget request with $33 million going to enhance reporting for drugs, devices and biologics.
The VA is going to increase their investment in patient safety and that includes the development of an order entry system and really disseminating a bar coding project that they did, which was originally based on the experience with Hertz Rent-A-Cars, where you walk through the door they. They bar coded you very quickly and you got in and out into the airport. It is a great way to collect very, very specific information very, very quickly and helps you avoid some checks on errors.
The Department of Defense is also going to be integrating some of these efforts into electronic medical record projects.
I wanted to leave you with a few issues that I felt, based on our discussions within the Quality Interagency Coordinating Task Force, were very germane to your role here in the NCVHS. The first are what are the changes in our data infrastructure that would be required to capture relevant information to the patient safety challenge? For example, coding; data for non-hospital settings because one of the most disturbing things that we know is they will know the number of patients who are harmed or die from medical errors is quite astounding. One must recognize that that data comes from the places where we can find it.
We are looking under the lamp post at hospitals. All of you will recognize that ambulatory settings, home health settings, nursing home settings, all certainly have additional patient safety problems and we haven't yet developed the ability to do that and it may be that there are some issues for the NCVHS to consider in helping us move forward in those other settings.
What are the lessons from the NCVHS to the reporting challenges based in patient safety? And, finally, what is the relationship of these proposed efforts to other NCVHS work in terms of confidentiality? With that, I will stop. Thank you.
DR. LUMPKIN: Thank you for your presentations. I think Kathy will answer all of those questions.
I actually have a couple of questions to sort of kick things off and, Janet, I know you have got to -- you have a time constraint.
To what extent have --
DR. MEYER: I can actually stay for a few minutes longer so if folks can think about that when they are directing their questions.
DR. LUMPKIN: To what extent did the study or subsequent thought go into patient-related errors? For instance, one of the issues we are struggling with are generic drugs and we have had a number of reports in Illinois of patients who are prescribed a generic drug, particularly those that are associated with critical dose, like warfarin(?) or others. Because the medication looks different, they overlap the old prescription with the new prescription because the pills don't look the same so there is confusion in how the patient takes the medications. To what extent was that looked at?
DR. CORRIGAN: Our literature review did cover -- our literature review went back about 12 to 15 years. We only looked at the publications in the U.S. for the most part and we did pick up some studies of medication related errors and some of them were out of hospitals. So, it did cover all types of adverse drug events.
It is really hard to separate out, though, the various factors that contributed to the adverse drug events. We could not from the studies reach any definitive conclusions about sort of what proportion or fraction was attributable to patient-related issues.
We highlighted it as an area that needs to be studied more because as the numbers of drugs people are on continues to increase with an aging population and a burgeoning pharmaceutical industry, it is just likely -- we do know that errors go up as the numbers of drugs that you are on go up. So, it just kind of makes sense that this is becoming a higher risk area and we need to take a much closer look at it.
MR. BLAIR: Janet, there was apparently a very significant difference in the results from the Colorado/Utah area -- I think it was like 48,000; whereas, the one on the East Coast was like 98,000 lives per year extrapolated. What observations or conclusions did you come to as to why there was such a significant difference between those two sites?
DR. CORRIGAN: Well, the two studies actually corroborated each other fairly closely in many ways. Each of them found that the proportion of hospital admissions experiencing an error was somewhere between 3 to 4 percent. So, they weren't that far off in terms of the number of people who were harmed as a result of an error. What differed was in Colorado and Utah was at a lower percentage that resulted in serious harm or in death.
We can really only speculate as to what some of the causes might be. The studies were conducted ten years apart. So, there is a time issue. They were in very different parts of the country. The methodologic approaches were very similar and they both included very large sample sizes. Clearly, the health care institutions and delivery systems differ somewhat between Colorado, Utah and New York, but beyond that, I am afraid we really can't say anymore.
The authors, Eric Thomas and his colleagues in the study from Colorado and Utah, the did take a look at the same issue as well and cited those kinds of possible factors that might explain some of the differences. But I should emphasize that they were very consistent in many other ways. It was only in the fraction of people who were injured as a result of errors that resulted in serious harm is really where the two studies differed, but they were actually very similar in terms -- remarkably similar in terms of the numbers of people who experienced harm as a result of errors.
MR. BLAIR: Was there any significant difference in the two sites in terms of the degree to which information systems were implemented that might have --
DR. CORRIGAN: Well, given that there was a ten year difference between -- a time frame difference, you would sort of expect that probably Colorado/Utah were a little bit more advanced. I guess the only one other thing we would point out is that, obviously, Intermountain is in Utah. So, there was clearly one institution that is known for very advanced information systems compared to others. That is certainly a possibility.
Of course, many of the studies, too, on the adverse drug events have come out of the Intermountain, Utah folks. So, they have -- I mean, that is an institution, a health system, that has spent an awful lot investigating errors and working on it. So, it may well be information systems. It may well be the effect of their quality improvement patient safety programs.
DR. LUMPKIN: Barbara.
DR. STARFIELD: Both of your efforts equated patient safety with reduction or elimination of errors. There is a whole other area of patient safety that has to do with the unanticipatable non-error adverse effects, which are described in the IOM report reasonably well, but not dealt with reasonably well. They have very different implications for the kind of data that you would need. In terms of error, you need to know a lot about the patient ahead of time or about the system ahead of time.
In terms of unanticipatable things, what you are interested in in terms of data is accumulating information over time so that you can identify the potential for adverse effects, you know, due to poor premarketing, testing.
Is there any attempt on either the IOM or AHRQ to really come to grips with the non-error adverse effects, which by my calculation count for at least as many deaths as the error ones?
DR. CORRIGAN: We don't have any studies underway in that area that I am aware of. It might be of interest to you, though, our committee had an active debate about whether all adverse event should be considered errors or whether you should make the distinction between preventable versus those that were non-preventable. They chose the more conservative route, you know, to stand on the ground that were clearly preventable simply because, you know, where we were headed with this report, but it was quite a debate because actually some actually view that all of adverse events, if we were devoting adequate resources in many other areas earlier on in the pipeline and into the research and testing, that we could avoid so many of those other ones as well.
DR. STARFIELD: Yes, just to pick up on that, you know, there is a difference between whether you look at the problem as one of a hospital problem which you did because that is where the data were, or whether you look at it as a continuum of care. Once you are in the ambulance, there are a lot fewer deaths, but a lot more
-- you know, less deaths, adverse events.
DR. CORRIGAN: Good point.
DR. MEYER: Just to respond, we also were relatively conservative and limited our discussions to preventable adverse events. With that said, we would look for -- and if you look at the now public wording in our congressional justification for fiscal 2001 budget, in terms of what we are asking for in fundamental research in medical errors and on patient safety, it would open up investigations into that exact question, which is what percentage of them are, in fact, preventable and how can you track patients through time and space to be able to capture those.
So, it is a research question that we are open to.
DR. STARFIELD: I mean, it is an issue of --
DR. MEYER: A number of the programs in the departments with delivery systems, for example, the VA and the DOD, those are expansions of things that have started. Many of these are sizeable expansions. So, it is not more of the same, but it is much more of what was hoped for.
In terms of the research agenda, which is, obviously, what is most near and dear to my heart, there, again, our reauthorization Agency for Health Care Research and Quality specifically asks us to fund research on patient safety issues. So, there is no need for an additional legislation there.
In terms of probably the most important ones, where there are important needs for the Congress to work is the extension of peer review protections to some of the results of the mandatory reporting, particularly in terms of protecting the analysis, the root cause analysis of medical errors. That will require congressional action.
Again, if you look at the report, if you download it and go through quickly, we tried to outline those that we can do under existing authority and many of the actions that I presented to you actually are under existing authority.
For example, the conditions of participation mechanism that the Health Care Financing Administration can use, the OPM's ability to work with the Federal Employees Health Benefits Plan to ask certain things of the plans that apply. So, there is much there that actually can be done under existing authority. However, the peer review protection issues are ones that require important congressional action.
MR. SCANLON: Could the center for patient -- for quality improvement and patient safety at AHRQ, would that require new authority is that within the administrative --
DR. MEYER: That is a wonderful question. I will share with you that that is something that we are thinking about and whether we should wait for authority
-- it ends up that the designation of the center actually is probably within our authority, the reauthorization, because it asks to specifically look at patient safety. The President of the United States announced a plan for the creation of that center at AHRQ on Tuesday. So, I am not sure if there is a need for further authority on that particular issue.
MS. BEEBE: I am Susie Beebe(?) with the National Center for Health Statistics.
Dr. Corrigan, you had mentioned about Dr. Lucian Leape and having read his articles about adverse events, I remember reading that one of the major barriers to reporting these is the legal environment dealing with reporting in the health care environment. He suggested that there should be more of an aviation model, where there is non-punitive action. So, I was curious as to seeing your mandatory reporting and three year time frame, I think, how the quality of the reporting would be affected dealing with the punitive arena within the health care arena.
DR. CORRIGAN: That is a good question.
Well, keeping in mind that there are two reporting systems here and the majority of things are going into the voluntary system that is not punitive and is protected, has peer review protections, so, what we are talking about reporting into the mandatory system is that very narrow band of incidence.
The criticism that has come out at times about this is that mandatory reporting can drive reporting underground. I guess we have to keep a couple of things in mind. Reporting is underground right now. We are in the foxholes. So, the question is really whether we can bring anything up and is the better way to bring it up through voluntary or is it to bring it up through mandatory? We probably couldn't drive anything more down than it currently is because nothing is coming out.
On the voluntary side, we hope that a lot comes out. On the mandatory side, what we have selected there is that small numbers of incidence, where it is really hard to hide those events because the results were so serious that you can't put them in the closet. Usually multiple know about them.
So, if there is any form of a penalty for not reporting and if you were going to have it be a punitive system on the mandatory side, the penalty should be for not reporting. What you want to do is discourage an individual from not reporting something that they know about.
So, you would want to structure such a system so that if you have, you know, a doctor or two doctors and two nurses and an administrator who have all observed that there was an avoidable death that just took place and that is probably due to a medical error or that there was wrong side surgery or an event like that, it generally is known to multiple people.
Should they for some reason or another not report that to the mandatory system within 24 hours or 48 hours, then the way one can deal with the underreporting issue is to have a penalty at that phase. Now, where you want to praise people, though, and respond positively is when they do report that incident and they conduct a root cause analysis and they demonstrate that they have taken appropriate actions to put the systems in place so it doesn't occur again.
So, a mandatory system doesn't have to be completely negative but it probably does have to be negative in terms of making it very clear that reporting is a responsibility when there has been a real serious injury or a death. It is because of the way it is structured, the limited focus of it, the kind of events that are occurring that it was felt that it was an appropriate action to take.
Now, keeping in mind, though, also that our committee really does view it more as accountability than it does as quality improvement and that is why they felt there needed to be these two directions working in parallel, one mandatory, one voluntary systems because on the mandatory side it is serving other purposes. It is serving the right of the public to know about what is going on in its institutions.
It is also providing a place for someone to go or for a family member to go when somebody has been seriously injured or killed. It provides them someplace to go to be able to report this incident and in some sense that there is going to be an investigation of it. So, it does serve a somewhat different purpose.
MS. FRAWLEY: Janet, I understood that under the voluntary system there are peer review protections in place, but the mandatory system what I am concerned about is why kind of protections are going to be in place as far as who that information will be disclosed to? Would I as an attorney be able to get a list of cases from a hospital? Would I as a patient be able to get information about other patients?
DR. CORRIGAN: The reporting systems as our committee envisioned them are organization level reporting and under the mandatory system, there would be no reason -- and I believe it is the -- the Administration has been very explicit about this -- it would not identify individual physicians or patients. So, we are not talking about this being something that would breach the confidentiality of nurses, doctors because we don't believe that it is the provider who is the -- should be the focus of this effort. It is really the institution where the resources exist and the responsibility resides to put those systems in place to make it safer. So, that is where the point of accountability is.
But you are correct that your question -- you know, a mandatory system as our committee envisioned it, that once it got up and running and was really moving along, that there would be some reports to the community periodically as to the number of incidents, the type of incidents by institution, but, once again, not identifying physicians or patients.
DR. MEYER: And the recommendations at present follow this very closely.
MR. FANNING: Can I follow up on Kathleen's question? I am John Fanning.
You would have to ensure that the information that is retained by the reporting institution is equally protected. Kathleen may not be able to go to the agency that is collecting it, but what is the status of that material in the hand of the reporting agency, will be one question that will have to be answered?
DR. MEYER: And, again, that is where -- an instance where there will be a need for further legislation in terms of protecting that information.
MR. SCANLON: Do other countries -- any other countries have success with these mandatory reporting or voluntary reporting or other systems? Can the U.S. learn what has transpired --
DR. MEYER: A number of countries are considering similar systems. The ones I know of best is a program in New Zealand, which is just getting started. The U.K. in the wake of the Bristol pediatric surgery problems is considering doing some work in this area and actually there is a collaboration developing between the Agency for Health Care Research and Quality and the Department of Health in the U.K., specifically looking at this -- or the Ministry of Health -- looking at this issue about how we go about collecting the data and trying to learn from each other.
MR. SCANLON: It sounds like there are not well-established systems in other countries. Is that correct? Or there are?
DR. CORRIGAN: We did not do an extensive review of other countries. We heard about some that existed. We also heard about some reporting that took place in the context of no fault insurance kinds of mechanisms, too. So, the liability issue is clearly there and everywhere and some countries have dealt with that somewhat differently as well, which will probably be something we will want to think about down the road in trying to deal with -- because we don't deny that the liability is clearly an issue here as it relates to medical errors and our ability to get at this problem.
Our committee just felt that there was still an awful lot that could happen but limited changes in affording some forms of protection, but still sees that as a major issue that needs to be dealt with at some point.
DR. STARFIELD: But other countries are taking the other approach; that is, you know, trying to identify the potential for adverse effects. For example, both in Spain and in France, you know, they do monitor medications for justifiability of the use of the medication.
DR. LUMPKIN: Again, Janet, we would like to thank you for coming and we know you have to leave. Thank you very much.
DR. FRIEDMAN: I guess by elimination this question for Gregg.
It is a point of information that I don't understand, which is what is the difference -- what would be the difference, for example, between mandatory reporting, where it is clearly a systems problem and one where one would need to protect the privacy of not only the patient but the physician and one where, for example, if you read The New York Times in the past few weeks, there is a very well publicized case in there, where it is a problem where somebody's license needs to be yanked. Are we still trying to protect that physician's privacy and how do we make that kind of a gradation?
DR. MEYER: One of the important points that we tried to emphasize very early on in our report is that, for example, a surgeon carving their initials into a patient's abdomen is criminal behavior.
DR. FRIEDMAN: That is not the one I was thinking about.
DR. MEYER: But I am using that as an example and that in no way would the mandatory systems and the protections that we are building around them negate or stop the need for those processes to take place. This is not supplanting those. This is additional to those that already exist. So, I think that is an important point, that is worth emphasizing. We try to do that in the report. We try to do that in our discussions since the report was released, but I am happy to beat that drum over and over again.
The other important thing to note is that in the IOM's call for action and is largely absent from the Federal Government's response, deals with the third tier of the approach of the IOM and that is in terms of working on accreditation issues. Again, we recognize that that is largely a private sector responsibility and a professional society responsibility. We are working with -- we will work with those organizations to supply them with information and provide technical assistance as possible. With that said, we recognize it is their -- that is their mantle to carry. We support them doing it, but that is for the private sector.
DR. LUMPKIN: Thank you very much for coming. It was a very thought-provoking presentation and we have a couple of charges for the Quality Committee. I see Kathy is still staring at that last slide.
Certainly, we will look forward to working with you on these issues.
DR. MEYER: Thank you.
DR. LUMPKIN: That concludes our business for today. I would like to thank everyone for coming and we will see you in June, except for those who are staying over.
[Whereupon, at 4:03 p.m., the meeting was concluded.]