Agenda Item: Leading Health Indicators

DR. MCGINNIS: Let me just thank you, David, and all of you for this opportunity to come back and share the progress that you have been making.

DR. SATCHER: I am sorry, I really didn't introduce you. Maybe you don't remember. Let me just say that Dr. McGinnis previously headed the Office of Disease Prevention and Health Promotion and has played a leadership role in this area, especially with Healthy People 2000 development. He was co-chair with me on the Public Health Functions work group and he has just done some outstanding work. He is now with the National Academy of Sciences.

DR. MCGINNIS: Thank you, David. It is a treat not only to be here in the same room with so many folks from whom I have learned so much over the years, and who perfected my thought processes about these issues, but also to have a chance to observe the maturation of the process.

There are certain imperatives that were clear, I think, to all of us about the objectives-setting process, if we were going to be able to reach the full potential of this endeavor. One of them was the imperative to have the enterprise become incorporated as part and parcel of agency strategic planning and priorities. It is clear from the discussion here today that that has moved along quite nicely.

The second was the imperative to address in some fashion the issue of quality of care, which is tricky, but clearly necessary to acknowledge. The discussion the last few minutes makes it clear that that is part of your deliberative process as well.

The third was to have us, as a nation, not only develop a large and multi-faceted, colorful fabric that represents our national health goals and objectives and has an element in that fabric for each of the potential participants but also, while doing that, while cutting the broad swath, have a targeted focus on certain key areas that could provide a repeated reminder of some of the priorities and help people to understand what the objective process is about. That is what I am going to talk with you about in the next few minutes.

Actually, I think I will reduce what I had initially intended to say, because I know you are behind schedule. As I look around the room, I see a lot of folks who were involved in the sentinel objectives working group, so they can fill in the gaps either in the question session or after the meeting, as well.

What I am here to talk with you about is the work of the sentinel objectives working group, which was an effort that got underway in the late fall at the request of Susanne Stoiber, who clearly had the sense from grappling with issues around the objectives that there was a need not only to address the broad focus, but also to develop this subset. It was a need that was perceived in each of the preceding efforts, the 1990 objectives and the year 2000 objectives, but yet not addressed as effectively as it might have been addressed. She wanted to make sure that the process for the Healthy People 2010 took full advantage of the energies of the Department, as well as the insight from outside the Department.

What I am going to talk with you about is a fairly blitzkrieg enterprise to get the ball rolling, leading indicators for the year 2010, sentinel objectives, if you want to use that term. I will go through very quickly some candidate sets that were identified as part of our work, and leave a few minutes for questions at the end.

In effect, the activity of the sentinel objectives work group is summarized in this question. That is, can a relatively small set of exemplary health indicators be identified which will reflect progress for the health goals of the nation, and do so in a manner which prompts public understanding and policy action related to the important determinants of that action? That is the leading sentence of what you have in your report.

The key words here, I think, are these: understanding, action and progress. Those are the key elements of the task that we set ourselves to.

Understanding -- we want to identify indicators that enhance the public understanding of this process as well as the priority issues.

Action -- we want to identify a set of indicators that can help, by virtue of their focus, to mobilize more effective action around the country toward those targets.

Progress -- we want to have indicators identified that can give an indication of the overall progress of the nation toward the whole set of objectives, by identifying those that provide some sense of the progress within the whole group.

The short answer to this question, the work group feels, is yes, that there are ways in which we can identify a subset of the overall set of objectives that will help facilitate progress. I underscore in that comment the plural. There are a variety of ways that might be chosen. We felt our task -- and specifically our task as charged by Susanne -- was to come up with a series of options that could then be carried forward into a deliberative process that involved outside people.

I will describe in a little bit the role of the National Academy of Sciences in taking the ball where we have left it and carrying it forward, and then ultimately to ensure that there was enough flexibility in the range of options provided to the Secretary that the Secretary would have some leeway in the choices that she makes in consultation with you and the Council.

The notion of identifying a subset of targets which was to provide a general sense of the overall progress is not novel. In the 1990 objectives process, there were 226 national objectives established. Antecedent to their issuance in 1980, actually in 1979 in the Surgeon General's report, there were five broad mortality-based goals that provided an opportunity for focus.

Indeed, to the extent that there was a consistent focus over the course of the decade from 1980 to 1990 on an element of the objectives process, it was probably found in the goal for infant mortality reduction. The goal there was to reduce infant mortality by 35 percent, to a level of nine deaths per 1,000 live births by the year 1990.

The other goals, by the way, were to reduce childhood death rates by 20 percent, to reduce adolescent and young adult death rates by 20 percent, to reduce adult death rates by 25 percent, and to reduce the average number of sick days for the older population by 20 percent. Those provided an opportunity for a repetitive focus over the course of the decade, but an opportunity we didn't quite capture.

As I said, the major focus was probably on the infant mortality target, which we did meet for the population as a whole. It fell short for minority groups.

In some ways, the most striking accomplishment on these particular targets -- which has gone largely unrecognized -- was the target for adults, reducing in a decade's time the death rate for adults by 25 percent, to a level of 400 deaths per 100,000 adult population. The result at the end of the decade was 400.4, very striking progress over that period, again an indicator that we haven't taken full advantage of.

In the year 2000 objectives process, there was also an effort to identify a subset, again recognizing at that point that 226 targets was too many and, as you are finding here, for every target somebody had to take away, they had two more to add to the list. So, we ended up with 319 objectives for the year 2000.

Because the goals for the year 2000 objectives had shifted from quantitative in nature to qualitative in nature, and therefore didn't lend themselves to the sort of sentinel indicator status that the previous five life-stage goals had, we identified in the context of the report a series of sentinel objectives. They were called representative objectives. There were 47 of them, a couple for each of the priority areas. Again, having stated them at the release of the report and in the body of the report, we then didn't follow up as effectively as we should have, or I should say, at all, with specific periodic reporting of how we had done in those dimensions.

In order to avoid repeating the missed opportunities of the last two decades, this process has gotten underway. Drawing upon the creative energies of virtually all the agencies around the table and some not around the table, the working group reviewed a variety of interesting indicator activities, both inside and outside the government -- the work of the Federal Interagency Forum of Child and Family Statistics, the work of the committee on 22.1, the committee within the department to look at common data indicators that could be collected at the federal, state and local levels, to track progress on certain key dimensions, and a number of others that have been undertaken around the country -- state-based activities to develop indicators -- North Carolina and Connecticut -- and some of the activities at the local level.

In fact, interestingly, just last week I was in California talking with folks in the healthy cities movement, which has drawn a great deal from the whole Healthy People process. You see the healthy cities and communities activities around the country using the setting of targets as an opportunity to pull together different sectors of the community and come up with an agreed-upon common agenda.

One of the folks who was in California actually had just been to Colorado and shared with me the community indicators of the Yampa Valley Colorado project. There are really dozens of these around the country. This expands a little bit beyond the health arena. In Yampa Valley, they have decided that one of their important indicators of progress is the number of elk in the community.

In fact, this is in many ways what this is to trigger. It is a little bit beside the point of this discussion on national targets and indicators. But what we also want to bear in mind is that what we are trying to do is stimulate activity at the local level that can help people identify their own priorities and, therefore, stimulate their own progress.

DR. SATCHER: Do they want to increase the number of elk or decrease the number of elk?

DR. MCGINNIS: You can imagine, where there is a strong hunting constituency, they are actually seeking to decrease the number of elk.

In any event, looking at the wide range of indicator activity across the country, we found that there were many, many models from which to draw. I won't go through this list in detail, but you can see there are about 15 different models, into which these various activities could be grouped.

A few examples. The mortality model is pretty straightforward. That, in effect, would be a model similar to the 1990 objectives, where you pick key causes of death and say those are our foci and that is what we are going to track and follow very closely.

Another would be the disparities model, which identifies outlier conditions for those in minority populations. That is to say, where is the disparity greatest between the general population and certain minority groups, or ethnic and cultural, low-income minority populations?

Another would be the leading contributors model, that is, focusing not just on causes of death, but what the factors are that lead to those deaths -- tobacco use, sexual behavior, physical activity, nutrition, environmental exposures and so forth, in order to focus attention on the issues that are actionable to some extent.

Another would be the index model, which takes a variety of indicators and, through some calculus, mathematically aggregates them and comes up with an index that allows a comparison from area to area.

Another would be the theories of change model, where you identify key issues that help areas or activities to move from one state to another. Where are the pressure points?

There are a variety of other approaches that are represented in these different models that we assessed. Based on that review, we came up with five candidate sets for your consideration, and also for the consideration, and as sort of a baseline, for the National Academy of Sciences endeavor to follow. Those sets are listed here and they are included in your red book. They are identified as approaches, which is really what they are. It is a series of approaches to the establishment of leading health indicators.

I should say that the notion, the terminology leading health indicators, is one in itself which was the subject of great discussion. What is the right thing to call these?

Obviously, we abandoned along the way, at least as our place holder, the term sentinel for a variety of reasons, and chose leading health indicators as a replacement place holder, because it was the closest to the model of leading economic indicators that had served as the focus of much of our discussion. The leading economic indicators have made their way reasonably fully into the public consciousness, if not in detail as to their contents, certainly in principle as to their existence.

The candidate sets can be grouped into two areas. The first three are approaches that are oriented around the indicators themselves. The last two are indicators that are oriented around the reporting mechanisms, the periodicity of the reporting. It is the reporting process that, as a strategic issue, serves as the motivating element of the framework.

Let me just go through the content of each of these five candidate sets or candidate approaches very quickly. The important issue is not the actual targets that are included in these approaches, but the nature of the approach itself. That is what you will need to grapple with over the next 12 to 18 months. I should mention that in each of these cases, the assumption is that the target that is actually taken will be taken from the full set of objectives. It will not be a target which is apart from the broader set of objectives.

First is the health status model, which was, in essence, a model drawn directly from committee 22.1. The committee 22.1 refers to the committee that was established under the leadership of CDC and the National Center for Health Statistics in order to help implement objective 22.1 of the Healthy People 2000 objectives, which calls for a common set of indicators that can be tracked at the national and sub-national level relating to the progress in Healthy People.

There are objectives in five different categories that have been identified by committee 22.1 -- years of healthy life, promoting healthy behaviors, protecting health -- that you see here. Again, without belaboring the content, you can see the years of healthy life has some surrogate issues identified as candidate risk foci -- lung cancer, breast cancer and cardiovascular disease deaths.

Behaviors -- a variety of behaviors that are important to health outcome to which attention would be drawn. For health protection, there were vehicle injuries, work-related injuries and air quality exposures.

Two other elements of the health status model are those related to assuring access to quality health care -- again, getting closer to the issues that John was talking about earlier -- as well as strengthening community-wide prevention programs. You see here the items identified as foci surrogates for the access issue, ranging from infant mortality to immunizations, cervical cancer screening, primary care linkages. These numbers you see in parentheses, by the way, are references to objectives that currently exist in the Healthy People 2000 targets.

Then strengthening community prevention activities -- the two issues that were identified by 22.1 in that area were homicides and childhood poverty.

DR. SATCHER: Can I raise a question?

DR. MCGINNIS: Yes, please.

DR. SATCHER: I am curious about how you got homicide under community prevention and suicide under access to quality health care.

DR. MCGINNIS: The assumption in this case was -- well, first, it is fairly arbitrary -- but the assumption was that there are instruments that can be applied in the clinical setting that will identify depressed patients, less likely to be effectively applied in identifying homicidal patients. It was assumed that that was a relevant issue for access. The taxonomy is obviously subject to great change.

DR. SATCHER: It is also something that we ought to talk about sometime. More and more, especially among teenagers, there is a question as to whether some of the risk factors for homicide and suicide are not the same -- hopelessness, and whether, in fact, some cases of homicide are not suicidal ideation being played out. It is just something we can talk about.

DR. BRANDT: Homicide and suicide going together, occurring at the same time --

DR. MCGINNIS: The next model or approach that was identified was the health disparities approach, which would have us focus as our motivating framework only on those targets for which there were substantial disparities between the general population and special populations which were either economic or geographic or ethnic minority populations. What you see here is a list. Actually, there are two other pages, that I will run quickly through for this model, that represent those areas for which there is a greater than 25 percent disparity between the general population and the rest of the population, again, grouped into the same categories that we had for the previous slide.

Years of healthy life, healthy behaviors -- I don't feel guilty about not going into detail through these, because, again, they are all in your report. You can look at them more thoroughly at your leisure.

Protecting health, assuring access to quality health care, and finally, the issues of strengthening community prevention -- again, there is the focus on the high/low gap for the experience in each of those areas.

The third model or approach that the working group has identified for your consideration is a combination of two of the models that were identified earlier, and that is the summary measures, or aggregate measures, and leading contributors approach. In that, a regular approach would be provided, let's say annually, on indicators in part, which represented leading contributors to health outcomes, such as tobacco use, overweight, sedentary life style, alcohol, immunizations, air quality exposure, firearms, condom use, motor vehicle deaths, illicit drug use, depression, the uninsured population, high school graduation rates, children living in poverty.

Each of those has a predictive element, to some extent, for the health fate of the population. In addition, there are some summary measures that are aggregate in nature and give an indication in the aggregation of the state of the population -- years of potential life lost before age 75, hospital days per 100,000, and reported disability for population groups. Reported disability or reported health status is probably the best predictor of the probability of death in a given year. Self-reporting seems to be a more accurate indicator of the likelihood of problems or of death than physician prediction. In any event, you see the general intent is to aggregate and identify these 17 or so foci for annual reporting.

That eases us into the notion -- the term annual eases us into the notion of the last two approaches, and that is the reporting-oriented approaches. The first of those is a monthly report approach in which, through great wringing of hands and gnashing of teeth, you would come to a conclusion about 12 specific targets that were the most important. Those would then be reported seriatim every year. In January, the focus would be, for example, on infant mortality; in February, on cardiovascular disease deaths; in March on alcohol and other drug use. Obviously, both the content of this list and the association with any particular month is completely subject to your discretion.

The key point is that, strategically, under this scenario, the assumption is that having a regular focused report that the press can expect, gear up for, and that the agencies can home in on, would provide an opportunity for educating the public in a more effective manner.

One of the things that David put in an observation earlier that isn't in here that I think would be a good candidate -- I would certainly vote for it -- is depression as an important issue for the health of the population. Again, you will have lots of opportunity to angst over that issue.

Then, the last approach was the quarterly or semi-annual report approach. It would have, on a quarterly or semi-annual basis, a report issued in a very public fashion toward the whole set of these. I think we had a total of around 20 or so indicators in three areas -- health status indicators, health risk indicators, and then the third list over here is a set of health service indicators.

There are substantial data challenges in this approach that need to be underscored, and I am sure will be in the course of your deliberations. The notion of coming out on a semi-annual basis, not to mention a quarterly basis, with timely information on this full set of issues in a manner that is analogous to some extent to the leading economic indicator approach is a substantial challenge. If it is possible, it could be a powerful statement on a regular basis.

That is what you have before you in the red book with all the supporting materials. I will just wrap up with a summary of the critical issues that the working group suggested you give special consideration as you are sorting through the options. The first is, obviously, as a primary motivation of this exercise, the need to ensure that it improves public understanding. Obviously, indicators should be chosen that really are reflective of the most important health challenges.

The reliability of the data sources underpin the reliability of the whole process. This has to be a credible exercise. The robustness of the indicators over time is important as well. We have to anticipate that these will be indicators that can be, and will be, collected not just now or next month but next decade as well.

The relevance of the indicators to the program -- for the very reasons I heard being addressed around the table a few minutes ago. The utility for sub-national comparisons can provide stimulus to activity at the state and local level that will allow them to identify their own priorities and do so in a fashion that will facilitate comparison among geographic areas.

In wrapping up, let me just give special acknowledgment to Linda Bailey, who was the heart and soul in staffing this effort, and special thanks to all the members of the working group, many of whom, as I said, are around the table. I would also introduce two members from the National Academy of Sciences, the Institute of Medicine, Kathleen Stratton and Carol Chrvla, who are right over here, and will be carrying this activity forward.

I actually have a handout that I will pass around the table to summarize the process from this point on, as it moves to the Academy and the committee they are pulling together.

DR. SATCHER: Thank you very much. Comments, questions?

DR. EISENBERG: Can you connect how this relates to the quality of care agenda that the IOM is undertaking? I understand that Janet Corrigan is moving over to work on the health sciences activities. Are these two projects connected, or are they going in parallel?

DR. MCGINNIS: I will actually defer that to Kathleen, but I will lead up to an intro by saying, even though they are contractually parallel, clearly they will be informing each other along the way. I know it will be a great effort to ensure as much consistency as possible in those activities.

MS. STRATTON: I think Mike probably answered that question as well as I could -- which is that we are separate units.

Janet Corrigan will be coming on board to run a health care services, health promotion and disease prevention unit. We do hope to work together. I think the questions are very different, but I think when it comes to the health care quality issues that come before our committee, we would depend upon her for advice and suggestions and review. Likewise, I hope that when their effort deals with public health and population-based things, she will come to us, and I expect will.

They are separate, even conceptually, in terms of the projects and time frames.

DR. SATCHER: Other questions or comments?

DR. SONDIK: I think this is a terrific report. I read it word for word, and I really think everybody did a wonderful job on this thing. It has got a lot of insight in it. It really should help enormously in helping us to understand what different measures can give us.

Do you think it is feasible for us to use more than one set? For example, I was taken by the disparities measure. I mean, this came up earlier this morning, as to how we can deal with that. We don't really have a lot of great ideas on the table at this point, and this could really point us in the right direction.

DR. MCGINNIS: I do think it is feasible to use more than one set. In particular, if you select a small enough group of categories, then when you report -- the periodicity in which you decide to report -- you can focus not just on the national aggregate piece, but on the disparity.

I think you can set targets that are disparity- based as well. In fact, I think there is a certain inherent obligation -- because that is where our greatest shortfall is -- to do that.

DR. SONDIK: I also recommend that people look at something you didn't talk about, which was section five, on the available data sources. It is a very nice summary of that.

Also, the background material, appendix material, is very, very useful. It gives you an insight into how this is used and reported on. Particularly, the information from North Carolina, I thought was very interesting and very useful.

DR. MCGINNIS: Thanks very much for that comment. It actually reminds me that I should point out to you, for those of you who are intimidated by the weight of this, the report itself is really only in the first 40 pages. So, it can actually be read word for word.

DR. SATCHER: Thank you very much. I think we are going to take a short break, 10 minutes.

[Brief recess.]

DR. SATCHER: We have two more very important presentations. One is from the Business Advisory Council.

You have probably noticed that we have changed the schedule a little bit. We thought you would probably like to end by talking about scenarios for the future. We are going to do it that way. We are going to have Lauren Leifer talk now about the Business Advisory Council perspective, and then Clem to finish up.

Back to Table of Contents