Morning Session 2 - Transcript

Agency for Health care Research and Quality
National Advisory Council (NAC) Meeting
Friday, July 13, 2012
Health Care Costs

CAROLYN CLANCY: Well we are going to get started. So first I have a request for all of you. When you’re speaking if you could get really close to the mics. This is just for the benefit who spent way too much time at rock concerts as younger people, although it is helpful for us. But it’s also for the Webcast. Some people are reporting that they can’t hear what people are saying. So just be robust, be proud, get right close to that mic.

So I’m now going to turn his over to two of my terrific colleagues. This is a session, which is a bit—it’s a bit misleading in being labeled “Health Care Costs.” It is about an incredibly important dimension of health care costs and expenditure, namely the concentration of medical expenditures. But I wanted you to be aware that there is a lot of other work that we do at the Agency related to costs. It is something that we are already hearing a great deal about, writ large, but we will be hearing even more about.

So you can count on future sessions at these meetings focusing on other aspects of health care costs. And with that I’m going to turn it over to Dr. Steven Cohen and David Meyers.

STEVEN COHEN: Carolyn, thank you. And I also want to thank Jane Crowley for suggesting this as a topic at the last meeting of the NAC. So we wanted to be responsive to that request. In terms of the significance of the issue, more than one of every six dollars is going to health care right now. And the rate of growth in terms of health care expenditures is clearly exceeding other sectors of the economy.

However, in the last couple of years, we’ve really seen a recent moderation in the rate of growth and that is very promising. In terms of the distribution of health care expenditures, that distribution is highly concentrated in that a very small segment of the US population are tied to very large health care expenditures. And we’ll be talking about some of the characteristics of those individuals. Health care expenditures continues to be one of the largest components of federal and state budgets. And cost containment continues to be a major concern for public and private payers.

In terms of the most recent statistics, overall expenditures for health care is $2.6 trillion dollars in the nation. It is actually 18 percent of GDP. That’s almost a four percent increase over the prior year. Growth is much lower than the prior decade. Again, that’s promising. Per capita that’s around $8,400 dollars. And with projections from our colleagues in CMS in the next decade, with the aging of the population and other factors, it looks like things are moving to one of every five dollars is going to health care—so clearly, a major consideration for public policy issues.

I’d like to now draw your attention, and many of you are familiar with this survey. But I love talking about AHRQ’s commitment to this important national data resource that’s incredibly important in informing policy and practice in terms of the issues of health care expenditures, trends in expenditures, and distributions and expenditures. The survey, actually, is a family of surveys. The core part is a household component where we have roughly 14,000 households, 32,000 individuals each year to get detailed information on their health care utilization, their expenditures, their access, their insurance coverage. And we have some quality metrics as well.

We actually have this as an in-person interview. There are five data collections to get two calendar years’ worth of information. And it’s actually a longitudinal survey. It comes from the department’s Health Interview Survey. We take a national subsample and then we have two years of longitudinal data to get details on the dynamics of changes in demographic composition, changes in employment, insurance coverage, where we could see the impact on health care expenditures, utilization and outcomes in treatment protocols.

Another dimension is that, rather than just aggregate statistics, this really is the survey that helps you look at all Americans and takes a rank ordering of their expenditures and helps you see the concentration of expenditures. In addition to the concentration, given that the survey’s longitudinal, you could see the persistence of the concentration of expenditures over time.

And being able to drill down for specific conditions has been a highlight of the survey. As Carolyn mentioned, our collaboration with the National Cancer Institute, the CDC in using this platform to enhance a representation of a group of individuals with chronic disease—and actually not only enhance sample but enhance content to get more details on their use expenditures and treatment protocols.

Now, in order to supplement the quality of the data on health care expenditures, for the household sample, we get permission to go to their medical providers. We go to 100 percent of their inpatient stays as well as all the separate billing doctors that are providing care in an inpatient setting. We also go to a nationally represented subsample of their ambulatory care providers. We go to all associated home care providers and all associated pharmacies.

We do this to get details on the dates of their visits, their diagnosis, procedure codes, charges, sources of payment, so that we improve the expenditure estimates for the survey. And it’s really the key imputation source for the remaining, missing data.

And to complement this resource, we have a separate, independent establishment survey every year, going out to 40,000 establishments to get detailed information on health care premiums and the availability of insurance coverage to the workforce, cost of coverage in terms of what the premiums are for the employees and the employer, and the benefit provisions.

And so with this rich, analytical tool, one could look at the impact of behavioral and economic factors, payment and health services demand on utilization and expenditures over time. We have the opportunity to go deep into the distribution of expenditures, the concentration, its persistence and, again, this focus on patients with multiple chronic conditions.

And a flavor of some of the impact of this resource the Agency produces, in terms of some of the premium data from the establishment component, it’s used by the Bureau of Economic Analysis to frame estimates of GDP for the nation, in particular, increase in premium estimates. The White House, the CBO, the Congressional Research Service, Congress and Treasury, frequently tap staff for some of the sentinel findings in terms of trends in coverage and cost and implications in terms of disparities in health care.

The data is extensively used by the Government Accountability Office to frame employee compensation issues. And as we speak right now, as the insurance component of data comes out on premium estimates and the survey, I might not have said, in addition to national estimates, it provides separate state level estimates. As part of the accountable, the Affordable Care Act, the acronym with Accountable Care organizations sometimes, but the Affordable Care Act is one benefit for small employers, making insurance available for their employees. And the get up to a 35 percent tax credit on the premiums.

And the Secretary of Health and Human Services with the IRS take the MEPS premium data by state for the average family premium. That said, is the mark for the states to actually provide the tax credit for small employers. So we are seeing a lot of use in terms of the data, particularly on the cost dimension.

More recently, some of the estimates from the survey on the concentration of expenditures makes the press—and the additional tidbit here is that in addition to knowing that one percent in 2009 were tied to 22 percent of the expenditures. If we look at the mean expenditures for that group, it was approximately $90,000 per capita. And even if we go down the top five percent, that was approximately $36,000 dollars. One caveat, the Medical Expenditure Panel Survey is for the civilian, non-institutionalized population. So we are not covering the institutional stays.

So, now talking to the concentration of expenditures and looking historically at AHRQ survey in terms of informing this issue, if we look back several decades to 1977, you could see the top one percent was tied to almost 30 percent of expenditures. And over the next 20 years that pattern held up. But as we move to 2008, there was modest diminution in the concentration of expenditures. And there are a number of papers that have come out from our staff in health affairs, attributing to this decline partially to the aging of the population and more pervasive use of prescribed medicines and other diffusion of other services.

So we’ve seen that modest decline. Yet, 20 percent of expenditures is pretty significant. And again, the top five percent, roughly 50 percent -- and the top 50 percent, 97 percent of expenditures. Some of the characteristics that align with looking at the concentration of expenditures would be individual with chronic conditions. And it’s really the complex patients that are really at the tail of the distribution, where not just one but it’s a whole syndrome of particular conditions.

Yes.

FEMALE: Before you leave this topic, can I ask you a question? Is it possible that part of the, one of the reasons that there is the decline, if you will, in the proportionality of expense is due to more use of services for otherwise healthy people—and overuse of services, multiple services and testing and things that some people might consider not adding clinical value. So the distribution looks in a way, better. But it’s in fact because we are spending more money, much of it unnecessarily.

STEVEN COHEN: Again, there’s so many competing dimensions of this. But that, certainly, is one aspect. If one looks at a segment of the population, because there have been analyses on the concentration of expenditures for the Medicare population. And you see a population using services in a more uniform way at higher levels. So they are nowhere near 20 percent of overall spending. So there are so many different factors going into that.

But it’s interesting to see this rather high rate for the last several decades and then more recently it’s around 10, 22 percent. So that’s a signal in terms of the overall conditions that are driving the differentials. But certainly we could explore that a little more.

FEMALE: Thank you.

STEVEN COHEN: Oh, sorry. So this, again, drives home the distributional factor. But here you could see a little more clearly that the bottom 50 percent, in terms of Americans and their spending are tied to only three percent of expenditures. But it’s important to know that you can’t just look at this for one year. Somebody who is in the bottom 50 percent could easily, with some sort of inpatient stay be in the top 10 percent in the next years. So it’s important to have a greater context than just looking at this one year and having the longitudinality to help improve the understanding of what’s going on.

A little bit more attention to the impact of expenditures for care for individuals with chronic disease—if you look at individuals with heart disease, of the overall expenditures in 2009, they’re spending almost $100,000 million dollars. Now, the next bar looks at the top five percent of the population and their expenditures for heart disease. And you can see they’re using approximately three quarters of the overall expenditure amounts for treatment for heart disease.

So there is this overlap in terms of the concentration of expenditures and individuals in the tail, certainly disproportionately consuming a higher component of overall expenditures. You see that for cancer care. And while it’s not three-quarters, it’s more like 50 percent for treatment of mental disorders or trauma or COPD.

Another analysis shows, as you move up from a given year, the number of treated conditions from zero or one, two to three, four or more, you could see a pretty radical step function going in terms of something like $2,400 dollars per capita up to $8,500 up to $16,000. So if you look down in terms of age distributions, particularly those four or more, you could see, it’s not really age differential, it is really the complexity in terms of if you go across the different categories, if you wanted to see what was salient in driving expenditures.

Now, framing this in terms of a longitudinal look at the concentration, what this analysis does is looks in the X-axis, looking at there a person was in terms of their distribution and health care expenditures in 2008. And as we go up the wires, it’s their position in terms of the persistence of expenditures. So those in the top one percent, 20 percent of them retain that ranking in a subsequent year. And if we go up to the top 10 percent it is roughly five percent.

So there is tremendous predictive capacity in where you were in a prior year. But, at the same time, there are transitions out. And it’s very important to do research to look at just like I was pointing out, those at the bottom 50 percent going into the top 10 percent. If we are looking for effective strategies that show efficiency, if we could pay more attention to those, say, in the top five or 10 percent, dropping to the bottom 50 percent in a subsequent year. And the treatment protocols and other characteristics that help achieve that, that certainly would be a signal for some promising activities.

Just to give you a flavor of some of the information that the survey uses to frame some cost projections, and the agency does have a division of modeling and simulation—so we take a lot of the rich data from the Medical Expenditure Panel Survey on socio-demographic factors, on health status measures. And the survey has taken the SF12 in addition to the SF1 to, you know, calculate health status dimensions.

And we have quite a bit of details on health insurance coverage in terms of coverage over the course of a year, details on health conditions, prior year utilization and prior year expenditures. And I could tell you if we modeled this without the expenditure variable, and then you do the alternative, just putting in the expenditure variable and not add in these other, very important predictors, that expenditure variable carries a tremendous amount of information. And other groups have seen similar patterns in terms of their projections.

So in terms of profiles for improvement, I’d just like to spend a minute or two on a very informative article that Atul Gwande did in the New Yorker where he talked about a physician, Jeffrey Brenner who was trying to follow what they actually did in New York with crime statistics. I believe it was Commissioner William Bratton who looked at where the hot spots were for major crime activity and put a lot of resources in terms of trying to minimize the incidence. And somehow this cascaded out to reducing crime in ancillary, approximate areas.

And so what he did was, in the Camden area there were, I believe, three hospitals. And he started actually mapping out where the individuals lived who had high utilization in terms of inpatient stays. And it turned out a very large component of the population were in two housing projects. And what he did was he turned over these patients to the Camden Coalition. And actually, they don’t even have a clinic.

But what they did was they made many phone calls and they had like in house visits. And they hired social workers. And they looked in terms of some sort of an impact on their top, 36 super utilizers. And apparently, before the impact of the care that was being received, these super utilizers were averaging 62 hospital and ER visits per month. And after the program it was down to 37 visits. Many of them were using the ER rather than like an office-based visit. And you could see a reduction in hospital bills.

Other areas of research relate to geographical difference where you made statements about a lot the health care is local. And some of the findings in terms of per capita Medicare expenditures, in an area like McAllen, Texas as opposed to El Paso or areas around the Mayo Clinic. That also provides signals. But more research in terms of finding the next hot spot. And the question is, can we really lower medical costs without hurting the value or the quality or the effectiveness of care, giving the neediest patients better care.

So, a lot of attention at the agency has been directed to our care management and prevention portfolio, patient safety portfolio. In the Affordable Care Act there is promise of Accountable Care organizations. The Agency has focused a lot of attention to reduction in medical errors. And I’m going to turn it over to David to say more about that.

DAVID MEYERS: Thank you, Steven. And thank you all for inviting me to be with you today. Dr. Clancy, in thinking about costs and really a separate issue for her of great importance of, how as a health care system do we improve health care for people with multiple, chronic conditions—invited me to participate today with Steve, to give you a little bit of flavor of the actual activities we’ve been working at her at AHRQ on improving care for people with multiple chronic conditions.

Before I go into that, though, it was important to say that we’re not doing this alone. The Department of Health and Human Services has a strategic vision and framework that I hope some of you are familiar with, and now all of you would be—that laid out four critical areas for us to work collectively on. And to some degree, AHRQ is taking our role from the bottom up here. We are leading research in—and you’ve already heard that we are leading data. But we are also doing research to fill the knowledge gaps about the clinical side and the system level side and building better tools for the clinical system to use.

[Next Slide]

The home for our work in multiple, chronic conditions has been our Prevention and Chronic Care Program. And what I’m going to do for you right now is try to turn our focus on this upside down. We’ve been motivated often to approach this because of these stories like Atul Gawande’s about those super users. And there’s a carrot out there that says, if we could just improve care for these people taking 60 visits a month, 20 people making 60 visits to the hospital a month (That’s impressive!) could we change the cost curve?

Well., we clearly could improve the quality curve. And there are probably something we could do for cost. But we want to suggest that the long term solution here is a bit broader. And so I’m going to share with you, coming out of our prevention focus, how we are thinking about multiple chronic conditions and improving care, using a primary, secondary and tertiary approach.

So let’s go first [Next slide, please]—

Primary prevention means stopping us as a country, not develop chronic conditions in the first place. AHRQ’s work in this portfolio includes working with the US Preventative Service Task Force and help know what are the most effective, clinical services, and then help people get them. We’re actually developing new metrics about the use of clinical, the use of appropriate clinical preventative services for all Americans, recognizing that if we can’t measure this we are not going to be able to improve it.

And we have also recently started using funds from the Prevention and Public Health Fund. Three Centers for Excellence in Clinical Preventative Services focusing specifically on how can we use these services to decrease disparities in the US within health care—how do we make sure that these, that harms are reduced and that thinking about safety and prevention as well in underutilized area, and also implementation in primary care and community centers. The three centers focus on those three areas.

[Next slide]

I posit to you, however, that the place most in need of attention and truly the place we’re going to make a long-term impact is on slowing the progression of those who begin developing chronic conditions, to stop them or slow, at least, their rate of going from having one chronic condition and having that accelerate, bringing them into many others.

We know that these chronic conditions, people living with multiple chronic conditions tend to cluster. It’s the person who, as they age, gain a little bit of weight, develop high blood pressure, then develop diabetes. Those two combine to give them heart disease. They have their first heart attack. Their exercise goes down. They now have arthritis. It starts a cascade. And that the long-term solution is helping people back away from that cliff.

And to do that it’s a—we believe it’s a needed focus on primary care as a foundation. And that integrated primary care system, with the rest of the health care system, is one way, in the long term, and this would be called secondary prevention—taking people who are beginning to develop chronic conditions and empowering them in their own health care to better manage within the context of their family and communities, their condition—so that they have longer, healthier lives.

We’re doing that here at AHRQ by focusing on the patient-centered medical home. I’ve been with you before. I’m happy to talk later about research, our evaluation, our implementation support, and our work in bringing together our federal partners to have a coordinated effort in this. We are also doing new work on care coordination between patients and primary care, within the primary care system and primary care as part of the medical neighborhood, with the rest of the health care system.

And recently a new focus on team-based care, recognizing that the old model of physicians doing everything for everyone, is not the way we help people with multiple, chronic conditions. There has to be a focus on how a whole set of skills are brought to helping people, including community services, the role of nursing in primary care, pharmacy in primary care, and bringing their specialist expertise to help patients and whatever their needs are.

[Next slide]

That said, we would be remiss to recognize that there are people in America and many of them already living with multiple chronic conditions. And it is going to take us a while to get the benefits from that primary and secondary prevention. And so we have started here at AHRQ beginning in 2008 a research network of grantees focused on how do we improve care? How do we improve the health care system for people living with multiple conditions?

We have three goals for it. [Flip to the next slide, please]

This work began in 2008. Actually, some of the first work in the department, this is pretty amazing. We looked at NIH, CDC and our other federal partners and very, very little research is done in the messy area of people who aren’t simple. And we said, could we start some new methods in this? Could we incite the research community to turn their attention to these folks. Eighteen exploratory grants, funded under ARA, generally three to four year grants were funded.

That was so successful and that community said, “We can do more.” We’ve followed that up with a series of announcements including further exploratory grants and 13 infrastructure grants, again, funded on—these are actually the ones funded under ARA. The first were actually appropriated funds from Dr. Clancy. The ARA grantees are looking at creating datasets, that are multiple use, that will be made publicly available for health services researchers to look at complex people with multiple conditions, focusing on mental health, diabetes, cancer, heart disease and many other combinations.

And to put all of this together, we’ve put in place a learning network among these grantees and a technical assistance center, with the idea of, that we, at AHRQ, we hope to compile this information to keep the field moving forward. I said these were three—

The first grants [Next slide, please]—

I said the first grants were three to four year grants. So the early results are just starting to trickle in. Doctor—I’ll highly one here for you. Dr. Gross and his team at Yale University were working on how prevention, and a lot of our early grants focused on incorporating preventive services for people with multiple, chronic conditions. Often once people have other health conditions, their preventive needs fall of. And at other times, when we should no longer be thinking about preventive services, certain (?) preventive services, they continue to get them when they don’t have a chance of benefitting.

So a lot of our initial grantees focused on this combination of chronic conditions and preventions. He looked at colon cancer screening, specifically colonoscopy. Were we using it effectively? His team found that there were substantial number of Medicare beneficiaries receiving screening even the harms clearly outweighed the benefits. And their work, which will soon be in publication will be suggesting ways that health care systems can better target colonoscopies to those most likely to benefit, including those who haven’t been getting it and should and those who have been getting it and shouldn’t.

So with all these important studies and promising initiatives there are still additional challenges that we are confronting in terms of the proportion of individuals with multiple, chronic conditions, is likely to continues as a continue as a consequence of the aging of the population and the obesity crisis in the nation. And another dimension of this issue is that, think of the family that has expenditures that are $5,000 dollars a year.

And so I talked about per capita being an overall, $8,400. So relative to that, that doesn’t sound like it’s a large component. But say the family income is something like $40,000 dollars. Issues in terms of burden of health care expenditure and affordability is something that is a critical dimension to be looking at, at the same. The priority populations for the agency clearly interact with some of the priorities on opportunities for potentially improving the effectiveness and efficiency for care for the high utilizer.

I would like to also point out, again, as Carolyn did, the agency does have a Value Portfolio that looks at the nexus of cost and quality to try to improve value of care for the nation. And there are a number of investments in that particular area. And the agency has another set of rich, analytical tools going very, very deep on the cost dimension for inpatient stays and emergency departments, being the HCUP (?) project, the family of data acquisition discharge records that they have, quite invaluable for advancing this type of research.

So thinking broadly in terms of where we potentially could go, issues that we would like to get some insights from you on how enhanced longitudinal profiles might help us advance some improvements in this area, what priorities should be placed on research initiatives and modeling efforts to help identify strategies to improve health outcomes and reduce expense for this population. So as we turn it back to you, we’re hoping to get some engagement from you in terms of really, where are the opportunities for the Agency, what is the role, the appropriate role for the Agency. And if you have some particular recommendations for data enhancements, research initiatives or modeling efforts, that would really be very, very helpful.

BRUCE SIEGEL: Thank you very much for the presentation. And we’ve been, I think, asked some questions to help guide the agency going forward. And so I’m going to start with Dr. Johnson and move around the table. And I’m going to ask everybody to please ponder and try to address these questions.

MIKE JOHNSON: It’s great. Thank you for the presentation, it was terrific. I’ll start by confessing that I used the MEPS for my doctoral dissertation and I’m very indebted. And it’s a delayed thank you to the folks at AHRQ who got on the phone and sort of helped me understand the data. [Laughter] I can’t tell you how much I appreciated that.

One of the things in spending time with the MEPS was just the idea of how much we can look at in that dataset, that it is really tremendous. And I looked at data from ’97 to 2000. And as you got a little further along with data, the SF12 got added.

So my question is, which I was really happy to see—so the question and maybe the opportunity here is, much of the expenditure that you report is related to or revolves around disease. So is there a way to add to that and capture function and the impact of function because does disease drive down function? And if so, can functions be the way back out of disease? We’ve seen it with diabetes and so forth.

Are there things, are there data elements, in addition to the SF12 that have been added, that really capture function, be them obviously self-reported or through the medical provided piece that you’ve given thought to—and is this an opportunity maybe to enhance the dataset and our understanding of population management in health.

STEVEN COHEN: We always take a look at the survey content and actually are self-administered questionnaire was brought about by the Agency’s role in producing the National Quality Report and some of the metrics that could be added to the survey. And in that process, we actually had a measure called the IRQAL-5D (?) in terms of some quality metrics that we put in at the time. There are opportunities where we re-assess, particularly in terms of the needs of the Affordable Care Act, what should be added to the survey.

Now, there is only a fixed amount of time we have. It is very costly and it’s burdensome to the household respondents. So if they are very clear-cut evidence based measures that clearly will add value that answers a very targeted series of questions, that certainly rises to the top. And then it’s the issue resources and being able to put that in the survey and potentially take out elements of the survey.

And we’ve had meetings within the Agency when we’ve had radical changes, like the quality dimension or when there are opportunities for collaboration with other departmental optives (?)—where, as you saw, the National Cancer Institute partnered with AHRQ, where for a one-time sample enhancement and a one-time, self-administered questionnaire on cancer survivors that was possible.

So there are avenues for those modifications. But there is a very big process in terms of viewing and prioritizing and assessing what’s the best changes to make.

MIKE JOHNSON: And spending a lot of time with that dataset, it is—I appreciate the households that participate because it’s a lot of data that they provide. Just one example that I would think of is with back pain, and the costs associated with back pain—looking at, utilizing something like the Oswestry Disability Index or the Roland-Morris as a way to capture some of the data and begin to look a little bit more fully at the population, just as a thought. But thanks very much.

BRUCE SIEGEL: Dr. Thompson.

JEFF THOMPSON: On the, we call it the 5-50 population, the five percent that spends 50 percent of the cost. But if they are always identified when they peak (?) out, you know, how do you sort out regression of the mean versus programmatic effect? I think that has always been a challenge for us. You know, I mean, a lot of times we are doing descriptive work that we have patients that go anywhere between 100 and 200 times to the Emergency Room, albeit there is just a recently reported article from Rutgers that said, “Medicaid is no different than employer populations when you look at symptoms in urgent versus non-urgent care.”

But from a strict cost standpoint, I mean we can be descriptive, is there any way of teasing out, as we look at the 5-50 or these types of patients, what the cost impacts could be.

STEVEN COHEN: No, again, the regression to the mean is definitely some of the transitions back down to lower gradients. What we’ve tried to do, and because this is really essential to have larger samples for people with specific complex conditions, we pooled several years of MEPS data and then we’ve inflated up. And to the extent one can see holding constant, doing very detailed risk adjustments, where there is still very significant differentials—that is certainly a signal in terms of where care, you know, care protocols make a difference and still impact on the differentials.

But, again, it is really the rich sample size and the precision to do that. But there is certainly opportunities with this data resource. And there will be with this All Claims Payer Database [00:38:42] the Department that was part of the ARA initiative. There is a lot of promise where. So, stay tuned.

MALE: We’ve recently done a randomized trial with Medical Home at Harbor View. And there was really not—the mortality, actually improved but the cost wasn’t over—the cost savings wasn’t over administrative expenses. We’re going another randomized trial with kids at Children’s Hospital to replicate the Milwaukee model. But I guess from Carolyn, I mean what can AHRQ do? And it is a tough one. We really need randomized trials for these types of interventions, which grits people’s teeth. But that’s the only way to, you know, determine what’s regression of the mean versus true, programmatic effect.

CAROLYN CLANCY: Well, where we can do them, we do. And to some extent some of these analyses I think can help refine important hypothesis to be tested in randomized trials. I wouldn’t be feeling too bad about Harbor View, though, because the VA tried this in a randomized trial, oh, about 15 years ago. We weren’t calling it a Patient-Centered Medical Home then. And there was a lot of rich debate in the New England Journal about was this unmet need or did this mean, this enriched, primary care model just wasn’t worth anything. The debate was actually more interesting than the study.

Plus, it would also occur in the context of a lot of transformative activity within the VA system. So we’ll be debating this for a while, I think.

HELEN HASKELL: So, thank you. That’s excellent and it is a wonderful dataset. And, obviously used for lots of very important purposes. I’m struck by how often we think about these data and experience as if the system, in a sense is, or the patient is passive and there is—for example, that evidence of high expense means that person has more health problems. Or something that is going to be fixed by the health system.

So what I’m getting to is the idea of the role of the patient and the accountability, what we’ve seen on the employer side, especially with some of the frequent flyers is moving to a model that says an awful lot of what’s happening in terms of use and cost is manageable or can be influenced by individual responsibility and accountability. And so there are a lot of benefit design changes around that.

But to the AHRQ role and opportunities, I just wonder if it is possible identify more of those conditions and procedures where there is clinical agreement, that there is generally no value. I mean the colonoscopies over 75, for example. And the patients either shouldn’t be doing something or the physicians may be involved. Either they are doing it for one reason or another. They own the lab or they are doing something else. Or they are just in agreement with the patient because they feel the patient is asking them for something.

So you add all those things up and that’s where a lot of money, we know, is being spent. So research on and more understanding, including through survey mechanisms about how consumers or patients are motivated to be engaged themselves, among other things to even use, for example, the US Preventative Services Task Force recommendations about colonoscopy. So why are 75-year olds in the Medicare program getting colonoscopies at the rate they are doing it when the US preventative services task force recommends against it and except, obviously, in risk situations.

And what is the gap between when what we know is going on or should go on and what the patient should know about what is going on. And how do we influence that, whether it’s behavioral economics, whether it’s neuroscience, whether it’s social psychology or the latest is social norming. And to understand—and I realize this survey’s role is to be much more quantitative. But if it is asking SF36 or SF12, whatever it is and information and doing other research at the same time, this might be a good opportunity to learn more about what would change people’s behavior in all these domains, which apparently are just numerically very substantial.

And if we’re going to avoid, among other things, having charges when we begin to try to change overuse, for example, charges that somehow the people are not getting the care they should be getting—or they are being denied care. I even heard a story the other day from someone who was very smart, very knowledgeable and her comment was, “Well, my doctor told me that I should have this test in this way. But she also told me my health plan wouldn’t pay for it.” And her reaction, which is what is scary, “Well, then it must be good and I should get it and pay for it. Because they won’t pay for it, it must mean”—that is sort of pretty distorted thinking, I think.

But sort of to learn more about that, kind of, how do we change the transactions and the understandings of the public around all these questions because we are talking about—I mean your numbers don’t yet show that it’s actually $2.8 trillion this year and probably close to $2.9 trillion. So we are already $0.3 trillion further along than these data and moving very fast.

CAROLYN CLANCY: Helen, the one comment I would make is, having had a reasonably close relative’s wife call me recently, did he need another colonoscopy at 80. He had been directed to go get one. I don’t think he was dying for one particularly because at 75 he had some major complications that required surgery and he is on a blood thinner and all this. But actually, the Task Force recommendation, I didn’t say, “Don’t do it.” I wasn’t even going to go there. And there was a spousal disagreement here involved.

I simply pointed out the Task Force recommendation. And the patient picked up the phone and canceled. So I do think that some element of figuring out how do we get this information out. I’m optimistic about the new information from the “Task Force for Consumers: What Does It Mean for you?” But we can’t actually do that kind of mass distribution alone. We need to be thinking through who are our partners who could help and so forth. Oh, David.

DMALE: Great question. And those are very much some of them that we have been thinking about. Two parts to share. One, I think the message comes best, that the public is going to respond best when the clinical community takes the lead. The government taking the lead on overuse—we’ve been trying but not surprisingly running into some obstacles. But when ABIM and Consumer’s Union led the recent, Choosing Wisely campaign, it’s a sea change in our national discussion. So that social norming and helping—what we were doing was helping provide them with the data behind it, that they could then use to make their own decisions, each specialty society themselves about low value care.

And that partnership between clinical specialty societies, consumer union and other consumer groups is, I think, going to be one of the keys. Another piece of research we did—we did our own and then partnered with consumer union about how do you talk to people about making health care decisions—really great stuff about underutilized good services. Decent discussions about appropriate use, absolutely nonsensical, just crazy making discussion about overuse.

The discussion broke down. They didn’t have the language yet to talk about, what would it mean to choose not to do something because it might be bad within health. This was specifically focused on prevention. But we’ve seen it—so we have a lot of work to do on those social sciences about how to have a discussion about what not to do.

HELEN HASKELL: And I wasn’t by any means suggesting that the government should do it. And I’m thrilled, I think, that Chris Cassell (?) and others and Ardis and Bernie Rosoff and others who have led this movement on the physicians’ deserve a Nobel Prize in my judgment. But what I was saying that in the—gathering information, whether it’s to help people to understand, to get what they should get—I didn’t use my other example, which is, the beta blockers after heart attacks—we know that the rate—here is somebody who has had a life altering event, who does not take the medication. I think it is something like, only 50 percent, Carolyn probably knows, yeah—

So it is very low. So after a heart attack—so to use this mechanism, other research to understand how do we help people understand better? What is it they want to know because we know this on the marketing side. I mean Procter & Gamble and Target and others study this stuff all the time. How do you change behavior for marketing purposes. We don’t do nearly enough of that on the health care

Side. And if we have that collected in a neutral, long established way, then everybody can use it, whether it’s to encourage more use where there is underuse or less overuse where there is overuse.

But that’s what we don’t know nearly as much about as we need to know. And that does seem to me a legitimate role, just to understand it, just to have the data.

CAROLYN CLANCY: Dr. Spitzer.

ALAN SPITZER: I’d like to introduce a slightly different perspective. I’m full of personal anecdotes today. Here is another one. But this one involves my father-in-law, who is 87 and is a very intelligent man. He was an engineer who worked with NASA for many years and in an advisory capacity with the space shuttle. He was a prisoner of war in World War II, was recently asked to give a speech on Memorial Day that made its way into the Congressional Record. Congressman Barney Frank picked up what he spoke about and introduced it into the record.

And he is an individual who I would put into the high utilizer category right now but through no fault of his own. Until about three years ago my father-in-law, who had a terrific internist, who took care of everything for him—he has, just so people are aware, he has chronic myelogenous leukemia. He is in remission. He has coronary artery disease. He has arthritis, diabetes long standing. So he has multiple chronic conditions. And he took very good care of himself with his internist.

His internist retired three years ago. And he was place into another practice where now the medical approach of that practice is to send him to one specialist after another, on and on and on. He hates that. He does not like going to the physician. He does not like being a high utilizer. And it seems to me that we have an obligation in many cases to somehow influence the physician practices and physician behavior for patients who do want to take care of themselves and not see multiple physicians.

The number of physicians that he has gone to or my wife and I have taken him to in the last three years, is almost endless, not through his own choice but because he has a problem. He’s concerned about the problem. He would like to have it taken care of appropriately by his internist. And from my perspective, and I have to put in a proviso here, I’m a neonatologist. So I can’t help him too much with many of these problems.

But from my perspective, a lot of this does not need specialist care, could easily be handled by an internist. You know, is he has an upset stomach, he does not need to see the gastroenterologist right away, which is what the response has often been when somebody does have these multiple problems. His internist right now seems to be totally adverse to taking care of somebody with chronic conditions. And maybe he needs to find another practice. We’ve talked to him about that. And I think that that is a big part of it. But trying to find that practice and to try to get him into the right practice is almost as challenging an effort.

And so I think that while we tend to discriminate a bit against these chronic utilizers, I think there is probably also a population of people who find themselves in the circumstance that I’ve just described who are almost being forced to over-utilize the system in a way that they would rather not.

MALE: So very quickly, I’m right with you. That is non-comprehensive primary care. You are describing the perfect set up. And the reason we have that in this country, according to our work is that we’ve set up incentives for how health care is financed that has driven primary care further and further into that model. And we have hope and we are starting to see that when you change the training of clinicians and the workforce of clinicians with realigned incentives, they go back to their roots and become comprehensiveness, comprehensivists again.

ALAN SPITZER: I’d like to make another comment there based on my other experience in training young physicians. And that’s the fact that with the work hour regulations that exist now, one of the things that I’ve observed over the past decade or more is that physicians in training are assuming less and less responsibility for patients. They no longer refer to a patient as “My patient with arthritis.” It’s “The patient with arthritis.”

That possessive pronoun, I think is a very important one in the way we practice this medicine. And I think that the way that training a has been structured now in medical school and in residency is creating, in part, this situation of lack of ownership of patients. And of the general internist, more and more I think we are training a population of people whose response is, “You have an upset stomach, you go to a gastroenterologist. You have a problem with urination, you go to the urologist.”

It’s nobody working through the problems. And I think we are fostering that in our training programs as well.

FEMALE: That was such a good lead in. I actually had on my pad here I would like to focus on the issue of provider behavior rather than that patient behavior, not that we shouldn’t provide appropriate incentives. But so you do and then—so the patient goes to the correct, ostensibly, provider and then the provider is sending them to all the most expensive specialists and causing all kinds of over utilization. I think that it would—and, actually, another thing that you said, Helen, when you talked about the way that physicians are looking at their patient as sort of atomized groups of symptoms or whatever—I’d like to suggest or ask to what extent our AHRQ is looking at PCMH’s as a whole entity rather than a place for a person with a minimum number of multiple chronic conditions can get the right care.

So I’m coming from the perspective of a health care payer, where we are trying to use our dollars most wisely for our mostly low income beneficiaries. And we get these people coming to us saying, “Give me your top one percent and I can do a real good job for them.” And I say, “What about my other 99 percent, who also deserve appropriate care, who deserve not to be pushed into over utilization, who deserve to get all the preventive care and recognize when they are about to tip over into the upper 50 percent or whatever.

So to the extent that we think that a patient-centered medical home is a really good idea, maybe a really old, good idea whose time has come again or whatever, I think it would be great if we could really look more as that institution as a whole and see what works in those institutions for care across the board, both for the high utilizers and for other people who also need to access health care.

And then I would suggest that when you’re looking at the ability of a patient-centered medical home to control expenditures, that price has a tremendous impact on cost. And I think there is some evidence of that coming out in various studies. Health Affairs has just published an updated study of the Alternative Quality Contract in Boston and, you know, the reiterated their conclusion that I think the single largest driver of an actual decline in cost was physicians referring patients to lower cost providers.

There were other important factors. Reducing utilization was an important factor but not by itself as big as referring to lower cost providers. And so I’d just like to offer a sort of emerging little, naturally occurring experiment that we’ve got going on. We have a group of about 28,000 low income workers, home care workers, actually who we are trying to provide health care for. And we offered them a financial incentive.

We eliminated copayments if they elected to receive their treatment, their principal treatment through a designated PCMH level 3 or FQHC because, you know, we weren’t going to go out and make those judgments ourselves. So places had already been judged to provide immediate access, care coordination, all of those great things. Only a quarter of our members elected to receive their care that way. We tried to incent the individual behavior to do the right thing.

And, indeed, we are seeing, so far, it is only about eight months into it—we are seeing noticeably lower costs among the patients who are going to the PCMH’s. But not because they are not going to the Emergency Room. We actually see like 40 percent higher rate of utilization of the Emergency Room. And hospitalization, not significantly lower. There’s a lot we don’t know. We don’t have a lot of the data, utilization data preceding this experiment. But it’s clear that price has a tremendous, tremendous impact on overall costs.

So, you know, I hope that the Agency can figure out a way to sort that out because obviously, quality ultimately is going to be able to depend on our ability to pay for it.

CAROLYN CLANCY: Newell.

NEWELL MC ELWEE: Thank you. I’ve been a big fan of MEPS. I really appreciate the presentation. So, my question is a little bit different from what you’ve heard before. I actually want to ask about some of the data from what you heard before. I actually want to ask about some of the data that you presented, which I find a little bit difficult for me to interpret because there is not a good context.

So I don't know whether five percent should be lower than 50 percent, what number it should be. But I do know that you are going to get a distribution of sicker patients using, spending more money. So as I think—one of the questions that you asked was about modeling. And I don't know if modeling can help provide a better context for interpretation or is you have looked at the data enough to provide that.

One of the things that I thought about was maybe seeing the same types of graphs and looking at maybe a state with high quality like Minnesota versus as state like Louisiana or something like that. Just help me understand the context of what the numbers should be.

STEVEN COHEN: Well, the intention was not to be prescriptive here. It was—one of the key attractions of the survey is it is producing the national statistics. And this is demonstrating in the nation what the overall trends are. But in terms of your question about modeling, what we can do is look at individuals under a particular chronic condition. And those that aren’t getting a particular treatment protocol, there would have to be sufficient numbers, one can try to simulate what would happen if their practice patterns were modestly different or significantly different—and what potentially would be the impact on use and expenditures.

We’ve done this also on changes in coverage. And much like the piece by Steve Hill in terms of, if, in fact, people with individual coverage got a sort of employer sponsored coverage, what would be the implications. But there are tremendous caveats. You know, the underlying simulations have a number or assumptions that you really won’t want to hang your hat on that. But it’s suggestive and informative. So that could be another line of analysis.

Again, the key thing is having very, very large data resources to actually have sufficient sample and all the core variables to actually produce the simulation. So, in no way was this trying to be prescriptive and saying, what is the right number. But show patterns and, in terms of who the driver of the expenditure or distribution is, there is a critical desire to know that. And is that changing? And if it’s becoming more concentrated, that certainly is a signal where action could possibly change that.

NEWELL MC ELWEE: So will any of that be addressed? I think there was an FOI recently, if I’m not mistaken, about the Value Portfolio. Or there is something that I saw published recently from AHRQ on the Value Portfolio. So will any of those sorts of contextual issues be addressed in that?

STEVEN COHEN: I do believe there was something related to modeling. And Irene, is there anything you want to add on that front in terms of --?

IRENE: There are a couple of things that you could be referring to. One was that there was an FOA on the [01:03:09] a new grant program. I don't know if it—

NEWELL MC ELWEE: So this had to do with the National Quality Strategy.

CAROLYN CLANCY: Why don’t we do this off line and I’ll get back to you.

NEWELL MC ELWEE: All right.

CAROLYN CLANCY: Dr. Schneider.

CATHERINE SCHNIEDER: So a little bit to that theme, you know, keeping, moving the top one percent [mic discussion]—Moving the top one percent or five percent out is a simpler challenge because you know who they are. But, obviously, what they are talking about is really predictive modeling to identify who in the rest of that population could be kept from moving into the top one percent or five percent or whatever you cut off is.

And I think—I’ve been starting to hear a lot of bubbling around the use of big data, which is sort of this trendy idea. But there is stuff going on in the private sector that is really—on the one hand it is a little scary. On the other hand it’s very interesting such as a credit score is the best predictor of a readmission which actually, if you think about it, makes perfect sense around the amount of chaos in someone’s life or whatnot being the predictor of whether you will be able to follow a plan of care and transition out in an orderly fashion and so forth.

Or things like measuring kids with depression and whether—if their texting rate goes down precipitously, it’s a red flag for depression episode starting to impact their safety risk, and so forth. And I’m just wondering whether that’s just a topic that this group might be interested in learning more about. I mean certainly—I’m not sure there is a lot of rigorous research going on in that area. I think it’s very entrepreneurial kind of big, black boxes. What kind of big data do people even know about us. Just wondering if you have any thoughts on that.

Because that’s really, ultimately, the predictive modeling tool that can give you real time, actionable information, whether it’s to the medical home or wherever, to help keep those folks from moving into that top tier right at the right time. That’s really the Holy Grail of all of this.

STEVEN COHEN: Carolyn [01:05:54]. That’s incredibly helpful. I could say, the Agency has supported a number of methodological advancements. The ones that come to mind are those that actually are using real data. Andrew Zhou, with University of Washington on better techniques for the highly skewed expenditure distribution and Lei Liu with—she was at the University of Virginia—but focusing on methodological innovations, on predictive modeling, I think would be in scope in terms of the method necessary to improve the accuracy of the prediction in the evaluative models—that would give, you know, more rigor to what you are getting out.

And, so, I think that is in scope in terms of—I think if we work—going to make that a priority, we have to make some changes there.

CAROLYN CLANCY: I for one would totally love to hear about some of these issues so maybe we can follow-up with you. Dr. Atkins.

DAVID ATKINS: I’ll try to keep my comments short because I know a lot of people want to talk. But this stimulated a number of things. So one is this issue of identifying the high utilizers, really mixes two issues. One is, do they represent preventable costs that we can intervene earlier. And I know at least in the VA, but I suspect elsewhere, a lot of that is mental health comorbidity. And so the push to integrate mental health and primary care has been a critical issue.

And thinking about models of care—and the Camden experience, I’m sure a lot of those interventions were non-medical. And so how do you integrate a care system that is addressing the non-medical issues that are causing people to seek medical care?

But the second piece of it is, is it a flag for a appropriateness? And that appropriateness issue isn’t confined to those high utilizers. They are just the potential, you know, to save more money with them. And so the comment I guess I had, you know, the value agenda looks like it is one percent of our budget when it’s reducing cost is one of the three goals of the National Quality Agenda. And so I think—I hope that there are ways to try to increase the attention to the value agenda.

Bob Brook had nice editorial about a year ago in JAMA, highlighting sort of five things that he thought health services research should focus on, related to value. And they were appropriateness, frequency, labor—you know, who do you need to do a colonoscopy? Do you need a gastroenterologist? Can it be a nurse practitioner? Transparency—I’m forgetting the fifth one.

So obviously, that’s a huge agenda to do on $3 million dollar budget. And so trying to think about how collectively—I mean the VA is—we’re not insulated from the same cost. We’ve had an explosion in costs that outpaces our increase in veteran’s population. So we’re sort of beginning to struggle with that. I mean I think for me one of the most interesting questions that Helen raised is, what’s involved in de-implementing something? You know, I think that is a very different process than implementing something.

And it’s behavioral issues among the providers, among the patients. It’s economic issues. And so with—and with the very little research that I could find on how you de-implement a practice that’s established other than very crude measures. And so, if we want to avoid the third rail of rationing, you know, how do we get to better appropriateness, so we’re not saying no, no patient over 70 can get a colonoscopy. But we’re saying, a lot of patients over 70, it’s not appropriate for, an how do we target that?

And so there’s a—I would love to see ways that AHRQ and the VA can work together on that issue. Because we have the same practice issues. We published a very similar study on the colonoscopy showing that—figures very similar, overuse of colonoscopy in older veterans with comorbidity. And that’s a nice control. We don’t have the financial incentives that might be driving it elsewhere. So it’s really an issue of sort of established practice.

Last comment—in the value agenda, is there—there needs to be research on sort of the policy end of driving value. And we keep seeing citations of the Rand health experiment to look at what is the expect of shifting costs to patients. It seems like are we updating that data on, you know, what happens to behavior in high deductible plans, and are we reducing inappropriate care? Are we reducing appropriate care? How do we get that mix right?

CAROLYN CLANCY: Those are all great points. Helen I wanted to put you on the spot for the moment because Catherine’s comment and also knowing about what some employers are doing that I’ve learned at your organization. I’m thinking of Boeing working with the medical groups directly in Seattle, where I they were focusing on the top 10 percent. And they kind of said, “This model ain’t working. And not only that, you’re not working with us anymore,” in so many words, “if you don’t work with us to create a different model.”

They had an external consultant to create something called an ambulatory ICU right up there in marketing terms, in patient-centered medical home. In other words it seemed like a not (?) wonderful label but—because they were treating a lot of the patients, Alan, like your father-in-law. That was actually the basic drift. And, you know, they worked together to do this.

How hard would it be for AHRQ to convene some of these folks? Obviously, we’re not going to provide the leverage that an employer is going to do. And we are probably not going to provide the opportunity and incentive that they would do. But we could certainly either advise or fund some of the evaluation. So if you would think about that, I think that would be really, really important. Because some of these models, actually you can’t do very much with unless you have enough data.

I’ve certainly looked at a lot of papers to review about high deductible plans. And the problem has been, not a lot of follow up and small numbers to be able to say very much that fairly obvious things like, “This year people did less stuff generally.”

FEMALE: I would say, actually, there is—we do know a lot already. It wouldn’t pass muster. The way we know it wouldn’t pass muster in an academic journal, not because there is anything wrong with it but it’s not designed for those purposes. You don’t have the data you need to actually do it in the thorough way that you would. But we know it is exactly like what happened in the Rand experiment and it is exactly what happened in managed care, which was HMOs reduced appropriate and inappropriate care.

We actually don’t even know exactly the dimensions of that. But we know that’s what happened. That is exactly what’s happening. For the most part, those plans are working pretty well. And they do what they are supposed to do. But even figuring out, for example, if you wanted to say, “We only want to reduce inappropriate care,” actually designing it to make that happen would be very hard. But for the most part, more and more employers now are moving not just offering them but to full replacement.

And I think we are going to be with consumer directed health plans in corporate America in three years, where we were at the maximum of the managed care movement. And, by the way, it’s happened in almost the identical timeframe. So it’s here to stay. And the only question is, how is it going to be refined once it is pretty much fully implemented across the country, at least in the employer world.

And our suspicion is that the exchanges are going to find pretty quickly that if they don’t offer something like that, they are going to have unaffordable choices. And we already know from Massachusetts and other places. We don’t need to go back 30 years to Rand or however many years it is, to know that the average consumer, unless they’ve got a lot of money, are picking the lowest cost plan that’s available to them, because they are all very expensive.

BRUCE SIEGEL: Dr. Selker?

HARRY SELKER: I want to mention something about the way you’ve used the economic information to get where you want to get. And then a question. But first of all I want to compliment you. Because as much as it is easy to look at those, as you call it, the concentration of expenses and people, it can distract. And you are doing the cross, the full thickness, which is what I think is really important. That is, it is fine to look at those as ways of case find if you will. But ultimately, we want to improve care. That’s the point.

In fact, I would even say suggest that your communication about the concentration of expenditures, brings in a perspective to the public, which has two problems with it, I think. One, as a communication vehicle—and you know the public can get quite riled up about health care, rationing it and so forth. The idea of seeing them as a concentration of it, when really they are sick people in most cases, I think may not be as productive as you’d wish. It can lead to some really strange and unpleasant conversations.

The other thing is, it leads to, among those who think about this a lot, like Helen’s comments and also Newell’s, is there are so many questions you can ask about the distribution of data. I think it is absolutely right that you might see less money in the expensive patients if, in fact, more of it is being used for other patients. And, in fact, not only is there regression to the mean, there is also regression to the morgue, where some of those people are falling off in another way. So, I think it is useful to use those data as a signal. But then I think we really probably want to not just go public with all of that.

The thing that really made me wonder—you are doing such an amazing job of data collection there. I’m reminded of a trial ongoing right now, I believe, in Scandinavia, the TASTE trial. Where they are doing a randomized control trial of thrombectomy. So when they are doing the stent, coronary stent insertion, in some of them they are sucking out the thrombus and some they are not. It sounds pretty dramatic. And they are doing about 10,000 patients in this.

And, you know how much this is going to cost them to do and how much data they have to collect? They have to collect one piece of data, which is whether they were randomized to a control or intervention. Everything else is collected out of the data systems they already had for giving health care. Now we are not there yet in this country. But we could do so much better research if we actually used the data we are collecting.

You are collecting really deep and full thickness data, that allows us to translate quickly from these issues of cost down to what we really care about, which is, is our family member getting the care they want. That is why these anecdotes move us because we care.

And so I think if you engage without violations of various sorts of information flow, doing trials of interventions in the places where you are already collecting those deep data, that would be really interesting. Because then you could do the link that you want. It is not just worrying about cost for the sake of improving care. Can you do that?

STEVEN COHEN: I certainly have to get permission. We have to get permission to do everything but you certainly could go forward with the resource [low mic] clear-cut—very clear-cut protocol. We have gone back to respondents to other health care surveys.

HARRY SELKER: But I’m talking about doing an intervention and using the data you are collecting.

STEVEN COHEN: Oh, an intervention with [simultaneous conversation]

HARRY SELKER: Doing an intervention in the people from whom you are already collecting these data, just like they are doing this huge trial. It would cost tens of millions of dollars. They are doing it for nothing.

STEVEN COHEN: Yeah. We would certainly explore if that’s possible.

MALE: I think what you are getting at it is what we are doing both in our CER portfolio and Health IT of thinking about how clinical datasets and places that are collecting data for clinical purposes can be tweaked to harvest data for research purposes to teach us about improvements in quality. And we have a whole effort engaged in how do we make that happen here. There are lots of privacy and HER sides of that to work through. But our work on registries, that the idea that you are not collecting everything through a new survey, that you are using clinical data systems to populate registries, which then can be used for research.

I work with distributed data networks. And the idea that we can ask clinical research questions of practices and have permission to go in and collect that data instantly, to answer new questions. So we have the flagship of MEPS going and we are implementing other projects and other ways to get to that Scandinavian type model.

BRUCE SIEGEL: Ms. Haskell

HELEN HASKELL: Well, this touches on questions that other people have asked along the way. But my question is, what percentage, what proportion of this five percent of high utilizers is due to health care associated conditions, hospital-acquired conditions? In my experience, a lot of the examples that I see cited, they tend to involve hospital acquired infection or something like that. And that’s a concern that we have. We see a lot of sort of patients who go into the hospital never suspecting that they are about to become a high utilizer or even about to embark on end-of-life care. So if that’s not something that’s being looked at, should it be.

MALE: We clearly have quite a bit of information in terms of the length of stay and issues tied to potentially hospital acquired infection. So to the extent we have the power—think of one person representing 10,000 people in the survey. You usually need about 100 people for an estimate. You would have to drill deep. And, you know, we will certainly will explore what we could do on that front. That is certainly an important question.

HELEN HASKELL: It’s our suspicion that it really is a high percentage. And that this is really skewing a lot of the statistics. And that’s a big area of potential savings of life and costs.

BRUCE SIEGEL: Dr. Pensen.

DAVID PENSEN: Thank you. I was struck by your last slide and the questions you asked about, what’s our role and what are the next steps with this initiative and specifically with the MEPS and economic modeling, which I think is really powerful. And the work is really excellent. And as you were thinking, I thought, well, where do you take this? And my thought is, is that, you know, you probably, you’ve taken it very far internally but it’s something which you would like to spread out.

And particularly Dr. Schneider’s comment about other data sources, which are out there, and different variables you have never thought about and different methods. And there’s actually, I think an applicable model out there already. I was thinking as you were speaking. And that’s the NCI SYSNET model. I don't know if you’re familiar with that network. And I’m part of it. So for self, complete disclosure. But if you look at it, you’ve got a group of investigators who’ve come in. They’re funded by NCI but have different perspectives. Most of them are biostatisticians or epidemiologists. Many of them bring their own data in.

In the prostate group we now have the European prostate cancer screening trial pooled with SIR (?) Medicare data. And we’re looking at different methods. Now, now saying this to brag about SYSNET. I’m saying this that you could potentially create something very similar around MEPS and bring in other datasets and create new models.

The other point that I would make that I think is unique about SYSNET, is I think a lot of the questions are not driven by the individual investigators. The methods are driven by the extra [01:23:03] folks but the questions are driven by NCI staff. And I think that that’s where AHRQ’s role in that would be. Instead of—if you go to the MEPS Website—now here’s the data. Feel free to use it. I would actually say, let’s convene a network. Let’s send some money out there. I hate to say that after I saw the first slide this morning. But that being said—and you guys say, “Here are the questions that need answering. And what are your methods?” Because I think that getting a bunch of minds together would go a long way. And I think that’s a great model. That’s been my experience anyways

MALE: Just a reaction to that. We’re somewhere along that curve in that beyond what’s on the Website, is that AHRQ has one of the few data centers in the department. And in the datacenter with the approved projects, researchers bring in other data sources. And they do merges and sometimes it’s statistical matches. And that still stays unless it’s proprietary within the datacenter.

And so, while we’re not—the researcher has to pay for their resources and time in using the data. So we having actually gone to like establishing a network. But the infrastructure is actually there in many ways. That’s very, very helpful.

DAVID PENSEN: If you are already there, then it is just a matter of, you know, creating a network, asking people, telling people the question you want answered and then saying, “Come on into the datacenter.” It really works with SYSNET, I think, you know, dumb urologist.

BRUCE SIEGEL: Dr. Hoven.

ARDIS DEE HOVEN: Thank you. I will be brief. My first comment goes to prevention and wellness. And after I’ve been out on the road talking about health care and prevention and wellness and how important it is and how cost effective it is, someone will then ask me the question, “So how are the doctors going to manage all these newly insured patients that we are going to have and how are we going to handle this? What’s the workflow going to be, etcetera, etcetera?”

And I remind people these folks are not new to the health care delivery system. They are going to have to be managed in a different way. And so I think the challenge to us as clinicians and everyone interested in this, and AHRQ in particular—so my ask to AHRQ is, you know, we are going to have to teach them how to use the health care system wisely. And this is going to be hard because we are talking about doing it at different levels, at different sites of care. It’s going to be local in many situations.

You can’t put something out on the Web and say, “This is what you need to be doing.” What we are going to have to do is look at a reasonable way for not only the patients, the people who need the care, but also, the whole health care community—to begin an education process that embeds the whole understanding together so that we can all work seamlessly and get this done appropriately.

So it goes to Helen’s point. It goes to everybody’s point around this table. But I think it’s a huge amount of work that we are going to have to do, how to use the health care system wisely.

The other point I would like to make—I do HIV care as most of you know. I’ll be sitting in the office, and it goes to Alan’s concern, you know I will say to the patient, “I’m your infectious disease specialist. Who is your primary care physician? Who is managing your diabetes, your hypertension?” They look at me and they say, “You are.” And this happens every day.

And I think as we are looking at, and it’s already underway, delivery of care models out there. There’s been a lot of work. Again, we are going to be seeing changes in delivery of care models, be it the patients under medical home or something else. And I think we are going to see variations on that in a variety of places, depending upon the region, the practice, the wants of the patient and so forth and so on. Because ultimately the payment is going to follow the delivery and how the care is being delivered. And I think, to Alan’s father-in-law’s issue, that ultimately will be the resolution.

But again, I’m very concerned because we’ve got Medicare populations out there right now that are not getting the preventative services we need. And what are we going to do to enhance that an make it better? Is it my fault as a physician? Is it a system issue? What is it that’s, you know, getting in the way of getting us to a better place on this. So my ask of AHRQ is to fix this.

MALE: We’re on. [Laughter]

BRUCE SIEGEL: I think they just said yes. Dr. Wyatt.

JAN WYATT: Thank you very much. It was an excellent presentation and very much appreciated. And I want to just piggyback onto Dr. Hoven’s comment and that is, successful interventions in primary care and chronic disease management really should be team driven. And I very much appreciated Dr. Meyers’ comment about that and highlighting that as a necessary intervention. I think that’s going to be a new delivery model that I hope begins to evolve.

Of course, I want to emphasize the role of nursing care as a very much an important ingredient along with other valuable health resources, but particularly in oncology care we have seen some tremendous innovations with nurse navigators as patients manage the multiple points of intervention that are necessary in that. And there have been also some great data to support the role of nurse practitioners in managing what, at one point was a very high hospital admission rate from nursing homes to hospitals that, with nurse practitioner care.

And I think United Health Care has adopted a great model that was invented by several family nurse practitioners. It is called and Elder Care—and implemented that throughout their system. And where nurse practitioners are the initial intervention of call from nursing resources in nursing homes to determine whether a patient needs transportation to a hospital for admission for a particular problem. It saved them millions of dollars.

But as we look at that, the value equation for patients I think is shared ownership of their health / illness experience by multiple clinicians within the health care system. And I would love to refer Dr. Spitzer to a wonderful geriatric nurse practitioner in the Massachusetts area—because I think there has been unambiguous information about the positive care that nurse practitioners can manage because they own the patient’s experience of health and illness through the lifespan.

The other issue I wanted to ask for some help with really concerns data enhancement in arthritis care. And I’ve been privileged to be part of a wonderful disciplinary health team that includes orthopedic surgeons as well as PT, nurse practitioners, bone and mineral specialists and what have you—within the US Bone and Joint Initiative. And we are working on an initiative to create a paradigm for early recognition of osteoarthritis.

Osteoarthritis is a chronic condition that we think can be amenable to some of the screening for risk factors and prevention oriented activities and ongoing monitoring that occurs right now with blood pressure and cardiovascular disease and so on and so forth. And we’re really looking for some help in creating a metric for early diagnosis of osteoarthritis that currently doesn’t exist. Patients are very well attuned to metrics in their health care. They know their blood pressure. They know about pack years. They know blood glucose. And they are very engaged in conversations with their providers about these kinds of metrics.

There is no metric for joint health and yet by 2030 there will be over 3.5 million knee replacements. And the cost for those replacements that you probably are aware of will increase over 600 percent according to the resources at the US Bone and Joint Initiative. So in terms of impacting cost or arthritis care at an early point would be tremendous because there are so many lifestyle issues that are associated with the progression of arthritis as a chronic illness that we think early intervention could be important.

But what we are finding is that patients—we have patients with knee and hip replacements that still don’t think they have arthritis, or that that was the reason for their admission. And, in fact, as you track that hospital admission criteria, they are admitted for the procedure, not necessarily for the related diagnosis of arthritis. So we can’t even really, truly count, you are talking about a registry, count the incidence or prevalence of arthritis because we are just counting procedures.

So I would love to see some data enhancements in that area. Thank you.

BRUCE SIEGEL: Dr. Thompson, I assume you already spoke.

JEFF THOMPSON: I just have one question on the cost about, the provider relation to the cost. When I looked at the 5-50, especially around narcotics I find that there are prescribers that, you know, they like their narcotics. And they have a very different idea about how much they want to give. And so, is there some way that we can start looking at what is the provider contribution to the 5-50 and looking for those as outliers. And I know there are, that believe, that have some altered sort of evidence that they believe that more is better, especially around narcotics.

And I’ve actually called up the top 20 prescribers in my state and talked to them and asked them about their narcotic prescribing for Medicaid—and was quite surprised at some of the reactions I got from that.

STEVEN COHEN: (?) Just a quick response to that.

BRUCE SIEGEL: Yes. Please.

STEVEN COHEN: (?) The Value Portfolio is actually sponsoring an activity that looks at data from providers that would address some of the policy issues of the day, that the federal sector is not capturing. So we definitely will consider that. I thing Herb Wong is leading that effort in the Agency. So there is like a lot of opportunities on that front. And stay tuned.

BRUCE SIEGEL: Thank you. Before we go there, so any members on the phone have any questions? Okay. Hearing none, Dr. Atkins.

DAVID ATKINS: Just one more comment about that medical home issue. The VA is in the, I guess about a year or so into a national rollout of the medical home. I think it’s the right thing to do. But it’s going to be very challenging to do it. And the returns on it are going to take a while to show out. So I think we need to be realistic about what we can expect in terms of cost. My understanding is the example at Group Health that is cited so often as the success of a medical home, when they implemented it they reduced the panel size for clinicians.

So they were able to get cost savings. They were able to improve satisfaction and reduce provider burnout but not by having them maintain the same panel size. And so I think doing a good medical home isn’t a simple issue. It takes work to work as teams. It is not necessarily less work. It can be more work. And we’re learning some of those painful experiences especially in the transition period when you have all the costs and the upheaval and none of the benefits yet.

So I guess the request for you is just to pay attention to those things and the more we can share those learnings so that people don’t get turned off by the experience because they don’t see this magical solution a year or two into this big transformations.

BRUCE SIEGEL: Great. I think we have no more comments. Thank you very much, much appreciated. This is generally our time for public comment. I don’t believe anybody signed up for public comment. Am I correct? Is there anybody who wants to comment in the audience? Please. Approach the microphone and if you could identify yourself we would be most grateful.

BONNIE HELMALE: Hi. My name is Bonnie Helm and I’m actually nutrition and dietetic student right now. And I just know there has been a lot of—it seems like a lot of the chronic issues that we are talking about and different things have a definite underlie in nutrition—a lot of the things we are seeing these days. That is a huge cost to our system. But I rarely hear of a dietician being part of an interdisciplinary teams that you talk about.

In fact, most of the physicians I volunteer with currently kind of view nutritionists as a bit of a waste or a thorn in their side. Do you see this changing a little bit or bringing nutrition more into an interdisciplinary group that you are talking about for helping with some of these chronic diseases and research you are doing on those things? Or do you just not look at that at all really?

CAROLYN CLANCY: I guess what I would personally see is, we don’t do such a great job nor do many other developed countries in terms of providing the best possible care for people with chronic illnesses, whether they’ve got one, more than one, and so forth. And there’s been a lot of discussion that all kind of converged here on the theme of patients owning this and so forth. I think the self-management support is part of that.

But do I think that the current medical practice probably underemphasizes the importance of nutrition? Hm-hmm. Particularly the kinds or practical steps and strategies that people need to make changes. But I think that’s true across a lot of other domains. So it’s an area where I think this nation is going to have to make a big, big—part of Ardis’ comments about re-teaching people who have gotten care, even if they have not had insurance and so forth, to engage and be a part of the team and so forth—is going to require the assistance of a lot of other professionals, community resources and so forth.

So, aspirationally, we’re there. In real life we’re at the very early phases. We recognize that we have some challenges.

BONNIE HELMALE: Thank you.

BRUCE SIEGEL: Thank you and thank you for your comment. I don’t see any other individuals for public comment at this time. So we’ve had a little change in our logistics. Let me go through this with our members as well as with our public guests. Cozy is running late. So our market-based solution has had some problems. I think they are looking for--

FEMALE: [01:37:58]

BRUCE SIEGEL: They are looking for a larger subsidy or something. I don't know. [Laughter] Yeah. Thank you. And so what we are going to do is that Jamie is going herd the members who are hear up for our photographs on the third floor and then after that she will herd us to the fifth floor for our executive session, which will be closed to the public. In that (?) we get our photographs taken, Cozy will be there. So those of you who want to buy your lunch can buy your lunch up there.

And then Cozy will come down back here for everybody else, for the public and others who want to buy lunch out in the foyer. So that’s how that will work. Does that make sense to everybody?

[Brief unintelligible conversation]

BRUCE SIEGEL: Hold on a second. I need to talk to people on the phone, also, before we all leave. For our friends, our members on the phone, we will ask you to hang up from this current call, our members, and dial in, in about 15 minutes for the executive session through the new number that you have been provided. And then, after that session, you can dial back into the existing number, the number you’re using now, to recommence with us during the public session this afternoon. Hope that made sense to our members on the phone.

FEMALE: Okay. Thank you.

BRUCE SIEGEL: Any questions from our members at this point? We are all set? Thank you very much. We will now break.

END OF SESSION