[This Transcript is Unedited]
DEPARTMENT OF HEALTH AND HUMAN SERVICES
NATIONAL COMMITTEE ON VITAL AND HEALTH STATISTICS
SUBCOMMITTEE ON POPULATION HEALTH
September 15, 2010
Embassy Suites Crystal City Hotel
1300 Jefferson Davis Highway
Arlington, VA 22202
CASET Associates, Ltd.
Fairfax, Virginia 22030
P R O C E E D I N G S 2:15 p.m.
DR. STEINWACHS: Welcome. Why don’t we go around the table and do introductions.
Don Steinwachs of Johns Hopkins.
DR. BOENNING: Doug Boenning from ASPE, no conflicts.
MR. HITCHCOCK: Doug Hitchcock from ASPE. No conflicts.
MS. JACKSON: Debbie Jackson, NCHS, CDC.
DR. HORNBROOK: Marc Hornbrook, Kaiser Permanente, member of the committee, no conflicts.
DR. GREEN: Larry Green, University of Colorado, Denver, member of the committee, no conflicts.
MS. JAMISON: Missy Jamison, NCHS staff.
DR. FRANCIS: Leslie Francis, member of the Full Committee and I am co-chair on the Privacy Subcommittee.
MR. LAND: Garland Land, member of the Committee, no conflicts.
DR. STEINWACHS: So, why don’t we just talk about the agenda. Larry and I put this together fairly quickly. Most of the discussion was about trying to decide what are the priorities for this subcommittee for the coming year.
Are there any other things that ought to be on the agenda?
Missy, do you want to start.
MS. JAMISON: I have one for number five and two for number two.
DR. STEINWACHS: It looked like that one, two and five, were the ones that we had the most positive endorsement.
I now have Garland’s – number one is five, number two is four, number three is two, number four is one, number five is three.
So let us talk about two ideas; one is that we can go through each of these but for the same since there seems to be a group clustering around one, two and five, sort of the discussion about what is most important, are there elements that you put together into one package – one high, five low?
Okay so reverse Garlands. Garland’s number one is one, number two is two, number three is four, number four is five, and number five is three.
So I put down for you, one is number five and two was number two.
So, I think it’s come through consistently, so one, two, and five on average got the highest endorsements. One and two, I don’t really remember what the whole statement was, but one was really looking at capacity to link health data and monitor population health benefits at a community level.
Two is sort of closely related, so again at a community level and state, the capacity to link onto data, information on health resources and the range of influences on health. So, it’s not just monitoring public health, but two, is to instead broaden that to include the influences on health. And, so let me say something and see whether anyone agrees.
In a way, one and two, one is maybe step one, and two is step two in that process. I don’t know, looking at Mark and others. So, you know, you might think of them together but you might start saying one could you do the basic function of monitoring and then two, you add on other data sets and information that would help you look at influences on it.
Then, number five is sort of a cross-cutting kind of issue that says you have both the metrics and the capacity to monitor the impacts of health reform which would pull off some of the same data that you might use for monitoring population health.
MR. HITCHCOCK: Was that a best practices issue? I don’t know that it would be, they probably are.
DR. STEINWACHS: So, I thought, what you’re saying Dale that, you know, it would be great to have a hearing where we could bring in some people who represent states and localities that have been able to carry out linkages. And, so I know they said but some brought a real public health infrastructure and to hear from them about how they did it, what barriers and challenges.
The other was, I guess, on one was sort of thinking about since we’re trying to push the envelope with electronic health records in places that have begun to tackle, in a sense, trying to put a denominator under some electronic health record data so that you could actually use it in the way that helps on population health statistics. And, so again, starting would probably be best practices of what’s out there, so it would be what are the things like in the Kaiser health plan that you can do in the enrolled population, that would you think in the general community population may be tougher, but maybe there are things to be learned on what’s done.
So, does quiet mean everyone agrees? Okay. So, let’s talk a little bit about when and how we might try and organize something in terms of a hearing. We’re now in September.
SPEAKER: Assume October.
DR. STEINWACHS: Definitely not.
DR. GREEN: I have another question. Did Leslie leave?
DR. STEINWACHS: No, I think she’s coming back. Her stuff is here.
DR. GREEN: In light of that adjudication you just did, I would like to inquire of the Privacy Subcommittee, about their level of interest in our items number three and four on this list, about education of the public. They had a number of claims and statements about that in some of their prior letters and reports, and it goes down our priority list, and I’d just like to at least entertain the possibility of asking them if that might go on their list.
DR. STEINWACHS: It’s always good to share ideas. Those are good ideas. Okay, so when Leslie gets back why don’t we suggest that.
MR. HITCHCOCK: We probably should avoid using the word “hearing” because we actually got advice from our Office of General Counsel at one point, that we don’t really do hearings. I don’t really know what the preferred —
DR. STEINWACHS: Forum?
MR. HITCHCOCK: Forum –
DR. GREEN: Meeting?
MR. HITCHCOCK: Yes, meeting. [Laughter.]
DR. STEINWACHS: We did workshops. Last time, we were doing workshops. Is “workshops” good? Workshops are something that sounds positive, but doesn’t, okay Dale, you’re going to keep me out of trouble.
So, think about the earliest would probably be trying to do something in November. December is usually very tough, unless it’s early December, to pull off.
So, let’s talk a little bit about who are the customers. We’ve tried in the past to think about who do we work with, in a sense, to shape workshops. So, since we have wonderful representation here from ASPE, are these issues, where do they sit, because these are also still very thick questions. They could be refined, shaped more too when you think about who do you select, and then who do you try and get to.
MR. HITCHCOCK: Have we heard from NCHS, Linda Bilheimer, about the community health data indicator warehouse?
DR. STEINWACHS: She gave a presentation, yes, because I was going to turn to Missy next on NCHS.
MS. JACKSON: More as an intro – as we are getting Todd Park(?), hopefully at the November meeting. I can get him to do an update on what the data warehouse is all about. So, something like that could lead into following up, to build up of what we want to do early in the year.
I just don’t know about workshops or hearings before the end of the year. It gets very difficult because logistically –
DR. STEINWACHS: So, you mean it might be more like February is what –-
MS. JACKSON: That is not the best weather time. It just kind of depends on where you can align what is going on in the department with other topics that come up as well.
MR. HITCHCOCK: I think that’s a good idea.
DR. STEINWACHS: Well, why don’t you share with Linda that we’re talking about having a workshop on best practices in community state level monitoring population health, so that’s both the issue of data systems but also what are the metrics? What are the things you’re using? And it would be appropriate, particularly at the community level.
MR. HITCHCOCK: I sort of like Debbie’s idea about Todd Park – so we can be sure to actually get Todd to come to the meeting. We probably can.
DR. STEINWACHS: Okay, so what we’re saying is to get Todd as the lead off.
DR. GREEN: What was the answer to the question of who we think our customers are?
DR. STEINWACHS: Well, I was trying to go around so I didn’t manage to get Dale into that, he shifted immediately to Missy and NCHS. [laughter]. And, Doug was unusually quiet about this.
MR. BOENNING: But, we’ve started to hear a little bit about this at Data Council, and NCHS is going to add some questions to upcoming surveys, to look at, at a minimum, emergency department utilization, anticipating that with greater coverage they’ll be misuse of emergency departments. I am not sure that is going to happen for a long time. But we have had some discussions about health reform and how to monitor it. That may be a group to involve with any kind of hearing, workshop, meeting, that we have that has got people from the different agencies, like CMS, AHRQ, all the data leads from the various HHS agencies come to that monthly meeting.
MR. HITCHCOCK: Or do you mean customers, people who would like the federal government to be able to provide more localized level for state and county health offices?
DR. STEINWACHS: We started, I think a couple years ago, trying to make sure that when we took on a topic that within HHS, there was interest. Certainly, that doesn’t have to be every topic we take on. And, so, as much together, we did the workshops on data for modeling health insurance coverage, and so on, and that particularly tied to an interest that you had with an ASPE, certainly an interest we had too.
So what I was looking for was whether or not within the Department there were places -the Office of the National Coordinator, Larry and I were talking about ONC and how they would see their interests. We’ll talk to them because that could help shape who and sort of what gets talked about specifically under this broader umbrella. That’s what I was looking for.
MR. HITCHCOCK: There’s the Healthy People Process – that is another one. Maybe they would be interested.
DR. STEINWACHS: And so, this is Healthy People 2020?
MR. HITCHCOCK: Yes.
MR. LAND: Several of this is talking about community, state and local, I think that is really where your customer base is – your state and local health departments. And, I would suggest an opportunity would be to invite people from Massachusetts where they had health care reform now for a couple of years. And, I think their health department has actually been using ED data to look at the impact of their health care reform.
MR. HORNBROOK: That would be good, especially if somebody has done some analysis, actually show something.
DR. GREEN: I’ll just step out of government, but the two other audiences, I think there’s an EHR vendor audience here that needs to start thinking in this direction. And, the second one is CTSA awardees. They all have a community engagement arm that cuts across the NIH Institutes. It might be a place – as I circulate among the institutions that have those, they are struggling for data sources to accomplish their state admissions. And, these roads tend to lead back to needing more granular data at the local level in order to make measures that they need related to proposed interventions.
And, so I think that the CTSA awardees, and there are about 57 of those institutions or so, I think they would be a great audience for this, and they would be very interested in this linkage. The word linkage in here, they are struggling with the linkages across the social determinants of health and the linkage to the delivery system at the intersection of the policies. That’s what they’re interested in.
DR. STEINWACHS: Are there some people particularly in the CTSA?
MS. JACKSON: There is a serious air conditioning loudness, so please speak up.
MR. HITCHCOCK: There is a committee, a national committee on the —
DR. GREEN: The Key Functional Committee for Community Engagement chaired by Lloyd Michener at Duke University.
DR. HORNBROOK: He’s out of NCRR at Duke, at their clinical research. And, of course, you know, they’re struggling like we are. It’s a committee of people with big ideas and no money and a lot of frustration against the cultures of medical schools that are embedded on their own imperial empires not about the people.
DR. STEINWACHS: Conversations, whenever I listen to them at opH
Hopkins, come back quickly to patient recruitment.
DR. GREEN: Yes, exactly.
DR. STEINWACHS: Engagement of the community is trying to find more people to put in randomized trials.
DR. GREEN: Well, that’s the definition of a data bank – a community – a tissue bank. (Laughter) That wasn’t serious.
The social determinants of these items, Nancy Breen, she’s just carried the water of this cause over and over again. I wish she were here to do it all over again. But, the linkage of those down to a more granular level is a rich opportunity that we have spotted and we need to figure out how to take advantage of this.
MR. HITCHOCK: Every time I hear that sort of talk I think of geospatial displays. That might be some sort of area that we might want to hear or see a demo on.
DR. GREEN: That’s a way to substantiate a relationship.
MR. HORNBROOK: Yes, right, you get everybody’s medical records and you put them back to their house, plot them on a map, diseases, drugs or whatever. You know, the FDA is going to an Informatics model which NCHS would just love to have, and of course, they don’t have the same money as the FDA has.
The FDA has got the sentinel network, distributed database, and so they have got contracts with data holders and academics that are not data holders, to access individual patient records where there’s data on a drug, or all their drugs, and then downstream events, so they can stroll through the data looking for correlations so you look at the signal detection, but there’s also specific hypotheses tables on the test, and you’re at the level of individual patients who have never consented to having this. There’s no IRB involved because there is public health surveillance. So, nobody has ever asked anybody other than the health fund executive do you want to be involved in this. Of course, they’re not paying anybody very much either.
So, the question is, if NCHS had the same kind of data model where they can get the data on person X, and then have it come up through a de-identification regime to a certain level where they could then aggregate it by track, or by state or by county, for both immunization, diseases, procedures, drugs, BMI, vital signs, certain lab test results, for example. They’d have a whole bunch of indicators so now they go through the old survey methodology for.
Then of course the problem is to find the set of people who have electronic medical records for whom you could create and reweight samples to the US population, although there are still people who just don’t ever get into – I don’t know, we are getting federally qualified health centers that have electronic medical records. You can get data on uninsured people, too.
So, the point is that right now it seems like NCHS has the vision but not the money.
MS. CAIN: Well, and also, we really value the nationally representative characteristic of our data. I’m not sure what you’re talking about right now would necessarily, at this point, that we would necessarily be able to get a nationally – know that it’s a nationally represented sample.
DR. GREEN: That captures what I see as the hot spot for the workshops. It’s about numerators and denominators in the new world.
MR. HORNBROOK: We know EHR use is going to go up, the question is when does it cross over so you can do it.
MS. CAIN: And, that’s a question that, I know, Jane and her group have been monitoring, when will it reach the point that we can start counting on it.
MR. HITCHCOCK: Something that we’ve talked about for years and years and years in terms of sort of a patient centered or a person centered, survey system rather than an episodic thing where you can actually follow the person.
MR. GREEN: Longitudinal data sets. The best practices idea, forgive me, you don’t have a gun, do you? [Laughter.] I’m not sure that we should restrict it to best practices. The envisioning possibilities, people that are working this territory now with the tools they’ve got and know what the best way to do it now is, what are they thinking as the way the new world of the National Health Information Network is coming? What are they thinking of the possibilities of how they can do better and push it back more locally to get better data back to the community level?
It’s the vision of what looks like it ought to be possible, and then that can spill over into an understanding of, and what’s keeping us back from doing that, and she says, well, it’s just not developed enough that where we can even start using it. We’re just going to have to wait a few years until this thing gets disseminated and hooked up right. But, we should be thinking a longer view like the next five to ten years, where do we think vital and health statistics starts landing out there and how are they going to be effected by this? I’d like to see us take that on as part of this type of workshop thinking.
MR. HITCHCOCK: I think I was naïve in talking about best practices, maybe there are not any, but maybe what we want to hear about is work in progress and obstacles.
DR. GREEN: Yes, work in progress. I like that.
MR. HORNBROOK: We are getting lobbied from the discussion this morning –
DR. STEINWACHS: Well, I was telling Judith that there was a chance that she’s going to make the list of the top 100 most influential people in American health care over this topic. People are going to get to know her, start sending her e-mails, starting to interact, this committee may hit the sweet spot of American policy.
Let’s try and break it up into major pieces. And then, we can figure out are these separate workshops or are they separate days or pieces of the same workshop? So, Larry you said one issue, and that was people who are working on trying to say how do we come up with a numerator and denominator when we’re working with electronic health records as the unit.
And, my greatest hope is we’re going to get the Census to work with us because if you have an address, then you start to develop real community denominator, then you’ll have something about people who are not represented in that, and that would be fun to see if the Census Bureau, okay.
Then, it seemed to me the other was to begin to draw in – so if you had a numerator and denominator of electronic health records, I guess that is one piece, and then certainly at the community level, too, you might link that to surveys like the behavioral risk factor survey, it’s more of a state level linkage. You could link it to administrative data too, if it is administrative data played out in a useful ways, and you have the administrative data in the public health sector, the births and deaths and so on.
Then it seemed to me that there was sort of the, there are people who have been working on building the infrastructure for a public health data set or data systems, which gets it somewhat at number two, that says linking the other factors. So, would it be worth spending some time on what’s being done in building public health infrastructure, look at population health issues? All I remember, in Florida they got quite a ways.
Jim Studnicki used to be at Hopkins, went down and really put together a whole huge number of data sets into a public data warehouse where you could then begin to look at the relationship between environmental welfare data and a variety of other things. I’m sure there must be some other examples.
MR. LAND: Several states have developed web-based data query systems that had multiple data systems available, Utah, Florida, Massachusetts.
DR. STEINWACHS: Another issue, I guess, is the metrics. What do we have that you could use at the community and state level that would be useful for monitoring population health? And, we also have health reform here. Maybe to some extent, those can dovetail.
MR. HORNBROOK: There are some metrics, I think that are relevant, except for illegal activities. You don’t need to ask how many counted cigarettes have smoked if you know all the cigarettes that have been sold – except for the stuff that has been smuggled in. The same thing way with liquor, beer, and wine. The spirits are taxed – those tax revenues, there are also sales revenues, and you just divide by the number of population, and you’ve got a consumption in your head now. What you don’t know is the fact that there are people who are tea-toddlers and that somebody else is doing most of the drinking.
So, in some ways, you can back into health indicators from industry and tax data. And, then there is the ER registries, the trauma registries, that track how many people come in who are drunk, or drug overdose, match those to two data points, and say yes, they are currently going into people and the people are getting injured.
The same way with air pollution and asthma attacks, showing up at the ER. Clearly, there only a real small part of the respiratory disease, but there kind of like a canary in a coal mine. If somebody was going in with a very serious asthma attack because the air is very dirty, there’s a whole bunch of people that are at home unable to breathe very well.
So, these are simple trivial examples, probably lots of other when you think about people not taking their statins, cardiac incident event rate, geographically varying, stroke rate, so small area variations in these events, falls, especially in the elderly, abuse.
MR. LAND: Dr. Frieden, director of CDC, has his six winnable battles – is that the right number – he calls them the winnable battles of public health, and now they’re developing metrics associated with each one of those and they’re going out field testing in the states to see what metrics can be collected and make sense and so forth. That activity that is underway right now. So, they include motor vehicle injuries, smoking, obesity.
MS. JAMISON: I think it might be three or something.
MR. LAND: I think it’s more than three, of the things that he feels that are important, actually there’s enough science behind making a difference on.
DR. HORNBROOK: And they make a difference because they have strong health –
MR. HITCHCOCK: That would be another set of people to hear from, the folks at CDC in Atlanta. They devoted a lot of time in the past to this.
DR. STEINWACHS: I’m taking notes and they’re all over the place. I can see this will be quite a write-up here.
MS. CAIN: Forgive me for being late, but have you talked about the community health data initiative?
MR. HITCHCOCK: [Nods yes.]
MS. CAIN: Okay, good. Thanks.
DR. STEINWACHS: So, we were talking about what we had online, our discussion here.
DR. GREEN: You’re talking about a specific thing – why don’t you say it.
MS. CAIN: Right. Well, I’m not the best person to talk about it, but it’s the departmental initiative to bring data to the communities and then follow-up on private industry and a number of people to get communities involved or providing information to communities.
And, there’s been a lot of activity. I know Dale, you could probably talk about it. And we also have, associated with that, is the data indicator warehouse which NCHS is still providing the data for. These are going to go live, I think, December 1st, and there will be county level data in that. It’s macro- level data, not micro data. But, that certainly should be a part of the discussion on this topic.
MR. HORNBROOK: In some sense, there’s also a question of we have deaths certificates that are universal, there’s something that death certificates can tell you, is the most important thing a death certificate in a system can tell you about – presumably changeable health outcome. You can’t change it for that person because they are dead all ready, but presumably there is something in the society that’s generating that pattern of death certificates. Same way with birth certificates, low birth weight babies, and then moms morbidities are often on birth certificates. Babies without father’s names on the birth certificate.
DR. STEINWACHS: So, as you were saying, I’m losing track of who said what here, but if we were trying to engage some people from communities that feel like working on this, you identified a set of states, Missouri, Utah, Florida, Massachusetts, where apparently they have been working on putting together a data set. Are those the same ones you’re sort of thinking of going to for people trying to tackle the issue of, well, how do I get something useful – I’ve got mortality data for my community.
MR. LAND: Vital statistics is included in the data sets.
DR. STEINWACHS: So, I was going to say, are these states what you think there would be people there who would be good to talk to us about what actually are the indicators that are developing out of this and how they can become more actionable? Have they tested them out to see which ones –
MR. LAND: [Nods yes.]
DR. STEINWACHS: Okay, so I’ve got three things now. I’ve got the numerator and denominator, issues around electronic health records, possibility of using census data or other kinds of data to try and understand who is represented and who is not. We talked about the public health data infrastructure and there the examples of web-based systems we put up in Missouri, Utah, Florida and Massachusetts, but that effort to construct that infrastructure that, I think, has been one of our concerns. We are seeing a lot of electronic health records, but we are not doing something comparable in the public health sector.
MR. HORNBROOK: There is also, within that, a subset of the reportable diseases. It is not infectious disease, of course, but cancer and a few other reportable diseases that public health has the data sitting there somewhere. Sometimes they cannot share it, but it is there.
DR. STEINWACHS: And, a third was metrics, which is what is being measured, what is the capacity for measuring things. So that flows then into both what is being done at NCHS and the community initiative. We were talking about kinds of indicators that are at the community level like air pollution, things that are derived from health care like ED. It counters for specific things. You mentioned small area variations that put your community in the context of other communities.
MR. LAND: One of the other Federal programs that relates to that is the Environmental Public Health Tracking Network, which they have funded probably half of the states to develop environmental indicators.
DR. STEINWACHS: Environmental public health tracking?
MR. HORNBROOK: These are risk factors to the environment that have health implications?
MR. LAND: They have both outcomes and they have environmental. So they will track the health status indicators and they also have specific environmental indicators.
DR. STEINWACHS: So, is this a CDC initiative?
MR. LAND: I think that you can get it from the CDC website as well as data.gov. It is at data.gov now. This just underscores we need to get somebody from CDC. Sometimes they do not have systems that are completely national, but they have something that works.
MS. JAMISON: Yes, I used to be involved in that for cancer. I was the cancer representative for that. It was still in the very preliminary stages when I left two years ago, it was fair to say. Definitely the concept is what we are talking about.
MR. LAND: They have a national web portal and then the states have their own state portals showing – they have specific metrics that they recommend all the states collect and then they collect at the national level, also.
DR. STEINWACHS: So those metrics –
DR. GREEN: Don, what in your opinion makes this timely?
DR. STEINWACHS: What I was particularly interested in was that we have a lot of momentum in electronic health records. We have a vision that we hope there is a benefit to health statistics and population health. I think a real concern is the public health sector generally lacks the infrastructure, may not have the metrics, may not have the vision or figure out how to put this together in a way in which communities and states can really benefit.
So it was partly a timing issue in my mind, but if we do not try and grab some of this momentum and say we have got to bring public health communities and population health along then ten years from now, we will still be having the same sort of discussions with a few areas of excellence and the rest lagging behind.
DR. GREEN: What are your reactions to that?
MR. HITCHCOCK: We hear a demand for this sort of thing in almost every, well, every meeting that we get that is related to these sorts of topics – more granularity, community data, state data. It has been a consistent theme lately. I think that is an indication of the demand and why the committee ought to be looking into this. That is my opinion.
MR. LAND: If this is going to end up either as a letter or as a report, what is it that we would hope would be the types of recommendations that would come out of it? We probably need to define those in advance to make sure the presenters are not just given their own experience, but what is it that is generalizable and useful to others.
MR. HITCHCOCK: How are recommendations sort of transferred to a federal role in all of this, maybe – the HHS role.
MR. HORNBROOK: There is also a legal aspect of it. I mean, the FDA has worked long and hard to get the legal right to go in and open anybody’s chart in the country to find out about drug safety. And not even ask the patient. You have to give it. You can stonewall it and see what happens to you.
They have about, maybe, ten million people in the population so far. They have started sending out data requests. They want literal frequencies of every drug that is being dispensed in Kaiser and WellPoint and other big data holders. So somebody has to go in and look at everybody’s record and pull the data out.
The other dimension I was thinking – two other dimensions conceptually, if we want to build a conceptual model, which we may or may not. One is the lifespan. So when you think about monitoring health, monitoring the health of a baby could be a different kind of data problem than monitoring a teenager, a woman of a childbearing age, a middle aged person, or an older person. So there is that notion of monitoring health risks as you go from very young to very old, and that the indicators would change over your lifespan as to where you want to define what kinds of questions are most serious for your health as a population. That does not get into the issue, of course, of frequent versus rare diseases because rare diseases are sometimes kind of random.
Then, the other dimension is family, where family is however the unit defines itself. It is not just a biologic unit and not necessarily a contract unit. It is whatever you find cohabiting together and they call themselves family and the notion of somebody – a unit that is caring for each other. And if there is a health problem with one person it affects the people in that unit or makes it disperse or break apart, if it cannot handle it. So there is the notion, of course, of caregiving going on inside that family. Families that have very strong caregiving skills, people are healthier than in families that cannot handle the disease. So, again, your family unit becomes, in a sense, a kind of a health indicator. Think about a homeless person who has no family. They are, by definition, pretty high risk. I do not know whether you want to, as part of our product, build this kind of conceptual framework as a vision for pushing the government forward, despite the fact that we are in an economic recession and nobody has any new money to do anything new. At least we are reminding ourselves of what we should be eventually doing.
DR. STEINWACHS: Well, one of the things, certainly, that NCVHS has done in other areas is to try and provide a vision. I think conceptual models sometimes helps people understand what that vision is so it is not just a bunch of paragraphs about this age group and this age group and this data set.
MR. HORNBROOK: Also, across the lifespan also brings a intergenerational – so if the intergenerations are alive, there is more support than if they are all so far apart, they are dead, the older generation. Or you do not have any children so there is no younger generation coming up to take care of you as you get old. So when you look at cancer in the family, you know, there can be generations up and below the cancer patient, and that has a big implication for how the – what is the implication of disease inside the family?
DR. STEINWACHS: So on the family, what – the source of information is most likely census, in the sense that you get some idea of what the units that are living together are?
MR. HORNBROOK: Right. So then if you could link into the medical records and get a disease or health problem list on each person in that unit, you have got then a mini-cluster of health problems. Then you can say, okay, how many of these people have chronic health problems? How many of these people are just getting short-term acute or they are perfectly healthy and never went to a doctor in five years? Eventually, of course, you get into preventive services. Are they getting access to healthcare to get their immunizations, to get their health counseling, to get taken care of before they get so fat and overweight and a bad diet so you cannot fix it anymore? The damage is all ready done.
DR. STEINWACHS: You are getting at something that if you had the information and you had linkages, then you could observe both, preventive engagement and some of the things that are on the President’s agenda, trying to push prevention in this country –
MR. HORNBROOK: So I would say, for example, you would want to know among, let us say, grade school and junior high kids, how many of them are smoking.
DR. STEINWACHS: I will never tell. I did not want to tell at that age.
MR. HORNBROOK: Evidently, the proportion of kids trying it is very high, however, they do not get addicted until they are in their twenties when they can buy them. If they cannot buy them, they do not use very much, but the fact that they had that first exposure and inside that exciting kind of how do I get to adulthood fast enough and how do I tell my parents to buzz off kind of thing that drives smoking.
DR. STEINWACHS: But I heard the modern generation is friends with their parents now. Not like us.
MR. HORNBROOK: Their parents buy the cigarettes and the liquor and hand it to their kids or make it available. Yes.
DR. GREEN: You know, I am hearing Mark flesh out more details on our priority list – one and two – the word linkage. You are talking about different types of linkages across different data sources and different typologies of data. If Nancy were here and she was championing the linkages to the social determinants of health, you are saying we need to think more broadly about that. We need to think about how we link things to families, people who say they are, and we also need to create linkages to individuals that are in cohorts that we follow over a lifespan to see how things evolve longitudinally.
MR. HORNBROOK: And employers, school systems, criminal justice, education. All the things that are connected to all of us at different points in our life, that are ways that we interact with each other and also ways that we come into contact with health problems.
MR. HITCHCOCK: SAMHSA has been monitoring youths’ recollection of prevention messages, and we just got a briefing the other day on their Health Survey on Drug Abuse and Health where there was gradual increases in the use of some drugs – ecstasy was one of them, I think probably marijuana too. While the use was going up, the recollection of seeing prevention messages was going down and also the perceived harm of abusing these substances – there was less perceived harm, as well. It is not causation, but there is something interesting going on there.
DR. BOENNING: So, the fried egg commercial did not work?
DR. STEINWACHS: Is that the brain – your brain is now like a fried egg? That was why I never used those drugs.
DR. GREEN: You work Friday nights.
DR. STEINWACHS: He sees me regularly.
DR. GREEN: Stumbling across the threshold.
DR. STEINWACHS: But, it is not much different, in the sense that when you look at the – you know, everyone is concerned about the economy growing back, which comes back to consumer attitudes and the idea that you would tap into us as consumers, broadly, and you would say, well, our positiveness and willingness to spend money has gone down, therefore we know the economy is not. I think this comes in the same sort of context. You cannot draw a straight line and say it is cause and effect, but it is a strong correlate of the kinds of things you are interested in.
People say – well, it is like cataract surgery. I remember, when it became outpatient, it grew tremendously. It was not because – the ophthalmologists were getting less money now than they were before, so it was not the ophthalmologists out there selling it. But, I think it sort of said, well, this is not very risky. You know, you do not have to be in the hospital to have cataract surgery. You just go outpatient and you it will be taken care of. So a kind of subtle message, but a very powerful one to many people.
MS. JACKSON: You invoked the name of Nancy Breen. I was just thinking of her as well because she would wave that mantra of social determinants of health. I see you trying to jump from that to the next stage as to just how we pull it all together logistically, when the technology is changing so fast. But, the convergence of the issues is really right on target. I want to make sure that you all recognize and appreciate the information in the agenda book, where your slides, Don’s slides, with the 50th anniversary – with the NCHS presentation, Linda Bilheimer’s summary is in here, as well. So a lot of this is just kind of converging. It is just getting the packaging together with the logistics and the technology.
DR. STEINWACHS: So, in some ways, it seems potentially very doable.
MR. HORNBROOK: I was thinking about the notion of beachheads. You know, someplace, somewhere where either we get ASPE to fund a demonstration project where you try to make these data links in ways that should have been made a long time ago, but there were so many barriers against sharing welfare data with criminal justice, with educational system. If you work out the right kinds of protections and solid encoding, encryption of data, you would be able to begin to get it to the indicator level where it is not just me, it is an aggregate number.
DR. GREEN: Is that concordant with the community data initiative?
MS. CAIN: Yes. I think they are putting in a lot of data. Now, what I do not know is to what extent you can link those things. I know you could look at individual things for a number of different communities. Presumably, you would be able to pull up a range of things across an individual community. But I am not sure the extent to which there is linkage within that initiative. I do not know.
MS. JANISEN: Excuse me, this is all NCHS survey data?
MS. CAIN: No, no, actually a lot of it is not because we do not have the county level data. So it is coming from other places. That is one of the questions is what comes in and what kind of standards should there be for data to be in it.
MR. HITCHCOCK: I do not know if we get it from EDA or Justice. We will at some point, I am sure.
MS. CAIN: Talking to other departments?
MR. HITCHCOCK: Yes, pulling data in from the other departments and collect it.
MS. CAIN: Right now – where they are right now, I cannot tell you exactly what is in it, but there working very hard to get this version 1.0 out by December 1st. So all kinds of people are coming and saying, well, you could add this, you could add that. If they start trying to do that, they will never get version 1 out. They are pulling a halt to everything – all the new things coming in, but they want to build on it. It’s just that —
MR. HITCHCOCK: No shortage of ideas.
MS. CAIN: Exactly, no shortage of ideas.
DR. STEINWACHS: So it used to be that the Agricultural Extension Service knew what was happening in every county. It could bring us back also to a kind of meeting like the data linkages. It gets the different parties together and talk about what is there, what could be done, what are the facilitators and the barriers.
MR. HORNBROOK: There is the notion of starting with a micro-level data like a health records system, and then taking the GO identifiers – so the person’s address, which could be linked to a track to block a county out and then pulling in the data that are available for each of those geographic units across that population. Then you have got a huge fishing pond.
MR. HITCHCOCK: There is another woman named Nancy at Harvard. I cannot think of her last name. She has been working with geospatial presentation of exactly those kinds of data.
SPEAKER: Nancy Krieger?
MR. HITCHCOCK: Nancy Krieger, yes.
MR. CAIN: That is the other idea is that you get all of your data and aggregate it up to a particular geophysical unit, and then you have homogeneous set of observations on census data.
MR. HITCHCOCK: That is great. I just wish I was 50 years younger for this reason only so I could fool around with geospatial data, look at, what is it called grocery store deserts or whatever they are where you really cannot go out and buy good food.
DR. STEINWACHS: But you can find a bar in those areas.
MR. HITCHCOCK: You can find a bar.
DR. GREEN: So I am being haunted by Garland’s last question where he said, what sort of recommendations would come out? The product that we would get out of this workshop – I think we made some real progress in thinking this question through, but can someone consolidate that a little bit more.
DR. STEINWACHS: I thought you were going to volunteer to consolidate that. So, let us just, the areas we identified were, first, the numerator and denominator issues. Part of that, it seems to me, is understanding what is possible in terms of linkages like between census data, electronic health records. Has anyone tried it? So is it just – I was hoping it was not purely theoretical. Because I thought New York City and Massachusetts and some of these other places, that they are doing things to try and utilize electronic health record data to do population health estimates.
MR. HORNBROOK: It seems to me there are geographic areas where, politically, it is just nearly impossible to share data. Oregon is one of those places. There are other places where they have already started saying data are safe. Let us share them. Let us open them up. They are getting much better able to put together the university medical center and the leading community hospitals and the leading, bigger physician practices, as well as the rural clinics and the inner city urban clinics. And they are all getting together and saying, we are in this together. We need to understand our health care needs of our population. In Oregon, everybody is out for themselves. They do not want to share any data. So the hospital association owns the statewide discharge abstract database. Nobody gets it unless you agree to all of the terms that the hospital association puts on you and pay for it. So you cannot link Medicare and Medicaid claims to the hospital discharge abstract data. Whereas, Maryland and Washington State are examples of, you know, the discharge abstract data are practically a public utility.
So I keep thinking, you know, just because I live in a data poor area, geographically. And I tried to convince our CTSA that one of the most important things that OHSU and Kaiser could do would be to build a statewide public health database based on our clinical data samples. But, OHSU said, no way, that is impossible, that is not in our best interest. Nothing came of it
DR. GREEN: But what a nice vignette to make the case that there needs to be – I am getting convinced that what we are messing around with doing here is in embarking on a revisioning project where that sort of event is like a trigger that is like a spark that says that is just not the way to think anymore.
MR. HITCHCOCK: How do we – how do you all operate as a committee? You are advising the Secretary, essentially. Whatever goes out from this effort would go presumably through the Secretary. I mean, what more can we do other than show her that there is a demand for these sorts of data on a granular level and recommend what, again, the role of the Federal government is – what the Federal government can do to help.
DR. STEINWACHS: I would agree. If you wake up in the middle of the night and have dreams about this, which I try to avoid, it seems to me that if you had the capacity to take in NCHS and sort of take it up to the next level, part of that would be working with states and communities and helping them both have an infrastructure, have the metrics, have some of the training and technical assistance maybe they need, so that they would have a kind of data system running. That is certainly – when you talk about CDC, we think of CDC as being the agency that works with the states and the community to do these things. Even though, generally, CDC has focused on some chronic diseases and certainly infectious diseases and injuries.
MS. CAIN: And, we do do that at Vital Statistics. I think that system is pretty well in place.
DR. STEINWACHS: Yes. So it would really be building out on something like that.
MS. CAIN. And we do do that to a certain extent with our community HANES program. There are issues with that in that all the community needs have been funded by the community because we do not have the funds to do that. Then we also have to think about, right now we do not have – right now the people who work on the community HANES are doing that in addition to their regular jobs. So if we were to build up this capacity to provide assistance to communities, it really would be fairly extensive building up the programs and the people in there, in terms of having a community assistance.
MR. BOENNING: How does Community HANES dovetail with the NHANES survey?
MS. CAIN: Well, it varies. It is the NHANES group that has been working with the Community HANES. We have done one – excuse me, New York City has done one. We provided a lot of technical assistance for that. Wisconsin is doing one. Oregon is doing one. Then we provide the technical assistance, and it really depends on what the community wants.
MR. BOENNING: Do they come to NCHS and we want to use your technology to get an assessment of health status in our community?
MS. CAIN: Can you help us, yes.
MR. BOENNING: And then, they take care of the funding?
MS. CAIN: They have to fund it, right. They get the funding from wherever they get it. When Frieden was in New York, he was very supportive of that and provided the funding for the New York City HANES. I am not sure where the other money is, but it is not from HHS. Other than built into these there actually is some contractual arrangement where the communities do provide some funds to NCHS to cover costs, but there tends to be a lot more given than is actually covered in the budget
MR. HITCHCOCK: It is NCHS equipment?
MS. CAIN: No.
MR. HITCHCOCK: It has got to be standardized, otherwise –-
MS. CAIN: Well, they standardize on their own. We do not give them our vans, I do not think. I know we have lent vans to Canada.
MR. HITCHCOCK: I do not know. The standardization is very important part of this process so everything can be directly comparable. The laboratory stuff can be all over the map unless you adhere to very strict protocols for processing and a lot of controls – that sort of thing.
MR. BOENNING: So do they make their data available or is it like Oregon, they sequester it?
MS. CAIN: It belongs to the community. They do what they do with it. As far as I know, nothing is publically available to just get in and analyze. It is up to the community.
MR. HORNBROOK: Where do you have to go to analyze it?
MS. CAIN: You have to be part of the system. As far as I know, the New York City Health Department analyzes the data and to produces reports.
MR. HORNBROOK: So if you are an employee you probably have good access. If you are not, you probably are going to have to go through some sort of IRB and pay for it and all that kind of stuff.
MS. CAIN: I do not know if they would. I am not sure whether they would give it out to other people.
DR. GREEN: To get local for just a second, I inquired of our School of Public Health. I make various inquiries to various folks that are knowledgeable about health reform, health policy, that sort of stuff in our state school of public health before I fill in my priority out here on this thing. By the way, theirs matched up with the way this came out. But we have a very active CTSA group for community engagement that have spent two years struggling with – running into the issue in Colorado, that the existence of the data bears little resemblance to the availability of the data for use.
I would like to link this back to Dale’s or someone else’s comment that we are not anticipating a boatload of money to do a bunch of other new stuff. But another potential answer to Garland’s question is to expose through this workshop, the very promising and almost immediately available opportunities for linkage of things we are already buying, already spending money on, already producing and already exist, which in a time of austerity and frozen budgets and everything – I am looking for a way that this workshop actually is not just dead on arrival because it aspires to do things that the Secretary or no one else has any way of doing it. And, this strikes me as being a sweet spot that, you know, you have invested this much already, surely we can invest this for gaining access to and getting this used, but it needs to be linked up properly.
I do not think this is answered. I do not think this is a redundant exercise from where I sit and I welcome for that to be refuted. If anyone can point to a place in the United States, maybe it is New York City or somewhere else, maybe it is Massachusetts, where they have done this, where the different potential users of health-related information actually have the means and the mechanisms in place where they can ask and answer the questions and, in our instance, and our last item of our survey of ourselves, in a way that could monitor what is going on during a period of health reform, I do not know where that is. I do not know where that exists. Billions and billions of dollars – now, I feel like I am playing Bill Scanlon.
DR. STEINWACHS: Well, Bill is gratified to know that his presence is here. [Laughter.]
DR. GREEN: Bill would just be sitting here saying, you know, how are we going to know did anything move in a proper direction or not? That just strikes me as being so timely. And, to go back to something you said, Don, I think the way you framed this about 30 minutes ago to get a lot of traction is now that we have decide to spend billions of dollars on electronic health records, where is a proper and proportionate investment in the public health information system to take advantage of this and to align with it properly. Isn’t that relevant? We should try to stimulate that. That is your dream. That is what keeps you up at night?
DR. STEINWACHS: Right, that is when I wake up in the middle of the night and I have run out of things to dream about.
DR. GREEN: I think we ought to decide to do this.
DR. STEINWACHS: Well, I agree with you. So what I was going to suggest is that Larry and I, since I think we have to do something special because we are co-chairs, is that it? Why don’t we try and draft sort of a paragraph on each of these three areas. I think, try and make a statement about why now. And, Dale, pull in some things you have said about why tackle this now and the investment we are making. Then, I think, back to Garland’s point and what Mark has been saying, too, we ought to talk about what near-term types of recommendations might come out of this, in terms of what might be the targets of recommendations that could come out of this if we look at these issues that could help accelerate it. And, that, it seems to me, merits some broad discussions, so Larry and I can start trying to put down what we think we heard here, but then have maybe one or two conference calls both to sort of solidify that, but I think also, to try and talk about what would – is this one workshop with multiple sessions and do we take one of these topics first and then move on to something else. And, then, I think Debbie was suggesting, this is probably something more toward February, which gives us some time, but it would be nice to think that by springtime we would have a product.
MR. HITCHCOCK: I think you are right, absolutely. But, I think a way to frame this, really, I think we would want to hear from at least three groups. One of them is Todd Park. It is a full committee?
MS. JACKSON: My note is that we wanted to hear from him, either, the full committee, December one and two, or if you are going to have anything before then. It is just hard trying to pull anything together in November, but he is definitely on board with us.
MR. HITCHCOCK: But, Todd, and then somebody from CDC in Atlanta that understands, part of the director’s office, just what they are trying to do in this area, and then, three would be to revisit with Linda Bilheimer on where they are. Although, if it is December 1st that it is going live, she might be all tied up with the going live on December 1st.
MS. JACKSON: Yes. I am looking and seeing that their products are due – they are posting a Health 2.0 Developer Challenge in October and then this thing is going live in December. A lot seems to be pulling together at the end of this year.
DR. STEINWACHS: Debbie, when our next meeting is —
MS. JACKSON: December 1 and 2. We have got an executive subcommittee November 30, the day before the full.
DR. STEINWACHS: Oh, I was thinking maybe we could do something in junction with that, particularly, if you wanted to patch together Todd, CDC, Linda and have a half day session.
MR. HORNBROOK: After Thanksgiving, right?
DR. STEINWACHS: Yes. So, Debbie, why don’t we talk about it. If we could do something like that and do what you were talking about, Dale, is get some of the key people that represent, within HHS, and their thinking. That would be very much help and probably finalize the molding of what these –
MR. HORNBROOK: Yes, I was just thinking about this. There is one scenario where let us say you ask a health plan, X, which has electronic medical records, to give you a vector of data on every single person, de-identified. You have got their age, their gender, and their health problem list. Okay. So you have got a morbidity profile of hundreds of thousands of people, maybe millions, and then you have got it geocoded, either to the health plan or to subdivisions within the health plan. So you have got a disease prevalence, plus you have also got people who have chronic diseases that are going to stick with them through their lifetime. So you have got their hypertension, even if it was diagnosed 20 years ago, they are still hypertensive. You may also get paraplegia or quadriplegia, in which they have been in a wheelchair now for 30 years. It is or is not part of the health status. That is one notion.
The other notion is that you go out and send out invitations to every single person saying, we want to create a data system to measure health in your community. Would you be willing to let your data be part of that? You do not have to do anything. Just sign this permission and we will go to the health plan and extract your data in a secure way and we will put you in this public health dataset, just like you do on a research study and assure them that they will be held confidentially. You will get a volunteer selection bias issue, which we always do with all of our research. That fits in with the traditional sort of NCHS population model because you already preselect who you are going to mail it to. So you already have some kind of sample design and you start off with some definitions of everybody you want to include in your sample and then go after them.
So I do not know how this kind of thinking around the fact that there are now these medical records that have been going on since 1997 or even in Utah earlier than that, is important or not. The VA, of course, is now nationwide. Hospitals have the Cerner and other companies, which allows you to download electronic data from the hospitals.
DR. STEINWACHS: It would be great to try and explore and be able to have some demos where you could actually test out and compare. A little bit like if you have the NHANES locally and you compare that to what you get out of lab tests and records and say well, what kind of biases are there with the VHR-based lab information versus what you with the representative population?
MR. HORNBROOK: Well, if you have the VRFS(?) data, you have the health problem list, and you have their certainty lab test values – kidney function, liver function, or medications they were taking. You could drill down pretty far without literally going in and putting them through the van and drawing blood, as you would do for the health examinations. So if you have BMI, you have a tremendous handle on obesity — if f you have weight and height over time
DR. STEINWACHS: So, Larry, do you think we ought to put Mark in charge of this project?
DR. GREEN: Maybe. That is exactly what I was thinking.
MR. HORNBROOK: Only if I have 70 billion dollars to try something.
DR. GREEN: Okay, I guess not.
DR. GREEN: So I would like to ask all of the staff and the people in the room here, having heard this, would you please speak up if this makes no sense to you or you are saying I cannot believe they are going to do something when they should have been thinking about such and such. Are sitting here with those sorts of reactions? [Laughter.]
MR. HITCHCOCK: Dale has never led me astray consciously.
MR. BOENNING: I think that getting the EHR vendors involved, which I think you mentioned very early on, is important because they design their systems sometimes I think almost in a vacuum. And, then, hospitals buy it, the whole kit and caboodle, and a lot of things could have been designed into the system earlier on, and that is not the case. So if there are ten key data elements that you want to monitor in the next five or ten years, I would figure out a way to incentivize the vendors to include them and make it easy to record the data as a process of normal caregiving, not, oh, and in addition, I have got to get your height now, and I have got to measure your shoe size. I mean, keep it relevant to the clinical visit.
MR. HITCHCOCK: We used to hear all about the business case for these additional items – any kind of encounter that the physician had to enter and how much it cost per key stroke. I mean, it got to be really ridiculous. It was like a ten cent job, but by the time you got through the encounter it was millions.
MS. JAMISEN: So you are saying, partly, to take advantage of the opportunity that we have now that there are people building their EHR systems right now? That is a timing opportunity that we have now that we may not have some other time. That is the only thing that I feel like is that we need to take advantage of this particular point in time where we are at, not something we could have chosen – I mean, we could have looked at the limitations of data. The limitations of data linkage are – exactly, many of them are not new.
MR. HORNBROOK: Well, NCI and FDA are paying millions of dollars to those of us who are willing to standardize our data so that you can put the query out across multiple systems, have the data all be in the same format, have one program that will run in Vermont, it will run in LA, it will run in Seattle, it will run in Florida.
MS. JAMISEN: Is this the caBIG?
MR. HORNBROOK: caBIG is another initiative, but I was thinking of the Cancer Research Network, HMO Reseach Network, and Mini-Sentinel Network. The VA is relatively standardized – relatively.
DR. GREEN: This four-way exchange is another good answer to Garland’s question. This is the type of recommendation that could derive by next spring. If we could get the right people defining it, we might be able to drill down a little bit to some of this potential specifications.
MR. HORNBROOK: So, the Mini-Sentinel is focusing on drug claims and dispensing records, depending on what kind of system you are on. So if you are an insurance company, you have got drug claims when drugs are covered. If drugs are not covered, it is another issue. But then dispensing data are the automated pharmacies or the automated medical record, where you have got the drug order. But dispensing is better because the patient actually picked the drug at the pharmacy.
DR. STEINWACHS: You know they got it.
MR. HORNBROOK: Whether it goes into them or not, at least they got it. The other thing, of course, is they want that to be linked to the chart and you have to provide FDA proof that the chart is accessible and you can open it. When they see an event, they are going to want to pull the charts in and audit them to find out all of the rest of the things that are going on around that event. You know, somebody had a heart attack or some serious kind of complication, expected or unexpected. So what we are talking about is trying to detect not necessarily an adverse drug event, but health status, health status improvement, health status decrements in a population with a lot of potential markers.
DR. STEINWACHS: Okay, well, I think we are coming close to the end of our time.
MS. JACKSON: The other item on your agenda, I appreciate (?) – NCVHS subcommittee, just to put on the table, the interest of privacy, that Leslie was trying to be here.
DR. STEINWACHS: So, Larry volunteered. You heard him volunteer, didn’t you, to talk to Leslie. As soon as she came back, she left again.
DR. STEINWACHS: She looked at us and said I am getting the heck out of here.
MS. JACKSON: We have got to salvage that letter. It will come back up. I just want to make sure it was in the record.
DR. STEINWACHS: So maybe, Larry, we could talk to Leslie about three and four.
MR. HORNBROOK: Each one of those – you are all aware of this – but each one of those elements that are classes of special information could be sequestered. You could make very strong cases for when it is at risk to the patient to sequester that data. It is not in the patient’s interest to do that and they are relying on the profession judgment of the provider at that point not to hold that data back, even though the patient’s requested it be sequestered. Or to let the patient agree that they are going to have to suffer the consequences. These are tough ethical issues all around.
DR. STEINWACHS: I think, particularly, it seems to me if you change, fundamentally, the structure, such as the certain structured areas of sequestering, then you get into usual practices. You always do it all the time or you say to the patient, well, I do not think you really should. It gets to be a very uneven system.
DR. STEINWACHS: Well, Mark’s having second thoughts about this already.
MR. HORNBROOK: No, no, no. It is how does the privacy issue jive with our indicators issue? That is what I was trying to figure out. And, it comes down to whether you are going to go to the informed consent avenue, where you have to ask everybody to let their data come forward for indicator. Or if you have some sort of public health authority that measures whether or not you have health insurance is something that should be reportable. Congress has to write a law and get it passed.
DR. GREEN: That is another concrete potential recommendation out of the work is, about an hour ago, you said maybe we look at some of the ways this could happen and then identify the barriers to it.
DR. STEINWACHS: Larry is available to go out for Congress not this year, but in two years from now. Right?
DR. GREEN: No.
(Whereupon, the subcommittee adjourned.)