[This Transcript is Unedited]
Department of Health and Human Services
National Committee on Vital and Health Statistics
February 25, 2015
National Center for Health Statistics
Auditorium
3311 Toledo Road
Hyattsville, MD 20782
TABLE OF CONTENTS
- Welcome – Jim Scanlon
- Population Health–Action – Bruce Cohen and William Stead
- NCHS Update – Charlie Rothwell
- NHIS—Early Release Program, Update – Robin Cohen
- Workgroup on Data Access and Use – Vickie Mays
- Public Comment
P R O C E E D I N G S (8:10 a.m.)
MR. SCANLON: We will call the meeting to order. What we will do is I will review the agenda in a minute. I guess we are having a couple of more members coming around the table, and we have folks on the phone.
The agenda for this morning is unchanged as we had planned it. We are going to spend about the first hour or so on population health issues and plans. We will take a break at 9:40.
We are fortunate to have Charlie Rothwell, who is the director for the National Center for Health Statistics, come over and abate us on some of the plans and activities of the National Center for Health Statistics. Right after that, we will have Robin Cohen, who is with the national health interview survey here at NCHS.
Remember at our December meeting, the Committee was interested in how we are measuring health insurance coverage and health reform impacts? We wanted to emphasize the population statistic side. So Robin is going to talk about some initiatives we made to the health interview survey, both in timeliness and in some of the information we are collecting to look at the impact of the Affordable Care Act.
At 11:00, we will have a report on the plans and activities of the Working Group on Data Access. Vickie and Damon Davis will join us. Then we will close it up from the full committee meeting at about 11:30. We will leave time for public comment. Then we will adjourn the Full Committee meeting at noon, and we will open up the Working Group on Data Access and the meeting of the working group.
What we will do is start with introductions around the table. Just indicate again if you have any conflicts. This is not emotional conflicts. This is financial potential. That would be a long report (laughter). Just financial potential conflicts of interest. We will introduce those in the audience, as well. Then we will go ahead.
I am Jim Scanlon. I am the Deputy Assistant Secretary for Planning and Evaluation at HHS. I am the staff director for the full committee, but for this meeting, I am the acting chair person, as well.
MS. JACKSON: I am Debbie Jackson. I am the interim executive secretary, also acting.
MS. JON PAUL: I am Tammara Jon Paul, CDC, NCHS.
MS. GOSS: Alix Goss, Pennsylvania eHealth Partnership Authority, member of the full, member of the sub committee on standards, no conflicts.
MS. KLOSS: Linda Kloss, member of the full committee, co-chair of the Privacy, Confidentiality and Security Subcommittee and member of the Standards Committee, no conflicts.
MS. LOVE: Denise Love, executive director National Association of Health Data Organizations and co-leader of the All Payer Claims Database Council, member of Full Committee, undeclared subcommittee, no conflict.
MR. ROSS: Dave Ross, Public Health Informatics Institute, Emory University, member of the full committee, new member of the population health subcommittee, no conflicts.
MS. KANAAN: Susan Kanaan, writer for the committee.
DR. CORNEILUS: Lee Cornelius, University of Maryland, population health subcommittee and full committee, no conflicts.
DR. COHEN: Bruce Cohen, Massachusetts Department of Public Health, member of the Full Committee, co-chair of the population health subcommittee, member of the data workgroup, no conflicts.
DR. STEAD: Bill Stead, Vanderbilt University, member of the Full Committee, co-chair of Population Health, no conflicts.
MR. SOONTHORNSIMA: Ob Soonthornsima, independent consultant, member of Full Committee, member and co-chair of Subcommittee on Standards, no conflicts.
DR. SUAREZ: Good morning, everyone. Walter Suarez, member of the Full Committee, co-chair of the Standards Subcommittee, and member of the Population Health Subcommittee, and I think all the other ones, and no conflict.
MR. BURKE: Jack Burke, member of the Committee, member of the Subcommittee on Privacy, Confidentiality, Security and Population Health, no conflicts.
MR. TANG: Paul Tang, member of the Data Access Workgroup.
MS. JACKSON: I have a couple of announcements before we get started. For callers, thank you for joining us. Please mute your phones when you are not speaking. The participants at the table, and anyone using a microphone, speak slowly and clearly, and I will do that myself right now. Please, as Jim has already demonstrated, move any and all portable devices away from the microphones. Thank you.
DR. CHANDERRAJ: Raj Chanderraj, member of the Full Committee.
MR. SCANLON: So I think we are ready to turn it over to the Population Health Subcommittee for continuation of yesterday, but plans and activities.
Agenda Item: Population Health–Action
DR. COHEN B.: We will start the morning with a really fun interactive exercise. Does everyone have the landscape colored version of the table entitled, NCVHS Strategy Triage Worksheet, February 2015? This one is color-coded.
Jack, Paul and Raj on the phone, I don’t know whether this was emailed out to you, Jeanine. If it wasn’t emailed out to you, it will be appearing in your inbox shortly.
The goal today is sort of carrying on from the discussion. The population block this morning will be composed of two halves. The first half will focus on roundtable recommendations for a Secretary letter and some guidance and some guidance for planning for the population health subcommittee. I will lead this discussion. The second half, Bill will lead on framework activities.
For the first half of the discussion, we will use this matrix. We have had a lot of discussions in the population block and in the general committee about the need essentially to reboot or reinvigorate HHS data enterprise. The big enterprise, not just data collection, but data dissemination. Sort of converging with the roundtable themes, our goal this morning is to identify short-term and longer-term strategies that we want to recommend or evaluate in terms of whether we recommend them to the Secretary, we work with partners or we think they are priorities for future work for either the entire NCVHS committee.
Our goal this morning will be to try to get a consensus for some of the recommendations and strategies that are in the left-hand column as to whether we think they are ready to move forward as recommendations to the Secretary in the short-term, or they need more work for future activity. Is everybody with me so far?
DR. STEAD: To state that in slightly different words, victory today would be if we come out with one, two or three checkmarks in the left-hand column where we have consensus. These are areas that we know enough, that we are ready to make a recommendation that would be in a letter that the committee would approve.
As we do that, we want to know where the ones that aren’t in that column land with a particular eye to picking, if we were going to just do one convenient activity in this space in 2015, which would it be? We might land that one in the fall. This is an attempt to, in essence, give us a little bit of a systematic way of coming to those two decisions. Is that a fair way to restate it, Bruce?
DR. COHEN, B: Thank you, my esteemed co-chair. It is really great having co-chairs.. So here is the list of possible activities. If there is something that does not appear on this list that you think should be raised to the level of a near-term or future priority, for the population subcommittee or for the committee as a whole, related to this particular space, please feel free to offer that up, as well.
I think the best thing for us to do is maybe spend five minutes of quiet time for individuals to go through this list to sort of evaluate where they would put checkmarks next to these activities, rather than us going through each line. Then after, folks can work alone or in pairs, if you want, to discuss after a couple of minutes, we will get back and discuss. I will ask for folks recommendations for activities that they think are ready to include in a Secretary’s letter. Okay?
DR. TANG: And I think on the phone, we haven’t received our copy yet.
DR. COHEN, B: So it is on its way, Paul. Sorry for the lack of pre-preparations. Take five minutes to look over this list and identify your priorities for short-term versus long-term.
(Pause)
DR. COHEN, B: So let’s begin. Let’s hear what some folks feel are ready for us to develop recommendations for the Secretary in May. Does anyone want to start with suggestions?
DR. SUAREZ: I would start with I don’t know why I have a lot of recommendations to the Secretary. Maybe create a community of community of practice would be one area. Now, some of the things, I guess, I check-marked two columns because it is not just a recommendation to the Secretary, but also a pitch to partner with.
DR. COHEN, B: Thank you for doing that. The lines are gray. They are not mutually exclusive. They can be certainly worked.
DR. SUAREZ: That is in the first block of the communication one. Then in the second block, I marked pretty much the first four as areas where I think we can make a recommendation to the Secretary there, finding a home within the federal government, aligning with HHS strategies, developing funding for new activities and including links with other federal departments. I think those are actionable.
Then in the other ones, align data and support platforms. That one I actually checked for recommendations to the Secretary and pitch to partner.
MS. GOSS: Is that platform to the reference architecture?
DR. SUAREZ: No, that is the align data for platforms.
MS. GOSS: You intended to use platform in that sense. Thank you.
MS. KANAAN: It is the intermediary entities.
DR. COHEN, B: There is a proliferation of intermediaries who present data in a variety of forms and communities. Should they go to healthdata.gov? Should they go to county health rankings? Should they go to community commons as their data source? It is figuring out how to create some consistency primarily among data suppliers.
MS. KLOSS: How would that be our job? I mean these are private groups.
DR. COHEN, B: For me, it would be more pitching and working with partners, rather than recommendation of the Secretary. We are just going through our list of what we think we should recommend to the Secretary.
DR. SUAREZ: That one, I actually checked both recommendations to the Secretary on page two.
DR. COHEN, B: Let’s finish Walter’s list, and then we will keep going around the table.
DR. SUAREZ: On the back page, recommendations I have reference architecture in the first block. In the second block, expand TA and system of regional centers. Again, the phrasing of the recommendation would be not necessarily create by June 1st system of regional centers, but more explore the possibility of creating. That is one.
The last three recommendations I have, which I check both recommend to secretary, to partners education, and data stewardship for state, for communities and for citizens. That was my list, I guess. I was just going around.
DR. ROSS: Just a point of clarification as a green bean to this committee. When we say recommendation, what does that mean? I was kind of expecting, when we said recommendation, that somebody has got wording prepared that what gets recommended is something that is clearly within the scope of the committee and is something that we are urging the Secretary to act upon. Is that right?
DR. STEAD: What this is, is an iterative process. Given the document we approved yesterday, which summarizes sort of what we have found and where we are, what we are hoping we can do this morning is to identify the handful that we think we understand well enough, that the Population Health Subcommittee can go offline and draft, iterate back and bring back to the Full Committee in the form that you just described as recommendations at the May meeting.
This is a generative task in that process. There have been two suggestions for process, as we move forward. I will let Bruce figure out what he wants to do with them. One was that after some of this initial discussion that instead of just going around one by one, we might have a show of hands to get some consensus. Another that it might be helpful if we had someone actually noting in the chart that is up, if need be. Those are just two process suggestions.
MR. SCANLON: Category two, federal hub for community health, it is a little out of our lane unless it is the data part of it. Do you mean for health data? Our lane is the data part of this, not the community health programs. Is that what you mean, a hub for community health data coordination or access? Is that what you mean there? If you are thinking about creating a program or something at CDC, it is a little out of our lane.
DR. COHEN, B: Susan and Bill and whoever was at the roundtable, please feel free to chime in. I think the notion is that across HHS, there is not a lot of coordination or understanding of community-oriented data programs. The focus here is trying to align those, so that different agencies aren’t recreating the wheel. It is focused around data and technical support. Is that your understanding?
DR. CORNELIUS: I have a recommendation that will get us through this a lot quicker. I think it would help if we would start first. Since Alex is up doing the chart, if we can just go through Walter’s list of checks and get that and get everybody, so that we have the boxes filled. Then we will deal with the consensus. Otherwise, we will be talking about this all morning.
DR. COHEN, B: Should we start there? Maybe the best way is just to go through these one by one and see how many are ready for prime time. Raj, Jack and Paul, remember to raise your hands high, so we can see them.
The first one is NCVHS learned more about existing federal activities. That is not really a recommendation. I think that is an ongoing activity.
Mutual awareness among federal agencies and intermediaries. Somehow to frame that as a recommendation to the Secretary. How many people think that is ready for primetime.
MS. LOVE: I didn’t think anything in this cluster was a recommendation. It was informational.
MR. SCANLON: Keep in mind who our client is, particularly HHS, when we go through these.
DR. COHEN, B: Well, certainly leverage opportunities to learn from partners is pitch to partners. Create a community or communities of practice. How many people feel that is ready?
MR. SCANLON: That would be in a partnership, presumably?
DR. COHEN, B: Yes.
DR. SUAZREZ: I marked it as first and second.
MS. LOVE: I don’t see that –
DR. COHEN, B: As a recommendation to the Secretary?
MS. LOVE: I see those two, the leveraging and creating community practice, as exploring opportunities. I am struggling a little bit with what creating community of practice would entail. More staff time, more money. What infrastructure or — see, I am new, and I missed a lot of stuff here.
DR. COHEN, B: Let me take thirty seconds to explain. We wanted to continue the fruitful discussions that occurred at the roundtable. One way to do that was to establish a listerv, a community of practice or something like that, or communities of practice.
Chris Fulcher from Community Commons offered to host and support some kind of interactive, online activity to continue the momentum of the roundtable. We have had further discussions about whether that makes sense or whether there are other vehicles, or whether we should explore a federal site to continue this conversation and other conversations. That is essentially the idea.
DR. ROSS: When I look at this block called communications in networking, what I am reading between the lines is that there is the possibility of recommending that NCVHS be staffed in a way and in a level to monitor, and therefore do our own surveillance, of what is going on in this whole space and be the place within the HHS that one could get information about all of these things. The conclusion would be that you have to facilitate a dialogue. That is the community of practice idea.
It really is the recommendation that this committee be staffed in a way to do that. No, that is not the recommendation.
DR. COHEN, B: This first block really is more work plan for us to explore these ideas than shovel-ready in many ways when I reexamine this, if that is okay. I don’t want to short-circuit the conversation, but I want us to be able to move on.
DR. CORNELIUS: I am just trying to be practical. I understand that your goal for this block is to focus on the recommendations for the Secretary. All those four columns are great and wonderful. Maybe we should go through this list with regards to the recommendations now, and then go back and deal with the other columns. Once again, we need to move forward.
DR. COHEN, B: So we have a consensus that no recommendations of the first block. Second block? The block is labeled federal hub for community health data. That is around activities, essentially aligning community efforts across HHS agencies. That would essentially be the nature of the recommendation to be expanded.
How many people think this is ready for primetime for a recommendation for the Secretary? I am sort of grouping this whole block.
MR. BURKE: I did.
MS. KLOSS: I checked number one and number two, and not the rest, assuming there is a home. The issue of funding links within the department, those other things would follow. I thought focusing on a home and aligning with strategic plans is kind of the levers.
DR. TANG: I checked the health statistics modernization initiative. At the end of your process, I could put together sort of a framework for organizing the recommendations.
DR. COHEN, B: Paul, I am sorry. Could you repeat?
DR. TANG: In this group, I checked health statistics modernization initiative. When you are done with your check process, then maybe I can follow up the discussion on how to group some of these recommendations.
DR. COHEN, B: Great, thank you. Again, I don’t want to debate each one of these. The consensus is there is something here that is worth recommending. We need to come up with language that embodies these ideas around modernization. But issues around funding might be out of scope at this point in time. It is going to be incumbent upon us to craft recommendations based on this consensus. There is a consensus that we should recommended some kind of better coordination around community data enterprise. Are people comfortable with that?
MS. GOSS: In regards to the meaningful use connection, that seems redundant to me about HHS strategic plans. It should be incorporated.
DR. SUAREZ: The same with links with other federal departments, that seems to be logical.
DR. COHEN, B: Yes, all to be incorporated into that concept.
DR. SUAREZ: With respect to the last one, the health statistics modernization initiative, NCVHS, of course, in June of last year wrote a letter on the public health information infrastructure development, recommending modernizing public health information infrastructure with five very strong recommendations. My thought is, as a national committee, we could consider linking this concept of the health statistic modernization initiative with the previous modernizations.
DR. COHEN, B: Good. Thank you. We should review our past letters to the Secretary to reinforce this.
MR. SCANLON: Remember, the health statistics modernization, there was some very specific — it wasn’t generally upgrading and improving. There were some very specific elements that I don’t think the committee members are familiar with.
DR. COHEN, B: I don’t remember the details.
MS. KLOSS: On that one, I thought that might be an area for NCVHS fruitful planning and convening activity. We have talked about it, but we haven’t really fleshed it out. I think I am intrigued by Paul’s suggestion.
This seems to be right in our wheelhouse to convene and flesh that out before we go further with recommendations.
DR. COHEN, B: In some sense, the whole modernization approach is consistent with some of the discussions we had at the last meeting and that we were having yesterday about trying to reorient and converge our work plan. We will need to think about how it fits into short-term recommendation versus longer-term, more thoughtful fleshed out.
DR. STEAD: I think we need to move. We have got good input on this. We clearly have agreement that we need to build a recommendation around the first two bullets. Whether we do anything beyond that, we will have to see.
DR. COHEN, B: I will try to keep the train running here. The next one is align data and support platforms. I guess we feel that this is probably an activity for partners rather than recommendation to the Secretary.
MR. SCANLON: I am hearing that it might require more work to figure out what this means before we recommend it.
DR. COHEN, B: Platform alignment. Aligning data in support platforms. We discussed this. This is more of a partnership.
DR. STEAD: It is going to be a partnership that would probably flow out of further convening activities.
DR. MAYS: I think we need to start talking about aligning platforms. That is maybe not a direction to go, in the sense that platforms change all the time. People have different reasons that they are doing it the way that they do. I think if you focus on just the data part, I would get rid of the platform part.
DR. COHEN, B: We are going to postpone this until we explore. I think that will come out. The next one is data visualization. More and better presentation of data by federal data. Should that be a recommendation? How many people feel that should be a recommendation? Is there a consensus for short-term? I see two, three. Is that ready, or do we need more work to make this more specific before we recommend it, or those that don’t feel that this is recommendation worthy?
DR. SUAREZ: The question is what is the recommendation itself?
MS. LOVE: I am for it, but I don’t know what —
MR. SOONTHORNSIMA: Is this tactical. It is so tactical that I am not sure it is a national committee work, unless it is more of a guidance, a strategic level, for example. You might combine that recommendation with align data and support platforms once we get there because they kind of link.
DR. SUAREZ: This is one of those that we can actually check all of the four blocks. We can say some statement as a recommendation. We can pitch that to the rest.
MR. SCANLON: I don’t know what we would do there honestly without an assessment for practices. That is professional practice. I don’t know what we would do particularly.
DR. COHEN, B: Let’s move on to some more of the technical issues on the second page. Recommend reference architecture for web-based query systems. People feel that is shovel-ready.
DR. ROSS: I think it is worthy of some convening. I don’t know if we have a recommendation for it. When you say recommendation —
DR. COHEN, B: It would be that the department develop a way to — it wouldn’t be a specific recommendation of that that looks like, but that the Secretary should do that. Maybe it is something that we should do.
MR. SCANLON: Open data, is it?
DR. MAYS: One of the issues is with the query systems because we have a lot of experience with CHIS doing this, is how expensive it is. But the detail that you can get down to without there be a privacy issue. What we are trying to do is give the community what they want. I am not sure in the query system, it will have to be very simplistic.
People are starting to try and think about whether the query system gives the communities what they want. They want it so detailed. They want the neighborhood this and that. I think CHIS is coming to understand that it is up here at these levels that query system works best. I thought the community wants a drill-down. I think you have to think about this.
DR. STEAD: Maybe it is a reference architecture for state and local data. In essence, that is part of what the framework is trying to do.
DR. MAYS: It is just that your community is asking for something very different. They want drill down my neighborhood. I think then that you want to figure out what the technical issues are. It really is what can the privacy people put into place that would allow you to have drill-downs.
DR. COHEN, B: There are some query systems that states have developed that generate aggregate neighborhood data. I guess my notion here was, rather than having all jurisdictions and all communities have to reinvent the wheel, if there was some common consistent guidelines of how people implement, that would be helpful in this space. I don’t know that we are ready for a specific recommendation about how to do this.
I would like us to think about discussing this more because I think ultimately, this will be a very actionable and beneficial thing to have happen. I don’t know that we are ready to make any specific recommendation now.
MS. KLOSS: I think that comment applies to everything in this category, just to move things along a little bit. It may be a possible recommendation after further development.
DR. VAUGHAN: Bruce, some of that piggybacks onto some existing initiatives. In the interoperability roadmap that some of that is embedded, Rachel yesterday referenced they are moving from the propriety system to the open system, in terms of how open data is going to be published. Some of that, in fact, already does link to active work going on within not just HHS, but within other agencies. I would point to those other partners as that moves forward.
DR. CHANDERRAJ: I am tried to talk to the technologist, and he was not able to help me. He said to intervene in the discussion and let me know that I am here to work.
DR. COHEN, B: So my sense is for these technical issues, well, the third one, let’s go down to, second, small area estimation. Are we ready to make any recommendation to the Secretary about improving small area estimation or focusing on that? I see heads shaking no. Dave?
DR. ROSS: Well, I think the small area estimation sort of statistical theory problem is the problem that drives the answer to the community level. What Vickie said is right. Communities want to know about them. It is a small area problem. I don’t think we have something that has both sound statistically, that is also then sound in terms of privacy policy.
If we could get there, then we open the gate for the kind of information communities want. At best, I think this to me is something that is really important, if we are going to move communities’ use of data forward. Maybe that is something this committee can act upon.
DR. MAYS: We had a discussion with Bob Kaplan, who indicated that HRQ is going to take this issue up. This is the kind of thing where it is too bad that our NIH person isn’t here, to get them to take it up. It is something that is lost within the budget of NCHS.
Here, it isn’t about community data. It is almost like it should be pulled out of this, and a separate recommendation needs to be made. I think there could be a convening about this topic. I think it is a different set of partners that should come to the table about this and start with HQ since he indicated that they are going to go down this path of funding work in this area.
MR. SCANLON: I am kind of heading that way, as well. I think HHS, it was probably a year ago, we had a workshop on small area estimation. We had all of the federal data producers and others. It was very interested what was underway. There was interesting research.
I think this might be a fact-finding developmental sort of thing. There are all sorts of reasons why you have or don’t have small area estimation. Privacy is only one of them. Maybe it is more of a developmental one, where we do a little more fact-finding. Maybe present the findings from our earlier workshop.
DR. COHEN, B: We are sort of running out of time. There are a couple of items here. One that we began discussing around standardization. I guess looking at the rest of the list, expanding TA, regional centers, educational and stewardship and security, sentential indicators and filling data gaps. Are there any in these areas that folks would recommend for inclusion in a letter for May?
MS. KLOSS: I think we are ready for some level of education on stewardship, so I checked those three. We have got framework. We have got a tool kit. We have got something to go forward.
MR. BURKE: I did, as well.
MS. KLOSS: I think it will cement the offer we had yesterday from the Office of Civil Rights, to help us disseminate the tool kit and perhaps get some additional humph behind that. That would be helpful.
MR. SOONTHORNSIMA: But the recommendation will be different. That is the question.
DR. COHEN, B: I guess at this point, this is an area to explore to see if we can develop a recommendation around it for the Secretary. The wording, we can’t create the specific recommendation now.
MS. KLOSS: I think we are ready to frame a recommendation.
DR. STEAD: What I am hearing is we can opt out. We can draft a recommendation in this space. If we draft it well, it has a high likelihood that we can get it across the line.
DR. STEAD: Are we ready with something around technical assistance or not?
DR. SUAREZ: I marked expand TA and training as one.
MR. SOONTHORNSIMA: That was our recommendation for public health.
DR. COHEN, B: Okay. Trying to craft a recommendation around expanding technical assistance and training.
DR. STEAD: So let me just make sure. We think we can get something around the federal hub for community health data, limited to the first two lines. Nothing else on that page. Then, TA and education, those could be packaged as a first tier of high-level related recommendations.
What I would advocate, Bruce, is that we will now set up a series of population health subcommittee calls to draft these recommendations and to take a first cut at bucketing where the key other things go in these other columns based on what we heard in this conversation.
DR. COHEN, B: I think that is a great summary, Bill. The one that we haven’t gone to, that I just wanted to hear what folks thought about, was the core set of sentential indicators, whether there is a convening or a recommendation, whether that is ready or not. Further work? Okay.
DR. TANG: This will tie in the other recommendations, at least from my suggestion. I wonder if I am looking at an organizing framework of why, what, how and for what. When you say the core set sentential indicators, I think that be the why that we can hang our hat on.
One example, a lot of people know that the IOM had the committee looking at core metrics. It doesn’t have to be that, some organizing priority for the country. It can be an outgrowth of the National Quality Strategy. That points to the new alternative or advanced health models that need to be formed, in order to serve these HHS priorities. That fits in with the Secretary’s January 26th sort of timetable for how quickly we move towards these alternative health models.
Then the health statistics modernization initiative could be the departments’ or NCVHS’s response to that. They have this core set of where do we want to go as a country that is somewhat different from where we have been because we have access to a different set of data in different ways. The health statistics modernization initiative is the way that the department can respond by somewhat recasting the data it has now to support that mission. That would mean that we would have better data presentation. It doesn’t necessarily mean different data sources.
How would we point the data? How would we help explain the use of the data we already have for these new and natural priorities? How would we expand the education and the regional TA potentially through these communities of practice to help local communities be able to take advantage of the new data presented in new ways towards this new goal, the triple aim essentially.
And mind the federal levers, everything from the payment, just like the Secretary announced, the meaningful use, all of the federal levers will point everything towards the new models, towards the new goals and use the data in somewhat different ways. That would allow us to then sort of think of, in the context of the statistics modernization initiative, what other data sources and ways we surfaced them, would we take advantage of?
We talked about before, well, now, actually the majority of the country now has EHRs. Actually the majority of the country has patients, people, connecting through their online systems. Are there different ways that we can tap into the data and the access to the data sources, i.e. consumers and caregivers, to produce the outcomes we are trying to measure in the new health model towards these new national goals?
That is sort of the way that I was trying to think of how do we build a compelling and cogent story around the kinds of initiatives we have on this checklist, as a way of communicating to others and ourselves. Why are we doing this? What is it we are saying that should be done? Who do we need to reach, and for what?
DR. COHEN, B: Paul, thank you. You are always great at being able to step back and present a unifying vision. This is really helpful. I think, as we write the letter, as we try to craft the letter to the Secretary, that context of presentation will be a really helpful way to think about things.
There are three comments here. I see Lee, Vickie and Susan. Then we need to move on from this discussion.
DR. MAYS: I want to fully support what Paul is saying. I think one of the things we want to do is to remember what our lane is and where we do best. I think the whole notion of us getting in a conversation of what the core set of sentential indicators are, that is not it. But instead, it is playing this role of making sure that we are offering guidance about we heard the community say that they are trying to solve a problem.
As Paul is saying, it is like pulling it together to give the Secretary advice of the things that need to be addressed in order to get the community what they need. We don’t want to get into the indicators. It is like that is a really big job. We want to say the reason they are having this problem is because we don’t have this alignment, and this is what we need. Then it can be done.
It either can come back to us as an assignment, it can go to the IOM as an assignment. I think we want to be careful about not taking on the work in this letter, but instead, pointing to what the needs are and a solution.
DR. CORNELIUS: My recommendation deals with creating a process for working through the rest of this chart. Given that we have the chart both here, and some folks may have it on the phone, I would recommend that we may set like a time like a couple of weeks down the road, where the committee can send their feedback about these other three columns. We can then take it as a committee to start working through the other items that are sitting here.
DR. COHEN, B: Great. So we all have a homework assignment to look at this chart and try to complete it, given this conversation this morning. The population health subcommittee will work on pulling all that together to present and create a conversation about what our priorities are. In many ways, this conversation is essentially we didn’t have a lot of elaborate slides around our work plan, but you are helping us create our work plan. Thank you all. Susan, last comment.
MS. KANAAN: I wanted to make sure that we enthusiastically accept Paul’s offer to create a structure and some language. That is number one.
DR. COHEN, B: Paul, there is a consensus vote here. Congratulations.
MR. TANG: I am sorry. The sound is broken up.
(Laughter)
MS. KANAAN: But we heard you loud and clear. The second thing is I don’t think you have an electronic copy of this worksheet. I will be glad to revise it before it is sent to you. Now that you have had a chance to see the origins, what I put in the column to the far right, just to kind of map where things came from, I thought maybe I would just delete that, so that you have that column for other things.
But if you can give me some guidance as to whether you would like to keep that sourcing information in the far-right column. We will send it to you as is. Send it as is, okay. All right.
MS. JACKSON: Beginning with the end in mind, thank you so much, Susan, for helping to pull all this together and then just understanding the organization of what members are trying to do. Just to hear this from the full committee membership, we do appreciate all the work. Thank you.
DR. STEAD: Can we now bring up the slide with the framework discussion points? We are going to switch gears to talk about the framework white paper that was in your agenda book. This is a reformatting of the workshop report. There has been a fair amount of additional of material to it.
For Dave and Denise’s benefit, the framework project is one of our looking ahead efforts to come up with a way of facilitating data classification and use. Our goal here is to develop a draft of a set of data and methods classification resources that we are now calling a data structure, which I think is a more approachable word than data continuum, and a methods taxonomy, which I think is more appropriate than data categorization. Thank you to Jim Walker for those two changes.
Our hope is that these resources will be helpful to communities and others as they work with data. One of the things we added to this issue of the white paper were three very short scenarios. One on page three, one on page eight, one on page thirteen. They are against gray backgrounds. They try to walk through how the framework might be used, first in its nascent form and then in a little bit further development form, and then in a vision when the framework was quite developed and a community of partners had built interactive tools to take advantage of it. At high level, that is what this is about.
Our goals for this block this morning were to walk through the discussion points that came up as people gave us feedback to what the earlier version that we circulated to the population health subcommittee and to the framework workgroup. Those points were highlighted as comments. They did not come from me, although they were my editing. That is why they have WS by them. I put at the top of this slide the list of them to sort of make it easy for us to walk through it.
Once we do that, if we have time, I want to then begin to discuss this one-pager, which has been passed out to you. It is a first draft, which for those of you that haven’t lived with me, means it is version 0.0 of what the third category in the methods taxonomy might begin to look like, the category that has the analytic and visualization techniques.
If we have time, I am hoping that, as we sit at the bottom there, what is the process we could use to begin to iterate this and to build it into the taxonomy. We will then need to come back and do the same thing for the stewardship block and the standards block. Before we have got to really a version one of the framework, we have to do those three things. I am just proposing we figure out how to do this one, if possible, this morning. That is what I am trying to do. Are people good with that?
Then let’s start with page seven in your agenda book. The first discussion point that came up, I believe maybe Lee, you may have raised this. Is there a better word for what we mean by the voice as a characteristic of a data source?
I will sort of explain what I was thinking about voice when I put it in the data categorization. I was thinking, is this the patient speaking, as if you will, in a self-report? Is this a clinical person observing and recording? Is this an administrative source of the data?
The purpose of knowing the voice that the dataset represents is one form of knowing, one source of understanding bias of that data source.
MR. SOONTHORNSIMA: Is that a perspective? Whose perspective?
DR. STEAD: This refers to the perspective of measurement. I think, Lee, you were the one that raised that. Is that better, perspective of measurement?
MS. LOVE: That would confuse me. To me, voice or perspective would confuse me. To me, I look at the type of data source. I think that is implied if it is a survey, if it is an observational study or secondary data source. Then I make assumptions, I suppose.
DR. STEAD: What we are trying to do with the framework is actually force ourselves to make those assumptions explicit because we are trying to get them out where people can see them.
If an observational trial could obtain data by a patient’s self-report or by a clinical observer, if the thing that is being measured is a social determinant, you are going to get two very different answers. If we don’t know whose perspective is being represented, then we have got a bias problem.
DR. COHEN, B: So on this table, I would re-label the first block, rather than to just call it data characteristics, and make E source. For me, the source would be record, clinical record.
DR. SUAREZ: When you look at data, there is data source, and then there is the method of the collection of the data itself. When I am seeing a patient, and I am collecting data from the patient by virtue of asking them a question, and they answered, the source, when I put it in the medical record, is me, as the doctor. The method of collecting it was the patient provided it to me.
The different story is when I measure blood pressure with a machine on a patient, and I record that.
DR. STEAD: Let me ask you to look at page 17 please, which has the appendix that has the one more level of detail than the table in the figure. It has this type of data source, electronic health record, personal journal, community datasets, national surveys, payer datasets, et cetera. That is where we put data source. I don’t know if that helps at all.
DR. SUAREZ: There is nothing about method of collection there.
DR. STEAD: We have not populated it yet.
MS. KLOSS: I would speak in favor of keeping voice. I think that, no pun intended, but it speaks to me. It connotes the flavor of the information that you are collecting.
MS. GOSS: To some degree, you have to take the table, at 17 and 18, to really get the context. I think it makes it clear to me.
DR. CORNELIUS: Under the method of collection, there is usually a quick cheat sheet face-to-face, phone, mail, internet. That is the continuum of mode of design of collecting data. That may be one way to overlay that against this other item of voice on page 17.
DR. STEAD: Ob.
MR. SOONTHORNSIMA: I am okay.
DR. STEAD: Vickie.
DR. MAYS: I am just getting lost a little bit in terms of the complexity of this, in the sense of for the audience out there. We are flipping back and forth. I am trying to understand how to make sure that it is going to be usable. We are trying to do use stuff, as well.
It is almost like it requires people who are very familiar with data to do this. Who is the primary audience for the framework, to make sure it is going to be usable? It is pretty complex, where we are going.
DR. STEAD: The user of the framework will be both people and computers. It is intended to pull out of the experts the information that is implicit in how we set down and structure analysis, visualizations, et cetera. And to set them side-by-side, so that when picks method A, we can then say, this is the strength of this method versus that method. It is, in essence, designed to give a systematic way of doing that.
The framework itself is going to be complex. The only good analogy I have for this is really the unified medical language system and its methathesaurus, which now maps thousands of vocabularies. It is semantic relationships, et cetera. We are talking about something that complex because we are trying to make explicit.
Think of all of the challenges we have in getting data providence explicit. This actually would give you a taxonomy to be able to go check, check, check what that providence was based on. Does that help at all?
DR. CORNELIUS: I actually come at this from a different vantage point, when I thought about voice. I think about community utility. Even as you look at the chart, I think about let’s say we are in a town hall meeting. The community is trying to use this data. How the framework and how the data comes to live is going to be different than what it may be mean if I am in a hospital institution, analyzing the data. The challenge we are trying to balance is the full spread of the usefulness of this tool kit.
From the other side, just like when I look at the charts, it is a keep it simple, sweetie. All these items on the chart are great. However, what is going to give that community the highest impact in using that data. We use the word “shovel-ready” right now.
When we think about both the voice and the chart, the question I would ask us is it is good that we are talking up here, how does this make the community feel, I can take this tool kit. I can make it come alive because these are where I hit the buttons to get what I need out of it.
MS. KLOSS: I think we should go back to the introduction. I don’t think this was ever intended to be something that we send out to communities and say go at it. This is a way for us to describe this complicated universe we are working with and have a common framework. I think the purpose statement is pretty clear.
MR. SOONTHORNSIMA: It is really not for community leaders per se, but program managers, project managers who are trying to cobble together. They have to be a little bit more technical. They have to understand. But it provides a common framework. Here are all the dimensions that you need to think about, depending on how deep you can get into it.
I think that is how you come alive to your point, Lee. At some point, how you present it back to the community. This is at the next layer or a couple of layers down, at a more technical, more problematic level.
DR. CORNELIUS: Is that where we really want to go?
MR. SOONTHORNSIMA: If you want to address public health, yes because there is no consistency across the board yet. That is why we started doing this.
MR. SCANLON: I am thinking just again, all of you from your disciplines, there are whole literatures of textbooks and courses on this, social science research, public health. We are not trying to duplicate that. This is, I think, an attempt to bring in informatics framework for the folks who actually are intermediaries and maybe data producers.
It is just sort of how then it relates. It is not a consumer-level. If you want a primer for social science research, that is different. That is not this. These are all known to social science researchers and vital statistics and so on. I think this is an attempt to put it into a framework that could lend itself to computable.
DR. VAUGHAN: Had you anticipated how this might be a part of or be associated with metadata that travels with the dataset throughout its very iterations?
DR. STEAD: Our assumption is that, for example, a dataset would be tagged with the metadata of where it fit in the data structure. It would be tagged with the metadata of voice, et cetera. Then analysis would be tagged with the metadata of the analytic techniques. And a visualization would be tagged with the metadata of the visualization techniques. That would, in essence, let people develop systems that could make that clear. It will not make it clear by itself. You could use it.
We are clearly not going to get to it today, but when I did the one-pager Sunday, I actually hadn’t seen that kind of taxonomy of data collect, of analytic techniques. It helped me just to sit down and think about putting them into a taxonomy. There are many rich descriptions of the pieces. But metadata is one of the things we would use it for.
DR. MAYS: We are trying to think about these things in the access and use group. What is the benefit of this? I am paying a programmer to do my work. I am paying them. I am trying to understand. Why do I want them to use this? What is it about public health? What is it that is going to make —
DR. STEAD: Let’s stop and walk through the scenarios because they were an attempt to try and describe that. If they don’t —
MS. LOVE: On the aggregators, before we get to that, couldn’t we just say limitations? Aggregators, judgment of fitness for the purpose? I have problems with that term. As a big data supplier in my past, we made a public dataset. We didn’t take responsibility for other people’s bad judgment.
We made it available to everybody, but there were certain principles. One of them is list the limitations and the data, which we made very clear. I am just saying, instead of aggregator’s judgment, wouldn’t you want limitations?
Every dataset should list or document its limitations, but not so much a judgment. I have a lot of judgments, but my limitation should drive it.
DR. MAYS: I think you really hit upon my point, which is if I am a sophisticated user, because that is what I kind of thought we were doing. If I am a sophisticated user, I am going to make these judgments myself. It seems like the way the judgments are being made, there is a framework to make the judgments. I am trying to understand, if I am used to using data and stuff, give me a sense of how this increases what is something.
Otherwise, it is almost like that is one — but I can decide. The dataset says this, but my creativity says I can use it for that. I am trying to understand.
DR. STEAD: Which is fine. Limitation is fine. What we want is a taxonomy that would let us systematically give you a way of saying, I am going to check limit one, limit four, limit six. They apply to this dataset.
In essence, trying to end up with some standardization about how you say what the limits are, so that we can then say, okay, this dataset has these two limits. This one has these three limits. I am actually getting ready to use them together. I need to pay attention to that. Does that help?
MS. LOVE: Yes. It is just I think I agree with Vickie. It is way beyond most community — again, I am coming late into this.
DR. ROSS: It is a taxonomy. It doesn’t mean that everybody has to check every box. It gives us a comprehensive, complete way of understanding datasets. I can see many benefits of having this.
DR. STEAD: Let’s sort of take a process check button. We are about out of time. I think we need to do a couple of things maybe to help with next steps. If these scenarios are not useful, can one or two people agree to take a point on drafting alternative scenarios that might communicate better?
I think we do need some vignettes, all the scenarios. These vignettes were anchored in appendix two, which was our deep dive in the workshop around the data needs of community health assessment and intervention. It is anchored in appendix two. We tried to cascade from the first scenario being something that would make sense if you really didn’t know much about the framework at all. The second one, you knew a little bit more.
The third, partner organizations or start-ups or whatever had built a rich array of applications that took advantage, if you will, of these taxonomies, something we are clearly not going to do. If the taxonomies were available, it would be possible to build those. Right now, they are impossible to build. We don’t have the knowledge sources to build. I think it would be helpful.
I keep looking at you, Vickie, because you are raising the point. It would be helpful if you could say, what would these scenarios need to look like to make this connect in a way that would make sense to you and the Data Council?
DR. SUAREZ: When you say scenarios, do you mean —
DR. STEAD: The scenarios are page 3-A and 14, I believe. The gray-shaded areas were three progressive scenarios. They sit on top of the detailed table in appendix two.
DR. COHEN, B: I think the real issue is really understanding who the audience is for this document and how it is going to be used. Maybe we need to go back and have everybody look at the introduction. If there is a way that we can clarify and make more explicit what the purpose of the framework is and who the intended audience is, I think that would help us move forward in fleshing out. And the scenarios are an effort to give examples of the intention. I think if everybody hasn’t had a chance to look at this document carefully and focuses on those particular sections, it might help clarify some of the conceptual issues that are being raised today.
DR. MAYS: Can I add one more component? That is valued added. To me, that is an important part. It is like I could sit and put this together, if I were using a dataset. It is like tell me how this increases the value of it.
I am going to give you an example. If I gave this to a student, and I wanted them to do some work for me, and they go and they check these boxes. Well, suppose race isn’t collected well. All these problems we have been talking about, they are kind of reduced to a check box.
I need to know something about quality or value enhanced or some limitations or something that you are going to say to the user, to make sure that the essence of the very things that we worry about, that they just don’t check the box, and then we are done. That is why I am struggling with what is added here.
I understand what is added for convenience. But what is added here in terms of us giving the community a better health outcome. There is something about this that we are doing, that should make something better. It is not just a scenario. What are we making better by choosing this?
DR. STEAD: What we are trying to make better is appropriate use and reuse of the data in a date set that has been captured by somebody else. That is, I think, what we are trying to make better.
DR. COHEN, B: The goal is to increase the understanding in a consistent way, so that you can compare across datasets and data elements in the dataset. The search here, again I don’t see it having a specific community application. It is more for people who are providing and serving communities with data, so that they can understand the limitations.
For instance, if you collect smoking data from a clinical record, NHANES collects smoking data in two different ways. You can ask people how much they smoke. You can pull a strand of hair and do biometrics. You can ask a household member.
In order to evaluate the data in a consistent fashion, this is an effort to develop a high-level framework that has that capacity. That would be its use. I guess we are struggling in our conversations about, at the community level, how communities use data and what we can do as data suppliers at some level is better explain the components and what the limitations and benefits are for particular data items. That is sort of the whole intention of this effort, which is very separate from other things that the population health subcommittee is doing. At least, that is my point of view.
MS. LOVE: So would it make it more alive to do an exercise of taking two common datasets and showing how the taxonomy works. I am just thinking out loud. That might bring it more alive for me to show a vital statistics dataset or maybe a cancer registry or hospital dataset.
I have done this where their timing is a little different. Their user purpose was a little different. That might bring it alive. When you are combining them, you see where the —
DR. COHEN, B: We started it at the other end, but I like that approach, as well. We said, let’s take a problem, childhood obesity. At the community level, how would schools begin? What data would they need? How would they classify it? Given that this is fleshed out, maybe we should take the other direction and try to apply this.
MS. LOVE: We are just giving a few examples, not across every data. I think that would be clearer to me. For instance, when you are using two datasets, these fields are important to know because, and give an example. Not do a huge crosswalk, but an example.
DR. ROSS: I think another example is every public health agency to be accredited has to do a community health assessment. I have been a part of working with some of those groups. They struggle with what do these data that we are looking at, really tell us to yield the assessment, which sets the goals for community health improvement over a three to five-year term.
This kind of taxonomy helps them be more precise and understanding what the data they have in front tell them. They still have to reach whatever judgments they reach. To Denise’s point, sort of lining up and understanding, well, gee whiz. I have got this dataset. I have got this other one.
My problem is, do we know enough to say our community childhood obesity is an issue. We are going to go after it. They have basic questions. Then they have to go find data sources and put it together. This taxonomy helps them really challenge their understanding of what their data tell them. I think it is helpful, from that point of view.
DR. STEAD: Linda, then Lee, and then we need to close this down to have a brief break.
MS. KLOSS: I was really going to suggest process. I viewed this work as coming out of population health, but having implications well beyond. We are just kind of at that critical point now.
We always have this come after we talk about other things. There hasn’t been enough time. I would really suggest that we allocate sometime in May to have this kind of applications for this work really get discussed more thoroughly.
I think our world now in health care is awash in data and big data. This does start to create some language and framing for all of this, not just what we are doing in the community health space. I think we should step back and really look at this more holistically in terms of what the national committee can do to illustrate how this is used in several venues. It is more than community. I think we should allocate agenda time to it because I think we have gotten kind of short drift at the committee of the whole time.
DR. CORNELIUS: Bruce, you had actually provided a comment a few moments ago. I kept thinking about what is the target for this? Who is the actor as it relates to all of this? When I saw the phrase, communities, in goal one, I came back to this issue where you talked about persons who are providing for data to help communities may help us to tweak this.
I think about persons and organizations doing that. They are in the community, but they are not always of the community. We want to be very careful since we are saying this at a high level to be transparent about the purpose of this document. I could easily see this running amok as this is disseminated.
MS. KLOSS: I think the conclusion is purpose is. I don’t think we should pigeon-hole this into one purpose. I guess that is my point.
DR. STEAD: I think we have gotten wonderful input. I do like the idea of maybe a plenary first-day session in May. We can fit that in. If you can help us build an example, if we could look back at the introduction and give us suggestions about how it could be tuned to come into line. Then my sense is we then need to take this offline to the framework workgroup and iterate it a little bit. We have got work to do. This has been very helpful. Thank you, everybody.
MR. SCANLON: Thank you, Bruce and Bill. We are going to take a 10-minute break. Then when we return, I think Charlie Rothwell from the National Center for Health Statistics will give us an update.
(Brief recess.)
MR. SCANLON: Let’s proceed. We are very pleased to have the Director of the National Center for Health Statistics with us here this morning. Charlie Rothwell is the Director. He is going to update us on plans and activities at NCHS. We have a full-year budget, Charlie, so I think we are starting off the right way.
MR. ROTHWELL: Absolutely. It is good to see some old friends. Actually, this is a pretty good year. I don’t think you will ever hear anybody that works for the government saying that they have all the money they need, and everything is just fine. If they do, there is something wrong with them.
Things had been a lot worse for us. I think what I have to tell you is pretty darn positive. I think the future looks good. I would like to share that with you.
First of all, I hope you saw the banners when you came in today. Sometimes you may not think it, and sometimes our staff doesn’t think it, but we are a part of CDC. We are a unique organization in that we do have protections around us, so that we are not programmatic-type of data collection activity. We monitor things that relate to CDC or elsewhere in the department. We are much like the Bureau of Census and BLS in that mode.
However, we have encouraged our staff to participate in activities. We have had a variety of staff, both epidemiologists, as well as statisticians and data mongers, to go over to West Africa. From physicians, again to statisticians, and we have had them over in each of the three countries. They have done well. Some have gone back for a second tour. Some are lobbying for a second and third tour. Some are over there, and they are extending. I am extremely proud of them.
These are people who volunteered. They did not have to go. Many of them are just plain GS employees. They are not commissioned core. They are not anything of that nature, and they raised their hand to go. I can’t tell you how proud I am of them. I think our country is coming to grips with what they have done and coming to grips with what they have to do when they come home. That was not the case initially.
Enough of that, let’s get onto statistics. Well, let’s get onto people. Nat Schenker is our new deputy director as of this month. Nat was formerly the head of the Office of Research and Methodology. He is a great statistician. He was past president of the American Statistical Association. We are really happy to have him.
Susan Queen, we have been able to steal away, will be next month to take over our office. Basically what it is, is our outward look. How we communicate with Congress, how do we deal with legislative issues? How do we deal with our friends and also planning for where we should be in the future?
We have not had these positions filled for a long time. It is about time that we have it, so that I think the office of center directors is more proactive than we have been in the past.
We are in the process. I have made a selection, but I can’t name it, someone who is going to be heading up our IT and data dissemination activities in NCHS. I think those three things are quite important from a personnel perspective.
From a budget perspective, we are looking good. We have basically been flat lined. That is good. There were some things that were good and bad. The president has, for a number of years, put in budget requests for us to use part of the prevention fund to fund our surveys, to expand them, so that among other things, we can look at what is going on in health insurance in the nation and be able to say something from a state perspective.
Recently, Congress has wanted to move those dollars elsewhere. Fortunately, CDC Atlanta stepped up last year and gave us some money. I thought that was a one-year time, but Tom Friedman decided to take pity on us. He gave us some money to help us again.
Our problem is that we are no longer, and this is getting into the weeds a bit, but just to let you know, we used to be funded by evaluation dollars. Those evaluation dollars come from, if you will, it is a gross term, but it is called a tab, if you will, to participate. Various organizations pay that tab in HHS. Then that money then is spread and was used to cover us.
That is no longer the case now. We are in the budget, which is good because that means that we are there. It is not something that could be just sort of brushed away or cut without anybody knowing about it. That is not to say it couldn’t be cut.
The other thing is that now that we are not in it, guess what? We can be tabbed. That means we are down about three and a half million bucks. So what Tom gave us, part of it will be taken away. Right now, I haven’t made the decision yet on where those dollars, the difference of those dollars, will be going. We really want to continue to with HIS.
You are going to be hearing from Rob right after me about coverage and health insurance. The reason why we can do that is because the questions were expanded considerably years ago with these dollars. We also expanded the sample size of HIS, so that we can give state estimates, which is important obviously for health insurance because that is where the action is, to see what is happening in various states and what the different coverage is going to be.
From a public health and health perspective, it is the first time we have been able to really give health estimates and be able to do that for states. I think, to me, long-term, it is much more important than the initial issue, which is how is health insurance coverage going to take place?
By the way, one would want the expanded coverage in the future in HIS, so that we can see whether it has really made a difference. It is not health insurance coverage, is it? What it is are we healthier? Are we receiving care earlier? Is that appropriate care? What is the outcome? That is something that we are going to have to be measuring for years to come.
I am happy to say that I was going to be coming here and giving you a different script, and saying that we weren’t going to be able to do that. We weren’t going to be able to say much about states, and that would really hurt us.
Robin is going to come in and talk to you about health insurance coverage. What she can’t say is what is going to happen in March. That is our next quarterly report. That quarterly report will be, for the first time, I think where we will be able to say something about not all states, but some of the larger states. Then the next time around, we will be able to really say something about states.
I think even though we were with great trepidation worried about the first two quarterly releases about 2014, these next two, and especially the next one after, the March release, will really tell a lot. I am looking forward to it.
For the coming fiscal year, the president’s budget looks good for us. They put in about $5 million to expand electronic death reporting. This has happened in many years. It has been couched in different terms. It has really not happened. That doesn’t mean that things haven’t happened. Whenever the president says, this is where I want to put my money, or I would like you to put your money, it gets attention. That certainly has been the case in vital statistics. This is one of the things that I really wanted to spend some time with you on.
We also, through some organization called ASPE, got money for this year and for some following years to again fund the collection of mortality data to speed it up even more, so that basically we would have some 22 states or more providing us data, 80 percent of their records within 10 days of the actual event of the death. I think that is very much a possibility.
They are also going to be providing us money to link up our health care surveys that include emergency departments with our national death index. I will talk about the timeliness of the national death index in a moment. I think that will be some interesting studies that can come up with that dataset, looking at what are the survival patterns that are taking place. That is greatly appreciated.
HIS has been the leader and NCHS has driven, I think, a lot of things that should have taken place years ago. For a variety of reasons, it hasn’t happened. What is going to be told to you after me comes from a set of quarterly reports that HIS has been doing for years. It is basically information that is being provided six months after the data collection period. That is the reason why they were able to say something in 2014 about 2014 data, which is highly unusual for NCHS.
At the other extreme, years ago we were lucky to get vital statistics out two and a half years after the close of the year. Last year, in 2014, we were able, for example, in mortality data, to put out the final data for 2011, 2012 and 2013. We were obviously able to put out natality data for 2013. This is not provisional. This is not preliminary. This is final, complete.
I was looking at some of the data. We have automated systems now that are tracking data coming in on a daily basis to us, some vitals. We are around 40 percent of all vital statistic records are coming to us now within 10 days of the event.
I was asked to go down and talk to the senior staff at CDC at invitation of Tom Friedan to talk about, of all things, how CDC could use vital statistics for surveillance purposes. All of a sudden, we are going to have data, and they had better be ready for it. We are going to be publishing it because we have to publish it. What are they going to do?
I don’t know if any of you are familiar with the MMWR. If you look, there has been a graph back there for years and years and years called the 122 city report. It was there when I started in public health. I can remember my health officer, Jake Cuban, being very proud of that. He was one of the people that was involved, at least he said he was, back when they came up with this idea. I am not sure that was ever the case.
Anyway, we are running dual right now. We are running that report. That report is given directly to our infectious disease partners in Atlanta. They publish it in the MMWR. We are also showing our data that is coming from the vital statistics reports. We are doing damn good.
In other words, not only are we doing damn good, we are faster than the 122 city report, more complete and will be able to in the future provide state-specific information that never has been provided before. I am not sure Ann and others would agree with me yet, but I think that they will probably shove that aside and use vitals as that.
We are also looking at vitals for monitoring a variety of other diseases and outcomes of public health interest. That is what I was down there talking about, as well as looking at high-risk birth outcomes that we need to do. Vitals is going to copy HIS. They are going to start putting out this year quarterly reports. They will be talking about things that happen quite recently. Those will be public-type reports that will be coming out. I am just tickled pink about this. I really think that we will meet the requirements of not just 20-some states, I think we are going to be better than that. I think we will have maybe 30 states providing us 80 percent of their data within 10 days of the event and very soon. At that point, we really, I think, have a gold mine.
I think this is one of the greatest success stories in my time in public health. It is about time that vitals and sanitation, which really started public health in the United States, is finally coming back to being able to use. The real importance here is the timeliness, but there is another thing. Have you ever heard a physician that believes in vitals reporting? They always complain about I was there in the night. I had to deal with this person whom I had never met before who died in the emergency room. The stories remain the same. Part of the problem with quality is that we never came back to them to ask them, did you really mean this?
Well now, we are going to be able to do that. We are going to be able to go back to them, either through the electronic death registration system that they are using then, or the next day and say, does this make sense to you? How we do that, how we balance that and not bother them to death will be the trick. What I am really hoping is that this is going to improve quality. That is, I think, from a public health perspective, the real importance here.
What else do I have? Talking about the future of where NCHS is thinking about going and will be going, we are looking at obviously how to deal with electronic health records. That is either going to be the saving point of our health care surveys or it could be our albatross. That is, how are we going to be able to develop standards, either at the beginning or at the end, so that we can make some use of electronic health records? Then, talk about big data. That will be with the millions of records that we could be receiving in regularly. How are we going to deal with that?
Part of it may be what we are looking at now is trying to consider working with vendors, so that they can build an intermediate type of piece of software that could at least standardize within their vendor community or within their hospital setting a standard set of records that could come to us. Whether that is successful or not, I don’t know. I think it is something that could not only be useful to NCHS, but could be useful to a variety of agencies throughout the country. That is something that we are looking at right now.
The other is, getting back to vitals, why not have that electronic health record of Charlie there for the physician to look at when he is doing my death certification? Why not, at some point in time, take that information and help the physician actually do the cause of death information from that, if that electronic health record is really complete.
Let’s go a little further. Why not have the birth certification, at least the statistical information, kick off the electronic health record? That is where we need to be, in my humble opinion. That is what I am going to be pushing. This is, I think, some exciting times if we handle it right.
I think we are going to need to tell our stories better. As a citizen, forgetting about what my job is, I was really dismayed, should have been surprised, but dismayed at what happened with the measles outbreak. Speaking as a citizen and not as a government employee, I really think that conversation that is now taking place, you don’t want people beating up on people. The conversation does have to take place in the public. The public has to decide where liberties extend to and when they conflict with other people’s lives.
We have a situation where we had physicians, as well as the citizens, not really understanding the complexity and the danger of this. We really need to be able to tell our story better. Here NCHS, and I am beating up on us because I didn’t think about it until one night when I got home and was discussing it at the table. My wife said to me, well, what are you doing about it?
Basically what she was getting at was you have the data, Charlie. You are not telling the story. You haven’t been telling the story right. We have mortality data to back at the turn of the past century. Why aren’t we telling that in a graphic way that gives people an appreciation not just for what we conquered, but what could happen. Why it was conquered and what could happen if we walk away from that.
I think that is a role that NCHS really needs to get involved in. That is, I think the term is, data visualization. We really need to look at that and see how we can better tell our stories, not just for other researchers and policymakers, but for the public. The reason why we can do it is because we are not advocates.
We are not trying to say that we are not from the cancer community, we are not from the – whatever community it is. We are from the data community. This is what happened. These are the things that took place during these times that made this happen. Then let others take it from there.
I think that is an area where NCHS needs to get involved in. There is some debate within my senior staff, as there should be, how far do we take that because we cannot be seen as advocates. I do think we need to be showing our data better. That will be one of the areas that we need to go in.
The other is how do we deal with surveillance data that we are providing to CDC. How do we make sure that we don’t embarrass states? For example, we are getting information, and we are publishing information, faster than now many of the states are. We are getting the information actually before the states get it because they are sending it to us to automatically code the data.
We need to make sure that we give that data back to them in a way and give them a heads-up, so that they can take action on it before our Atlanta partners come down and say, what the heck is going on. All of that is an interesting dilemma, and I am so happy for it. This is a type of problem that I wanted to have.
Other than that, that is about all I have got. Any questions you might have?
DR. COHEN, B: Charlie, I really need to applaud you and all of the folks here. It is a revolution, not an evolution, in e-vitals. For folks who haven’t immersed their lives in the vital statistics system, it is actually 57 independent systems that need to be pulled together at some level to make some coherence of what is going on. Your leadership has really been dramatic in improving how quickly things get published and supporting these jurisdictional efforts.
I thank you from my past job, and I thank you from this committee’s perspective. I wanted to suggest, I have talked a couple of times to Michelle about demoing e-vitals, e-death reporting and e-birth reporting. I think it will be a phenomenal education for all committee members to really understand what it takes to compile these records.
When you are talking about getting these in actionable form 10 days from the event, in Massachusetts, we are getting a little better, but it has taken over six months because of the chain of command and processing. The dramatic turnaround in time is phenomenal. I would like to renew the request to see if we can build in a demo of e-vitals for the national committee.
One other comment, Massachusetts is getting better. I know we are. Is there a relationship? Speaking of Massachusetts and graphics, our annual mortality report graphic cover is a timeline back to 1842, 40 years before the germ theory was proposed around trends and mortality. I think understanding trend data more significantly and using it would be a huge adjunct to what we are already doing in vitals and other areas, as well. I really applaud your efforts to try to promote that.
MR. ROTHWELL: Our staff had a lot to do with this, but it is the states. They got the message. They made the investment. They are sending us the data. I am just really pleased.
DR. MAYS: Much like Bruce, I want to say thank you. Also, I think this notion of the ways in which, I want to say it is a revolution. Just even my personal work in the ways in which the mortality data release has made my ability to link and do things and answer scientific questions just has been incredible. I am so excited about the things that you are doing, both from the perspective of a user and also from the committee.
MR. ROTHWELL: Speaking of that, the NDI, I didn’t mention that, will be updated for 2014 probably the end of this month. That is much quicker than we ever promised. That won’t be just for fact of death. That will be cause of death.
DR. MAYS: The other thing that I want to raise, some data that you have and what it is that you can do about a problem that is a very serious problem. I have been contacted by different members of Congress more than once to help them understand how they could get data on violent deaths.
All of this comes up within the context of the black lives matter and all the different things that are going on around the country. One of the things that they looked at was the fact that only 32 of the states are participating in that particular data collection. They were asking all kinds of questions. I am just kind of sharing with you that this may be a very good time to think about having a violent death dataset that can include more states and figuring out a way to be able to do this, where instead of it being state-by-state, opting in and out.
Congress and others are looking at we need better data. We need better data about the cost. I was saying to some people I went to a data hackathon where it was on police brutality. It was interesting. One of the datasets that they had up there, they had the kids doing the mash-up and all this other stuff.
Right now, if we take this issue of what is happening in terms of mortality and death data, there is a great cry right now for having a much better consistent dataset. What they did was after meeting with them, they showed the southern states are the ones that are not under the federal datasets. They are doing their own. They are not giving the data that is necessary to actually address what is a major public health problem.
MR. ROTHWELL: I agree completely. Our mortality data only tells a portion of the story, though. It has to be enriched by a lot of other datasets like medical examiner records and a variety of records to really flesh out what is going on here.
In this regard, though, we have a project going on with DOD and the VA, and I think I may have mentioned this to you before, with suicide data. The issue there is to make sure we are not double-counting. How do we break these folks out, so that we know that we are looking strictly at veterans?
I am a veteran. If I decide to take my life, it is probably not because I am a veteran. It is probably because of this job. If we are looking at veterans, current people in the military, and people who are civilians, we need to look at the right age groups and make sure that we are not cross-pollinating. That is a problem. We are trying to tease that out by matching up records.
Right now, for example, DOD has given us data so that we can put veterans’ status in the mortality data for the first time that we can believe. At that point, we will be able to make sure that we are looking at non-veterans when we are looking at the civilian population and look at their suicide rates and homicide rates relative to the current, those people who are in the military and veterans.
My guess is the differences, and this is a guess because I don’t know, might not be as high as you think. We are having an increase in suicides. Suicide is a major problem, period. I think the results might prove to be interesting.
DR. SUAREZ: About a year and a half ago, actually Michelle presented here to the full committee the status of the development of E-standards for vital records. I think the demo would be a really great next step. The other question I had was more about the work that we, as a committee, can do to help advance the recommendations that we made back in June to the Secretary with respect to improving and enhancing the public health infrastructure. We had about five recommendations.
I think it is more the question about what the committee could potentially do to help move those recommendations forward, and to what extent there could be some follow-up to the work that we did with the recommendations. That is one other question.
Lastly, if you have any comments that you can make about the health statistics modernization initiative, I guess, I don’t know if there is any concrete that can be mentioned. Health statistics modernization initiative that we have been talking about.
MR. ROTHWELL: I am ignorant about a lot of things, and that is one of them. At least I know I am ignorant about it, so I won’t say anything about it.
DR. SUAREZ: Maybe it is more of an idea at this point than anything else.
MR. ROTHWELL: On how to make your recommendations more known, I think Jim is probably much more in touch with the department than I am. I think that through ONC is a good avenue. CMS is the big gorilla in town. They certainly, I think, have something to gain in this.
I think in Atlanta, Tom Friedman is a data geek and a very impatient one, I might add. I think he is a terrific advocate. There is also another person in CDC, Chesley Richards, who heads up the surveillance activities. He is a great advocate. Those are some thoughts in my mind.
NCHS has to break out of the mode of being seen as reporting history. I think as we break out of that mode, we will be listened to more. I am not sure I carry much weight at this point or NCHS does, to be blunt.
DR. ROSS: Charlie, first, I have to say thank you for your service here. The American public has no idea how lucky they are to have you and your leadership team running this agency. Thank you.
I echo a number of remarks you made. If you want to comment more, fine, but just to say that you talked about storytelling and get the story out there. I think it is a fact that we all agree that facts should matter. They often don’t. They often don’t because they aren’t knowable in a way that people could say, well, wait a minute here. This is what the facts tell us.
I think you are absolutely right. The agency plays a real important role. It should play a very important role in that. I just encourage you that not only the data visualization is important, but then the linkage with the people who turn that into the stories. It is the stories that will actually help the public.
I am one of those who believes that health statistics, this is the bedrock of the public health system. If we can’t count births and deaths in a way that matters, then you almost don’t have a public health system. What you do is, in my view, at the rock center of public health.
A good on ya for speeding it up. We have been in these trenches for a long time. It is great to hear this progress, absolutely great. Anything you can do, or that we could recommend, that helps tell those stories, to educate the public.
The measles outbreak is just an example. We are now a generation of people who don’t understand basic public health measures. I would fear that not understanding how health statistics help us understand the health policy is another risk we run in this country. Whatever you can do to move us down that road towards public literacy would be good.
The one question that I would have to ask you to think about is this area a small area estimate. As you work around the country, communities want information that is both very timely and very specific. These are huge challenges. They are methodological challenges, too.
I am wondering if NCHS is doing anything on the methodology side that will eventually help influence how data get understood and used, to answer community questions.
MR. ROTHWELL: Nat and some of his folks have been working a little bit in this area on how we can do small area analysis with surveys of, at best, state level information, at worst, national. But it is not real strong.
I was very uncomfortable when I was sitting in here and listening to what was going on here. We have very little to say about communities. That is one of our problems. It has been one of our greatest weaknesses.
This is not an excuse, but we have looked at issues of telling the national story as best we can by looking at a variety of racial groups, ethnic groups, as best as we can. That is an expensive thing to do. Collect and to tell an in-depth story about the nation. As that takes place, you don’t have the money, you don’t have the resources to say something very specific and pointed at the community level, or I might add at the state level for some of our surveys.
I think probably the whole electronic record issue in the end will help us at it flowers, hopefully flowers, and hopefully gets some good geocoding connected to it, and that we are able to tease out what is really going on versus what is being paid for. I also think that we are going to have to relook at issues of different types of surveys that are done on the web and that type of thing, whether they are representative or not. I think you need to give folks capability to look at their communities. Maybe what we need is community-based web surveys that people will participate long-term.
I am not sure. I think you need to talk to a younger person. I am too steeped in old technology and old solutions.
DR. STEAD: I share your wife’s passion about telling the story. Visualization and presentation is one of the things that has been coming up, that we have got to begin to wrap our head around. I am just trying to know where you are thinking about.
By telling the story, I could imagine what we would do with measles was this is the death rate for measles. This is the risk rate on a population basis. This is the gap in vaccination. Then is the risk rate for an epidemic. The latter would be the thing that would communicate, to me at least. Is that where you are headed?
MR. ROTHWELL: Yes. I am not sure we will ever be able to get people to evaluate what risk really is. I was talking to my church about Ebola when everybody was concerned that it was right around the corner from them. My comment to them was, take some personal responsibility. They looked at me, what does that have to do with it?
I said, well, drive safely. Don’t text while you drive. I believe in drinking, but don’t drink to excess and don’t drink while you are driving. Get your damn flu shot. These are types of things that you are at much greater risk than the potential for Ebola coming and striking you.
It is just like if you are in the Marine Corps, the last thing you want to do is be worried about the next incoming shell. You get used to that. That can be also, by the way, a danger. We get used to the risks that we have taken.
The problem with measles was is that it seemed like there was no risk. Not only that, there was no risk, and I don’t even know what it is. The other problem was, that is the first time I have hurt my child. I have taken my child in. They got this shot. They cried like heck. Lo and behold, a year later, there is some sort of problem with the child. What happened? It wasn’t me. It wasn’t my genes. It wasn’t whatever it was. It must have been —
Anyway, all those types of things, I don’t know if we will ever be able to solve. But we can as heck try.
DR. COHEN, B: I wanted to revisit the small area estimation issue more generally. From personal experience, I know NCHS is the US expert and probably the world expert on a lot of health statistical techniques. Thirty years ago, I was working with Joe on statistical notes for health planners, which I thought was just a phenomenal product.
I think what we are looking for is not necessarily NCHS generating small area estimates, but essentially not only NCHS, but other federal agencies using their technical resources to help teach people how to fish. I think there is growing expertise at the state and community level, with the proper guidance and direction around doing community surveys, evolving from using some of, certainly not as sophisticated as HIS, but some of the techniques. Making small area estimates, how to use data properly, how to present and publish data.
I think the next generation of statistics is not necessarily collecting more data, but making the data more useful. I think NCHS has been focusing on a variety of larger national issues. My sense is, from this committee, we want to push not only NCHS, but all the federal health and human service agencies to sort of step back and think about how we can better provide this kind of support for folks and comminutes to do their work better. I think this is what I see as a real opportunity for NCHS to assert its leadership in this area.
MR. ROTHWELL: Criticism accepted. There are two areas, well, there are many areas, but there are two areas that we have walked away from. That is that type of information that we are providing to communities and states on how to use data and that type of thing. The other is internationally. Our international program is outside of Sam Notzon and a few others, really is not very strong. That needs to be improved. Those are two areas that we used to be much stronger in.
MR. SCANLON: Denise, you have the final question.
MS. LOVE: I would echo what Bruce said and take it a step further. Helping these states and communities with methods to forecast what is happening. We have morbidity data. I work with hospitalization data and another dataset system that I will mention.
We aren’t looking at that data in a forecasting way. Based on five years of trends, what is likely to happen as far as outcomes and utilization, so that states and prepare for whatever they are planning.
A couple of other things. We have orphan data sets springing up in states. I just start thinking about this at NCHS because we have over 20 states with all-payer claims databases that include pharmacy. I just think that helping us or setting methods for mining and filling these data gaps, even if it is not a taking of data or sentential data. But we really need an NCHS perspective on how to look at it. We have a lot of actuaries who are eager for the data, but not from a health statistics standpoint perhaps.
The other thing that is looming, I remember Charlie, many, many presentations at NAFSIS(?) and other meetings, about when we adopted ICD-10 for cause of death and the bridging. Well, we have got ICD-10 coming. We have a lot of states asking me and others, who may not know. I think this is another leadership opportunity. How do you bridge these indicators? How do you bridge the morbidity data? That will be something else that states will be looking for some guidance on coming forward.
MR. ROTHWELL: Thank you. That will be huge. I know you had the last comment, but if I could have a comment. This is a what-if. We are looking at it. We are just beginning to look at it now. You think about it. Technology isn’t the solution to everything, but this is sort of interesting.
I think I may have reported to you on this before. We were looking at how we could maybe add to HIS through doing measurements, like we do in NHANES. What could we do? What could an interviewer do?
For a variety of reasons, it is very difficult. Whether you are going to lug the equipment up, how it is going to stand up, the liability issues that are involved, a variety of things. Now, we are thinking about what about wearables? This is just huge. It seems like it is just a mushroom pushing up right now.
What if we provided wearables to HIS folks and NHANES? Talk about a longitudinal component within those surveys. They might actually help us keep our response rates up because people want the wearables. Anyway, there is a meeting, which unfortunately I can’t go to, in San Diego, I think, in May that is looking at where this industry is going. Some of our staff will be there with the idea of how can we use this? I think it could be huge.
MR. SCANLON: Thank you. We will go right into Robin’s presentation now. These are some of the, as Charlie was talking about, enhancements that were made to the health interview survey, starting with back in 2011, I think it was. Probably over 100 questions were added. Sample size was increased. The program was initiated to move up the ability to make estimates on pretty much a quarterly basis with a six-month lag. Robin is going to talk about that program.
Agenda Item: NHIS—Early Release Program, Update
DR. COHEN, R: Good morning. I am going to talk primarily about health insurance, but I am also going to touch a little bit about the early release program in general for you.
The next early release report is going to be released the 24th of March. This highlights where the health insurance estimates are. We have a health insurance early release report. We also have a separate, it is eight tables now, that provide quarterly estimates. We can look at things in three-month intervals. We also produce a preliminary micro data file for those who are interested in doing their own analyses. This is made available through the research data centers.
I would be remiss if I didn’t mention that we also have another report that comes out at the same time, which we call our main indicators report. This is 15 indicators. They range from health status to having your usual source of care, flu immunizations, a variety of topics.
In addition, I want to mention that we have been producing these early release reports, I think, for about 15 years now. It is on a quarterly basis. It has only been recently that we have been getting a lot more media attention for them.
Sometimes we also produce some tables of reports outside of our usual quarterly reports. I am proud to announce that tomorrow, we are going to be producing a special report that is going to look at problems paying medical bills. That uses some of those new questions that were introduced in 2011. It will be tracking problems paying medical bills for persons under 65 years of age, from 2011 through the first six months of 2014.
I am going to jump right into the health insurance estimates, which I am really very excited about. In the first six months of 2014, this is what the report looks like. We have a brand new look for this year. We are pretty excited about it. It is very attractive.
38 million persons of all ages, or 12.2 percent, were uninsured at the time of interview. 54 million, or 17.3 percent, had been uninsured for at least part of the year prior to interview, and 27.3 million, 8.7 percent, had been uninsured for more than a year at the time of interview.
Notice that we do provide three estimates of being uninsured. One is the point in time estimate, which is probably your most accurate. It gives a picture of what is happening at the time of the interview. Being uninsured for at least part of the year gives a look at people who are uninsured currently, as well as people who may be insured presently, but had a period of uninsurance in the past year. The picture of those who are chronically uninsured would be presented in those who had been uninsured for more than a year.
This slide looks at the estimates comparing the 2013 full year estimates with the first six months of 2014 for the three measures of being uninsured. This is for persons of all ages. You can see for those who are uninsured at the time of the interview, there has been the statistical decrease. There is a sharp decrease of 2.2 percentage points between 2013 and the first six months of 2014.
There has also been a decrease in those who have been uninsured for more than a year. Again, a similar sort of decrease in magnitude. It is not statistically significant between 2013 and 2014 for those who had been uninsured for at least part of the past year. This suggests that the recent increases or reductions that we see for the uninsured at the time of interview are due to recent increases in coverage.
This next slide looks at these differences by age groups. Again, we saw no change for children. But we did see a significant change for adults 18 to 64. This is for persons under 65 years of age. We saw an even greater change within the young adult subgroup. That was a 6.4 percentage point decrease in the percentage who had been uninsured at the time of interview.
This looks at the number of millions without health insurance coverage at the time of interview by age group. You can see that we saw between 2013 and 2014 for those under 65 a 6.7 million fewer persons who were uninsured. Again, we see almost 2 million amongst those age 19 to 25, which is another group that we saw earlier back in 2011, also saw a dramatic increase in the percentage of uninsured for that age group, with one of the early provisions of the Affordable Care Act.
Among persons under age 65, 62.8 percent or 168.3 million were covered by private health plans at the time of interview. We are really proud that we were able to produce a percentage and population estimate of those who were covered by private plans that were obtain through the health insurance market place or the state-based exchanges.
Amongst that 62.8 percent, 1.9 percent, or 5 million, were covered by private plans through the health system marketplace or state-based exchange. It is important that you understand that this is an average over the first six months of data collection. We may have included a little bit of the surge. Really hold your breath for next month because we will able to see the data from July through September, when things should be stable in terms of that, in terms of looking at the surge. The midpoint for that number is probably about the end of March.
We also looked at selected estimates by selected demographics. I know these slides are fairly busy, but they actually show a lot of very interesting results. We are looking at data for adults. Again, this is with the six-month estimates.
We are looking at those who were poor, near poor, not poor. Poor would be under 100 percent of the federal poverty level, near poor is between 100 and 200 percent of the federal poverty level. Not poor is 200 percent or more of the federal poverty level.
The declines in the uninsurance rate were greater for poor and near poor adults. A majority that declined for poor adults is associated with an increase in public plans. A majority of the decline for near poor adults and not poor adults was related to increases in private coverage, which is I think what we would expect to happen.
This is looking for health insurance coverage at the time of an interview, again amongst adults age 18 to 64. We are looking for Hispanic adults, non-Hispanic white adults and non-Hispanic black adults. Again, declines of uninsurance were significant for Hispanic and non-Hispanic black adults. The majority of decline for the Hispanic adults was associated with increases in private coverage.
More than half of the decline for non-Hispanic black adults was related to increases in public coverage. What we are really excited about is we also produced some quarterly estimates. This way, we can track coverage changes throughout the year. The other estimates were averaged over six months, but these will be quarterly estimates.
When we look at quarterly estimates, and this slide shows percentage of persons under 65 with and without health insurance coverage at the time of interview by quarter. We are looking at the last quarter of 2013, which is quarter four, and that would be data collected between October and December.
The first quarter of 2014, which is data collected between January and March, and quarter two of 2014 is data collected between April and June. We can see from here that among those under age 65, the percentage of uninsured decreased 3.3 percentage points from the fourth quarter of 2013 to the second quarter of 2014.
If you base on a full year of 2013 estimate, 16.6 percentage without coverage. We estimated that 9.8 million persons under age 65 gained coverage between 2013 and the second quarter of 2014. That is almost 10 million persons gained coverage.
If we compared it to the first three months of 2013, we went back even further in 2013, it would be 11.2 million persons. We are in still in that range with 10 million persons gaining coverage.
We broke out by age group. These age groups are a little bit different than those that were presented earlier. The young adult age group is a little bit more expanded. It includes persons 18 to 29. These are the young adults that are eligible for the catastrophic plans. That is why they were separated out a little separately. I think the thought was that the young adults would pick up some of those plans.
When we look at the age group amongst those 18 to 29, and that is the middle group of estimates, about a third of the decrease in the uninsured is due to an increase in public. Two-thirds of the decrease in the uninsured is due to private. Pretty dramatic decrease in the percentage. This is comparing the fourth quarter of 2013 to the second quarter of 2014.
What you can see that these slides show is that the increase in the public coverage or the uptake in the public coverage was relatively early. You really can start seeing that surge that happened. All the activities is happening with the private coverage. There are some differences between quarter one 2014 and quarter two 2014 for private, but not so much for public coverage.
This next slide shows for those adults age 18 to 64. Again, those with and without health insurance coverage at the time of interview by race and ethnicity. Again, we saw Hispanic and non-Hispanic blacks, a third of the decrease in coverage is due to public, and two-thirds is due to private increase. Amongst those who are in non-Hispanic white, it was all due to private coverage.
We also broke out those 18 to 64 by poverty groups. Amongst the poor, the decrease is due to an increase in public, although it was not statistically significant. But there is a suggestion there. Amongst those who are near poor and not poor, the decreases in uninsured are due to increases in private. You can see that there was a 10 percentage point decrease in uninsured rates between quarter four 2013 and quarter two 2014 amongst the near poor.
I know they are a little dense and a lot of information. As I said before, we also provide some data on exchange plans. Of course, we didn’t have the exchange plans except for our state in 2013. We are showing our exchange estimates for 2014 in quarter one in 2014 and quarter two. You really can see that increase in the second quarter, or the surge, that happened at the end of March and beginning of April in terms of pick-up of those exchange plans. It happened in all age groups.
Again, I only showed those under 65 because there are very few persons over age 65 who are eligible for exchange plans. We concentrated the analysis on those under age 65.
This is looking at the number of millions, and it shows the same thing. Again, so in quarter two 2014, we estimated that 6.3 million persons were covered by exchange plans. Again, it captures most of the search in those.
This is looking at exchange enrollment by race ethnicity. We see that there were increases amongst those who were Hispanic, non-Hispanic white, non-Hispanic black. Again, all poverty groups were impacted or you see an increase between 2014 in quarter one and 2014 quarter two across the board. Especially look at that near poor. It went from 2.2 percent were enrolled in an exchange plan to 5.1 percent. These are the numbers again across the board increases.
That really sums what we had in those products that came out last December. We are looking forward to producing another set of results coming out in a month.
MR. SCANLON: Thank you. The March release coming up will be quarter three?
DR. COHEN, R: In the March release, we are going to have in the main health insurance report, we are going to have predominantly, I think we have one chart that shows some quarterly estimates for exchanges. It will show estimates based on the first three quarters of 2014.
In December, we also put out some state-based estimates for, I think, 12 states. We will be repeating that again with the same 12 states for the March report. Whether they were Medicaid expansion states versus non-Medicaid expansion states, we also have some estimates about what type of exchange they had, whether they were in a state-based exchange, a federal exchange or a partnership exchange.
We have some regional tables, as well. We have a really nice map that shows the regionalization of where the uninsured are in the country. You can kind of get a picture of the country as a whole visually, which is really nice to see.
All of that, really it has 15 tables, I think, charts. It is a really packed report with a lot of really interesting and useful information. That is the health insurance report.
Then we also produced some other quarterly tables. You can look at things as they have tracked through the year. With those quarterly tables, we would go back to 201. You can kind of see how things may have been very level or hadn’t changed, and then you can see really the activity that is happening in 2014. As I said, we also sometimes put out some other special reports or tables. Of course, those preliminary micro data files are available through use in the research data center.
MR. SOONTHORNSIMA: Thank you, Robin. I think you touched on it, but I just wanted to confirm. I was going to ask you if you actually separated out some of these states, whether they are expanded Medicaid or whether they are FFM, Federal Facilitated Marketplace.
DR. COHEN, R: All that is available in the preliminary micro data files.
MR. SOONTHORNSIMA: But it is only 12 states?
DR. COHEN, R: No, we have that indicator. That is based on all the states. I have to say it was as of whatever it was, October 31, 2013. I know there are a few states that have switched. We made a decision for this to keep everything. We are looking at trends for this data year to keep everything constant.
MR. SOONTHORNSIMA: So you have the data, but you can actually dissect out of it more and just say are there significant differences between those states that have actually expanded Medicaid?
DR. COHEN, R: Yes, absolutely.
MR. SOONTHORNSIMA: You can draw a conclusion without looking at the data. I have a different question then. They actually have real data in terms of enrollment, and that is what health plans have to report. I wonder how that reconciled with the report here.
MR. SCANLON: I think we are in the ballpark. They are consistent. What we are getting through enrollment is not exactly the same number. What we are getting through enrollment cumulatively is similar to what we are getting there.
For the enrollment data, we actually have those application data and the enrollment. When we publish the 11 million, for example, in QE2, that is from the applications and the administrative data.
MR. SOONTHORNSIMA: Then the Medicaid enrollment is another separate set.
DR. COHEN, R: That was one thing. People seem to understand when we ask them about their health insurance, that they may have, let’s say, signed up, but they may not actually be enrolled yet. That is why I was saying with that quarter two 2014 measure, we have captured some of the surge, but we may have people there who may have signed up, but it is not an active policy insurance for them yet because they haven’t paid.
There was a lag between signing up and payment, is my understanding. That is why we are really looking forward to looking at those quarter three 2014 numbers to see. By that point, that is data collected between July and September. There is not that issue.
You also have to keep in mind that people may have signed up, but they may have also dropped out, as well. They didn’t pay, whatever reason. There is always movement in health insurance. People go in and out of having coverage.
MR. SCANLON: Bruce?
DR. COHEN, B: I have a question and a comment. The question is related to consistency issues. In the past, it is great to see NHIS getting into state estimates. It is fantastic. At the same time, states have been relying on the BRFSS to generate state estimates. A little on CPS, but mainly BRFSS to look at trends in uninsurance and access to care.
My question is as the state NHIS estimates become more into play, how they compare, I recognize the differences in data collection methods and sampling. A burden gets placed on policymakers at the state when multiple estimates of similar measures are released.
My question is more policy-oriented or my comment is making sure that these get processed with states and other parts of CDC, so that any differences that occur in the overall state estimates or subdomains in the state estimates can be discussed and explained clearly because it will create an enormous firestorm in states when these estimates become more into play.
DR. COHEN, R: I know one thing that we do. We put out the state estimates. For the larger states, I know we have looked at how they compared with CPS in the past. Everything tracks very closely. However, when we make our comparisons, we also compare them to how that state compares to the nation as a whole. We make a point. We really are very careful about how pitting one state against another.
DR. COHEN, B: It is not one state against another. It is within a state. I get two estimates of uninsurance actually from the same federal agency, CDC.
DR. COHEN, R: So when you look at BRFSS, they do not collect data on children. We have data on children. We collect all ages. There is no age cutoff on the national health interview survey.
DR. COHEN, B: So actually for insurance questions, BRFSS does collect over — but the action is really in 18 to 64. They also collect household coverages, as well.
DR. COHEN, R: We also have a more complete set of health insurance questions than the BRFSS does. We have lots and lots of questions. We really get our hands down and dirty with the data to really understand what people are telling us. We have a lot of follow-up and detail.
DR. COHEN, B: I am not arguing that. It is just there will be two estimates of uninsured persons aged 18 to 64 for blacks and for Hispanics. Which one is right? That will be the question that the governors and the commissioners of state health departments will be asking. CDC needs to be ready with a reasonable response.
My other is a comment. When we began our initial analysis of Massachusetts insurance data, we found a huge race-age interaction. We found particularly among Hispanics dramatic increases in insurance coverage in age 30 to 44, but not in younger Hispanic adults or older Hispanic adults. We saw the increase in younger persons, mainly among whites.
I know Massachusetts might be atypical, but I really encourage you to look at the interaction between race and age because it really has helped us target populations as we see changes in insurance coverage.
DR. COHEN, R: I think that is something that we do when we have a full year of data. We are trying to get some numbers out on a partial year of data and quarterly estimates. We are more limited in how much we can slice and dice the data.
One can connect the dots, but I think when we have a full year of data, there is going to be lots of opportunity to kind of look at some of those interactions between poverty and race and all sorts of other issues.
MR. SCANLON: CPS actually changed some of its questions to be more like the point in time. We did that. We sort of published both, and they are going to publish that for a while. It probably wasn’t the best time to do that.
DR. COHEN, B: At the state level, I don’t think CPS has as much currency in health departments for policy as the BRFSS, quite frankly.
MR. SCANLON: We are actually tending to use the HIS now as our authoritative estimates at HHS. I mean we have a lot of other commercial estimates and so on. Thank you, Robin, that was very helpful. I think we want to turn now to Vickie, who is going to update us on the activities of the working group on data access and use.
Agenda Item: Workgroup on Data Access and Use
DR. MAYS: Okay. What I did was also put in some extra slides, so that we could talk a little bit about what the workgroup is, so that you kind of have a sense of it. What I am going to do today, in terms of the goal, is just a quick overview of the charge and kind of a little bit of how the workgroup kind of differs a little bit from the subcommittee, so that you kind of have a sense. Our customer is a little bit different. The nimbleness by which we move is a little different. I just want to make sure that you kind of get a sense of that.
Then we are going to talk about where we are with the guiding principles on data access and use, kind of our current work. I am saying think alignment. You are going to see where we are relative to some of the things that you are doing.
I think what is different is that we may be thinking about it differently, but we want to see where we will end up. Our strategic plan for 2015, and now you all are talking 2016, we will talk just very briefly about that. I have kind of put it on hold because of staffing stuff. Then a little bit about challenges and concerns.
The charge for us in terms of this is we are a workgroup. We should be, remember, assist and advise. I thought this would be good for everybody because we are not just advise, but we also assist. Just in case I have to go through this quickly or something, this is actually online. You can see this. We assist and advise HHS.
Part of what we are trying to do is to think about recommendations to promote and expand access and innovative uses. It is not for us to just say, do what we usually do. The reason we bring the consultants together that we have, which there is data entrepreneurs, there is data developers, I call them all little wonder kids 1:25:53. It is a little different than us sitting around the table. Sometimes you need a playbook to understand some of the terminology. They are great and innovation and solutions. That is why they are there.
Part of what it is to do is to talk about these innovations and applications, to improve health and health care. We have both health status issues, as well as health care.
What are we supposed to do? Well, the workgroup is supposed to monitor and identify, so there are some things we should be paying attention to. Other things we should be trying to come up with is new.
Where there are opportunities to make recommendations to HHS. Now, when we say that, usually the subcommittees are making recommendations to the Secretary. We actually make recommendations to lots of different groups. For example, we will make recommendations, and we get to do it directly, though, we do come back through the full group.
For example, to NHIS. We brought in the NHIS people and asked them, tell us about your releases and how you put your data out. Our job was to, as we listen to them, think nimbly on our feet and give them some advice of ways in which they might do it differently. We are also going to kind of come up with more formal ways to do that. We can do in the room, on our feet thinking, top of the head. Again, the wonder kids do the top of the head, and they are brilliant. That is okay for us to do. Sometimes we don’t have this long process of debating through things. That is good. That is bad. But for the most part, it has been good.
What do we focus on? Content. We could talk about technology, media and audiences. Part of that is because we are supposed to learn about and have some sense of what the HHS data portfolio is. Sometimes we need to get ourselves up to speed.
We also should do this based on kind of thinking about traditional and new dissemination strategies. For example, that is what we were talking about. That is our job. That is what we were talking to our own groups about when privacy came. It is like, well, think about these things. We will give a lot of these. Whether you can do them all will be a different story. That is really what we should be doing is kind of pushing the envelope.
Then also think about the needs for data and information by participants in the health system. We have been trying to do this use case model. We are also supposed to look at the portfolio and think about. We tend to kind of stay focused with the surveys. This was a good reminder in terms of looking at this. We are supposed to do administrative, operational, public health and research data, and really kind of think about what the current policies are and figure out whether we have some commentary we could make for promoting access and innovative use of that kind of data.
We are also supposed to again identify and monitor trends and capabilities for again some of the new information dissemination and data access strategies. We are supposed to think about technologies that could be used. This includes social media.
We have been coming into that space. We are very happy, for example, to have privacy with us. They are trying to make sure as we come in that space, we do it in a very responsible way. It has been great. Leslie has joined us, and Linda is there to represent the subcommittee.
We are also supposed to identify and monitor types of data and information needed. Notice what our charge says. By all participants in the health system, so we are really supposed to think about use case across. We are not just at the level of the data makers or the data users. We are supposed to think about the consumers, the patients, the plans.
Again, you can see big bucket for us. It is like we are stretched a little thin, but we have got a big bucket. Again, I think as we think, we try and make sure. We are starting to pay more attention to when we make recommendations, who does that fit for and who doesn’t it fit for? Who does that leave out? Is there something else we can say to make sure that the people who are being left out can be pulled in, in terms of our recommendations? Just that alone has been helpful to us.
Identify and study areas of opportunity to improve access and application. Again, this is where to think about privacy, technology and data policy issues.
Again, we can be a form for promoting and facilitating creative communication to the public. We need to think about kind of how to do that. Key stakeholders, technology, innovators. Again, what we need them to know is what HHS has and kind of what the opportunities are for its use. Again, when we push things out the door, and we do it in a way in which it reaches new audiences, then I think we are probably doing what we are supposed to do.
We are also supposed to provide access to expert opinion. We do that by the consultants that we have, as well as people that they suggest that we bring in at times. That is to help us in terms of these various areas, policies, infrastructure, et cetera.
Again, back to improving data access and innovative use. Again, to advise HHS in understanding and evaluation of how their data is being applied and the value that it is generating. I don’t think we have done much of that, but I think we are on our way. We are kind of in this phase, too. We are starting to try and live up to our charge.
That is the background. David and Denise, in case you have any questions, please feel free as we go through to ask, or if you have them now. What is the goal? What is the problem? You really started to say who is our audience and what is the problem we are trying to solve?
Right now, what we are working on is developing some guiding principles on data access and use. The reason for that is we are trying to take the datasets that HHS already has and even starting with this kind of philosophy of as you produce your dataset, there are things we want you to think about. You probably can’t do all of them, but even if you do some of them, you will probably increase access of use.
We want to kind of look at those datasets and ask, who probably is benefiting from this? Are you leaving some people behind? As we begin to collect more information about that, we can then help over time to increase that use.
We have tried to have this three-use case scenario. When you say something and recommend something, is it that recommendation is good for somebody who is a designer or a data entrepreneur or the data warehouse? Is that something that fits in terms of people who have a skill set to use data? That is like researchers and others that use the data? Then at that third level, it is the consumers.
We recognize that HHS datasets are not always going to be equally as usable and accessible across the three. I think we need to get to a place of trying to figure out can we improve it across the three. It may not be perfect, but we are at least wanting to see if we can give some advice.
In developing these principals, one of the first things that we talked about is HHS already has some guides. I think it comes from OMB. We are not going to recreate the wheel. That will be out there kind of first. We will append ourselves to that.
The other thing we have been doing is looking at the literature and some of the issues of what happens when you try and have open data. What are some of the challenges? We have been working on this. It is part of what we will talk about today.
Kind of some of that impediment that we are finding solutions to and having suggestions for is availability access, findability, usability, whether or not it is comprehensive, the quality of it, linking and combining datasets, what kind of support should we think about from the data provider, and community building and learning. We are starting to see alignment, I hope.
When the field talks about what the problems are in open data challenges, and we are trying to think about things like how we can solve these. Findability, it may be social tagging may help. We started talking about social tagging, which is we call a dataset the NHIS. It means a lot to us. The public is like, huh uh, alphabet soup again. Maybe the social tagging will be that the public will say, it is the blood pressure dataset.
There are ways in which, and that is what we talk about, is the users may need to start doing some social tagging. In terms of finding it, it is not what we call it, but it is what is designated for people to find it and use it a lot easier.
What are some of the things that we are talking about as solutions? This is kind of what is up here. For example, we have been talking about metadata. We know that metadata typically top-down approach. It comes from whoever produces it.
We are thinking about, as we say to HIS or NHANES or something like that, okay, you have metadata. But part of what people really want is, for example, frequency of collection. A dataset gets put out. Somebody who runs a data warehouse, and they are thinking about developing some apps. They are only going to do it from a business case perspective if they know that this data is going to come out again. They want to know how frequently it is going to come out. They want to know when the updates are.
For them, they are trying to make a business case as to whether to invest in an app on that dataset. They need to know something like that. It is very simple, but a lot of times, this stuff isn’t really evident and available.
When we start talking about that taxonomy over there, alignment, this is the kind of stuff that, to some extent, you want to make sure is there. That is what the public is going to make decisions on. A researcher already knows the releases are always in blah blah blah. We are trying to get this to the entrepreneurs, who are going to get it to the consumer.
The issue of providence, the metadata, we said it should include the data sources. Be clearer. This is administrative data. This is survey data. Who owns the data. We talk about data linkages. Again, we don’t think HHS has to do everything, but there may be ways that the more information that you give, this is what entrepreneurs are interested in doing. They are interested in saying, I can do this for you, and link X to Y to Y.
I know we are getting freaked out because we are thinking of data linkages like the mortality data that is going to go to NHIS. Somebody is going to be like unblended. People are talking about, for example, putting data with data that may come from the grocery store chain. They are thinking differently than we are thinking about it.
The way for us to support them is to be able to put information out there that allows that creativity that you see in the workgroup and outside, to be able to make that data, to do mashups and all the other terms that we will hear. Part of our job, we think, is advising HHS groups about how to do some of this, to make the data be able to usable in ways we haven’t thought about.
Is the data more analysis ready? Josh calls this often smart data. That is where I can take this piece and this piece, and then I can find something, as opposed to what we do as researchers, which is we clean it up, we develop the variable in some way, and then we can do it. Is there a way that you can have stuff more analysis ready?
Glossary, kind of a simple thing, we thought about that. Data lexicon, again, it is like names. There are technical names. We get people to think about whether or not that is the best way to have it is just by this very technical name. Even saying NHIS, can you spell the whole thing out even? Data dictionary, kind of more of a common user explanation. We are all very familiar with the use of data dictionaries, but they are very formal in terms of how we use them.
The entity relationship diagrams, how this relates to other things. Can we put some of that out there? Again, to help people be able to see that connections can be made. We say that the mortality data is out. Tell the person what that dataset can be connected to and how to use it.
Social tagging, now, remember before I talked about things as top down. We are starting now to talk about more bottom up. In social tagging, what we are looking at is the ability for that user to kind of generate some information about this dataset. That may end up being in things like do you have a most frequently-used data point? How do people talk about it? Do you have FAQs? Do you have blogs?
What we are really talking about here is telling the person, you know, you put your dataset out there. But how about having something that says the 10 most frequently asked questions? Begin to get this discussion going in ways in which then the users are more informed and that the users then begin to say what they think this dataset is great for and how they used it.
Community building, again, we are eager to have Bruce with us, so we can even do more of this. In building the community of users across those user cases, it is like can we talk about best practices on how to do this. Can we go back to the concept of the learning health system?
We are still kind of embryonic here. We are working on this, but that is one of the principles is that you have a dataset out there. Can we help you with suggestions of how to build the community of users up, particularly making sure that it is not just the typical researcher or data warehouse person? Is the state using it? Is that consumer using it? Does the provider want to use this? That is something we will also be talking about.
As I said, top down. A lot of this is being done. We want to spend our time a little bit more talking about bottom up. This is where I think some of the ideas that have been talked about, about this community of practice or something like that, comes in. Because we are saying that part of what we need to know in the release of data is what is the problem that the person wants to solve from the perspective of the user.
We realize we don’t really have a good way to do that. How do we hear from that community that we had come in the hearing? Is there a way that, for example, if you put up a blog, you would find out what it is that people want?
Is there a way to build something in that allows you to collect the information? Can it be analytics that are behind the dataset that says, you know, some datasets, before you get to them, they ask you to sign on. You say who you are. Not a lot of detailed information, but you can say I am a researcher, I am a consumer, something as simple as that. You have analytics behind, in which you can tell how long did the person stay on? Where did they kind of fall out?
We want to think a bit more about how to build that and how, for example, to make sure that we know what the issues are. Then what we are talking about is taking those issues and, again, this is Bruce, do we need a data query system? What is it that we can use in order to be able to discover what is relevant for that data user?
Also, one of the things we talked about in terms of getting more information about this, and we usually don’t use it, is social media. Again, privacy comes in and they say, oh, you have to think about when you have enriched datasets what that means. We are good in terms of what that means.
The strategic plan, you have the strategic plan that we produced. It is within the RE agenda book. I am going to put it on hold a little bit. The reason for that is the next slide. We have to kind of figure out what we are going to be able to do. We just lost Lily, in terms of our lead staff person. We are trying to figure it out. A lot of this, I think we will learn today. We will hopefully learn something today from Damon.
I don’t think today is the time for us to go over the strategic plan because I think we have got to kind of figure out kind of what we can produce in which quarter type of thing. We are pulling back a little bit, even though you see we have a great agenda. We have great things we want to do. We just have to be, I think, reasonable in what we can do.
The other thing I want to talk about is our need for renewal of consultants. I think it is really time for us to look at this. We started off, as you know, Justine was the original chair for this. There was time that was used to kind of build this, and I think even with a different agenda.
We have a large number of consultants. But the problem is, for example, we have a consultant how has said he is going to retire. I think he really did because we haven’t seen him. We haven’t seen him. I think some people thought, oh, gosh, that is just talk, but we haven’t seen him back.
We have someone else who just said, call me when you need me, but I can’t be on the calls. We have a big workload. We are really only coming down to about, on a regular basis on our phone call, four of those consultants are actually working. What happens is that we send out, okay, here is the big picture. For each of the consultants, here is what we need you to do. Then they are just not able to make the calls. They are not turning stuff in.
We are between a rock and a hard place. On the calls, the excitement, the passion is really there for moving ahead. I just don’t want to run everybody in the ground. It is like we have Paul, so you have Paul, we have Paul. Paul is working really hard, for example.
I just want to see us kind of rethink about our consultant base and maybe have some handshakes and give pins. I am sure we don’t give gold watches, give pins or whatever it is that we do, certificates and nod our heads, and yes, a lot of gratitude.
Part of the work we are doing, we also need some people that are a little different than individuals who are there before. I love the fact that both Dave and Bill have this informatics background. That is kind of needing to be there. We have some other needs now, in order to get our work done, that I think would help us if we can get the consultants.
We learn with these consultants, they have a choice to just kind of contribute at the table, so we don’t push that away. At the same time, they made a commitment. That is what is happening to really get work done. They didn’t like just being left for, was it, three months. Then they come back again, and they are really curious. That is why having Damon with us has been absolutely great because Damon comes in. He tells us, this is what is going on in terms of HHS. This is what my problems are.
They want to know, when he comes back the next time, did we help? Did we solve anything? Did we make it better? They really want kind of more of a thread that goes through. That is kind of an ongoing activity.
I am going to stop here because I knew we were running late. I was trying to do this in a race. There are a couple of things. I would like to hear from the subcommittees, whether kind of where you see alignment, don’t see alignment, things that you would like to see us trying to get answers to. Then after that, just any questions.
MR. SOONTHORNSIMA: I have a comment and maybe also a question. A couple of slides back, you were talking about guiding principles. I think you started very clear and crisp to me like usability and findablity and those things. Those are kind of high-level guiding principles. Then you got into some of the principles, which seems to me some are more methodology. For example, data warehousing and mass data management as methodologies. Am I making sense? Then you got into metadata, and then you get down to that granular level.
The question I have is twofold. One, some of those really are tactical methodology things. They are pretty standard stuff in my mind. The question is, are these being used or the recommendation, are they being provided to HHS because you have got a ton of data out there, so the department can begin to harmonize using some of these common methodologies. I assume those are actionable items. That is the first question.
Maybe really more from the workgroup perspective, the second question is if you took out those tactical things, what are some strategic things then that we could focus on in a workgroup that clearly can align A, with the committee, B with the overall HHS strategy.
DR. MAYS: For the first question, it would appear that, of course, all of those things were done by everyone, but it is who benefits and the level at which it is done. In our job, part of what we are trying to figure out is, for example, in terms of something like social tagging, can you get it so that it is beyond just the typical researchers or the typical data warehouse individuals?
What we are doing is coming up with suggestions to try and expand kind of how to do this. That is why we started saying it is not just the item itself, but it is also the use case. Who does this apply to? There are basic things that they do. But we are trying to bring in innovation and other avenues that typically aren’t used. It may be that, in terms of the metadata, something as simple as is there a glossary? Does the glossary use terms that would expand down, as opposed to up?
I think part of what we are saying, is that, yes, those are things that are kind of technical management. But how to expand them in ways to have a different audience is what we think we can opine on that might be —
MR. SOONTHORNSIMA: — what the methodologies are being used within the department.
DR. MAYS: Yes. It depends on, I think, which dataset it is and who has what resources. We think that if we put it out there, you can just use what applies to you. If the 10 things that we say are ready, then you should get a gold star. If not, then here are some other ways for you to think about it.
I think for the second one, I think Damon is probably better.
MR. SOONTHORNSIMA: The second question is really centered around, based on the guiding principles that you have in the early slide, are there things that are less tactical that are more strategically aligned with A, HHS strategy, B, the work of the committee. The other subcommittees that are trying to bring all different body of work together.
MR. DAVIS: Could I have the benefit of that slide, just to sort of recap? I apologize to the group for coming in late. I would love to see that slide, if we cold.
While we are getting back to that slide, I will just say a couple of quick things. One of which, I think, Vickie has already mentioned. It goes back to, I think, the usability bullet. One of the things that I think is going to be really important for the department, but therefore challenging as a department, is the improvement of the metadata that we have.
I think it would be incredibly beneficial for those who are using the data, both our internal departmental customers, as well as the innovators and entrepreneurs, researchers and the press, who are trying to locate our data to have the improvement of our metadata that is going to allow them to know before even downloading, before even accessing, what it is that might be valuable out of the datasets.
I say that is going to be challenging as a department because you could pick any two datasets that are somewhat related and find different ways that we are describing the same kinds of measures and fields and things. Therefore, that sort of strategic alignment, I think, is going to be incredibly challenging, but would provide enormous value.
We have got a lot of mechanisms for our data collection. It is challenging to therefore change some of those mechanisms for that data collection in order to create that harmonized output at the backend. I think that guidance as to where it could be strategically valuable to create some of that metadata alignment is going to be very helpful.
MR. SOONTHORNSIMA: Is that then the single most important thing that could come out of this workgroup?
MR. DAVIS: I don’t know if it is necessarily the singular most important thing. I could use a lot of different things. As I traverse the department, talking to folks about the datasets that they are collecting and curating, there are folks that feel pretty good about the products that they have produced. I have heard conversations with the workgroup about the development of the scorecard that would help folks to understand. You are producing high-quality PDFs. You might feel great about the data that is being produced out of that report. But in fact, it is not nearly at the level quote unquote liquidity that we would all desire.
One of the challenges that I think we have there then is the resource allocation that Vickie also alluded to. The fact that we need to very much prioritize where it is that we would, in fact, see a real strategic benefit from focusing our limited resources on changing that PDF into a strong and well-supported API.
We need to spend a little bit of time sort of strategizing where it is that limited resources can be applied most effectively in order to make one set of data that much stronger and hopefully in support of many other sort of separate and related datasets, data resources, query tools, what have you, that could then also be uplifted by the improved quality of that one.
I am hard-pressed to point to one single thing to say that the group should do because there is such a universe.
MR. SOONTHORNSIMA: It would be nice to prioritize based on maybe one, two or three things that can actually be the tide that lifts several boats.
DR. MAYS: I think a little different in terms of the approach that we take, which is we have lots of customers. It is like for NHIS, if they come in or we bring them in to tell us, we may give them lots of advice. What they do and how they do it is a little different.
I think for us, it is the development of this information. In his office is the issue of which of those, would they like to come back and maybe ask us to help them work a little bit more. But for us, it is like, these are some principles to think about across the groups for their datasets. Our customers will come back and ask us for more. Then we look at our bandwidth to figure what it is that we can really do more of. Right now, we are kind of on hold to figure that out.
MR. SCANLON: I would say that. The workgroup has started on much of this. We have already had some very good recommendations on tagging some of our datasets that we have begun to employ. The workgroup also acts as a reactor panel. Like when Charlie was here, you were giving him ideas about what would be a good idea. That is kind of the way the panel works, as well because you have got the folks here.
This all comes out of our initiative to make our data more accessible, findable and usable to support health and health care in the US. It is part of our open data initiative. It is more than that. It is part of big data initiative.
Again, the research community, as Vickie said, they are aficionados. They know where the HIS is. They know how to get it. They know NHANES. They know how to use CMS data. This is not that we don’t have a need for improvement there, as well. This is more to provide the perspective of the developer and the technology community, the innovator community, to merely be a multiple for our data.
In other words, we put things on Data.gov. We continue to do our agency by agency public-use files and reports. That is fine. But again, you have to be an aficionado to know how to do that.
If you really want to have a multiplier effect of democratizing the data, again, we will still have to protect privacy and all of that. That is why we don’t have the HIS up on the web, for example. They are different audiences. They are different datasets that we make differently.
Again, this is not the research community that we are aiming at. This is how do we use the innovative community, the developer community, to look at the way we make this data available, what we have, and to support the development applications and really to have others help us through their own work to get the data around. That is, I think, particularly where the workgroup comes in.
These are folks who live in that community. They make a living in that community. That is what we are really looking for is their perspective. They have looked at our datasets. We have had some of our individual bigger survey folks come in.
In that instance, I think we are asking them to act as expert reactors from that point of view. CMS, I think we would like to get, as well and see what more we could do. But again, the overall goal is to promote the availability, the findablity, accessibility and so on of our data, to the broadest community to improve health and health care. We are trying to get beyond the research community, where we focused a lot of efforts in the past.
DR. COHEN, B: A couple of comments. I slightly disagree with you, Jim, that the target is not the research community. I think the research community knows where to get the data, but doesn’t truly understand what they are using.
I think this was essentially the discussion we had around the framework. If we could develop consistent, whether we call it metadata, principles, characteristics of the data, that everybody can go to and use, that describe the data that we provide, it will help the researchers, ultimately it will help the users who support community efforts. Certainly, it is a requirement for developers.
I think the metadata tagging and these principles that you laid out are very consistent with the framework activities that we are trying to describe. Perhaps we didn’t do a really good job describing the framework. But I see these as really convergent efforts, targeting the idea, just figuring out what the hell is in these data and being able to describe it in one consistent fashion, so everybody gets it. That is the bottom line.
Open data aren’t valuable unless people really know what is there and what its characteristics are and what its limitations are, and all these other kinds of things. You talked about metadata and what Bill tried to present in his taxonomy. I think Vickie, when you asked does this converge with the other subgroups, I think we have to figure out how to break the log jam in words, so that this truly converges with the framework activity. I think they are really very much related. That is my overall comment.
My other comment is when I look at the workgroup activity, I see two large buckets. One that is responding to specific questions for folks like Damon and the data holders and curators for us to work on and analyze. Another is the stepping back function that we are describing here, how to promote open data by standardizing the understanding of the data we release. Those are, I would say, the two areas that the data workgroup is trying to pursue simultaneously.
It is a bandwidth problem, but I think you have laid it out in the slides very succinctly.
DR. MAYS: I would agree. I think what we have done before is like the response stuff. What we are trying now to do is to build the consistent and steady. To do that, the consistent and steady is the interleafing that we want to do in terms of the subcommittee. This is the newer bucket that we are in the middle of. We haven’t before produced things, but that is kind of what we are working on now.
MS. GOSS: This conversation and your setup, Bruce and Ob, really trips into the conversation we were having yesterday about needing an overarching roadmap and how were we all going to get coordinated. There has been a lot of work that has gone on, busses, trains, dinners and side bars, related to this topic.
Debbie asked me to make sure I got it on the record some of the work that we are going to do and some of the coordination that I have accomplished in the last 24 hours. Terry and I are going to build a baseline document, based upon some thinking from yesterday and this morning, which will help us understanding our categories, our goals, our measurable objectives that track into our work plan tactics.
What we will then do is create a straw man that we will run by an ad hoc group, which will consist of Vickie, Bill, Walter and Linda, probably with Jim’s support, as well, to at least get some initial feedback, so that we can then have something of more substance for the full committee to respond to, and that the framework for this roadmap will be to leverage some of the approaches we have seen in the federal health IT strategic plan layout.
I am not talking about graphic representation, with content that is harnessed from our NCVHS charter, coupled with a lot of the thinking that we are already doing inside of our committee work. What it will hopefully give us is a picture about how to drive that convergence across multiple workgroups and committees, and help us manage our workload more effectively to deliver success in the subsequent years.
MR. SCANLON: Denise, I think you are going to get the last word again.
MS. LOVE: I can say that I am officially overwhelmed. Because I am the new kid, I will ask the dumb questions because do I get two meetings of dumb questions, and then I can’t ask them? It sort of reminds me of back to the future issues we did with the data web. Is that still out there, the DataWeb? It had a taxonomy and a cataloging of federal datasets. Some years ago, I did some development.
It has probably converted to something else. It was Ferrett, then DataWeb, and then something else. Also, I don’t know how many years, it has been a while but some of this work went on.
Another initiative that we have been involved with, and I am just wondering if you can piggy back on, is USHIK, the United States Health Information Knowledgebase. There are some data elements. Is there anything to piggyback on, or are these two very separate things?
MR. SCANLON: We can take a look at it. The health information knowledgebase, (USHIK) it is an activity that I think AHRQ is now managing, I think. The attempt there is just to record and define, while a lot of these terms I don’t know.
MS. LOVE: They harmonized on the all-payer side – they harmonized across the United States. There is some metadata out there and some common feedback.
PARTICIPANT: Is this something that they are tasked for doing or is this just a project you are doing?
MR. SCANLON: They have been doing it for quite a while. Let me send you something around about it, about what it is. Honestly, there is no shortage of attempts to do data dictionaries and metadata. You really have to decide if that is where, what is going to be the value in the long run. There are a number of them around.
The focus here is to get the data, which we have, which we made available already. We will continue to make it available, to use this community as a multiplier, to the extent that we can, to give us ideas about how we can do that better. There is some overall framework, but it basically just to improve health and health care.
We can define and categorize, but the older view is basically, I think Todd used to say, data Google versus data Vatican. It is getting data democratized and out there for everybody. That is the overall goal. How we get there, the tools, we can talk about.
MS. LOVE: One thing that helped in my world, which may or may not translate because I am learning, is use cases. Really getting down into those use cases may be one of the critical —
DR. MAYS: I think that is where we end up being out of our usual way of working because it is usually like we are making sure that the researchers it is working for. We have to realize that for a data warehouse, it is about a business case that needs to be made. If it is at the consumer, again, it is different.
I agree. That is why we kind of try. As people talk, which of those groups are you speaking on behalf of that the solution would work?
MR. SCANLON: We are quite happy with incremental improvements. That is really where we are now.
MS. LOVE: That helps. I was thinking way back when we did work on the DataWeb, which was a huge undertaking.
MR. SCANLON: No. We can make improvements of an incremental nature fairly readily. I think Damon gets the last word.
MR. DAVIS: I don’t want to scoop my colleague, David, who is going to come and present later. I don’t know if he will be here in the afternoon. One of the things that he is going to talk about is the collection of use cases to help us from what he has termed, demand-driven open data, so that we can really start to have a more strategic approach to prioritizing what it is that we are doing in the open data space.
One of the things that our current CTO, Bryan Sivak, has said is that it would really be valuable for us to build community around the data, so that we can keep a finger on the pulse of what people are actually interested in utilizing.
If we were to stop liberating data right now, and just start attracting and magnetizing people to the data, we could probably add a tremendous amount of value by having just that many more eyes and users and use cases be plugged into it, versus just continuing to put out data that potentially very few people actually want.
That is going to be sort of the shift, I think, that you are going to begin to see out of HHS is the desire to focus limited resources on more strategic approaches to the data you actually want in the ways that you want it. Again, demand-driven open data is going to be something that we will cover this afternoon.
MR. SCANLON: Other questions? Okay, so we are not that far behind the time. So we finished the regular Full Committee meeting, except for the public comment period. Let me ask if there are any questions from the audience in the room or comments or on the phone.
MS. JACKSON: I have a follow-up for the Full Committee, just so you will be aware of timing. The Executive Subcommittee will have a conference call probably in April. What I heard in the May meeting, we have got an action expected from the standards based on what they will be covering tomorrow for the administrative simplification and other items, as well as supporting community engagement from populations. We are working on recommendations on that. Thank you so much for your help in helping to prioritize that. I will be checking in with Jim on Affordable Care update.
There was a discussion about bundling, care, coordination, information along that line, as well as time on the Full Committee for a framework discussion. There has been so much that came up, as Vickie and others were saying from yesterday and today, and looking at the framework and autonomy and so much that we need to have on the full.
Also on the back burner is still this demo that we are working with Vital Statistics with Michelle Williamson. We have been trying to do this for about a year and a half. It takes so much person power effort, but it is definitely on our back burner.
I am really excited about the June 16th and 17th, I think. It will be for the ACA review committee. We have gotten some nice confirmation from the Full Committee membership for that.
DR. STEAD: We also will want a block to discuss the planning straw person that Alix was just describing.
DR. SUAREZ: To that point actually, I was going to suggest we need an Executive Committee to meet probably almost monthly, probably in March and in April, at least, to prepare for the May meeting. In May, we are going to have what Alix was suggesting.
We also were thinking about reserving at least the full half-day of the first day, the second half of the first day, entirely for the entire committee working session to discuss sort of like the larger plan. I am just thinking we need some prep time of the Executive Committee to do that.
MR. SCANLON: If we all agree, I will try to get a briefing to the Full Committee at the May meeting on this whole health care delivery system reform. What is the direction it is heading, and what the measures are, what the major parts of it are and what the metrics might look like in terms of data. We will do that, as well.
Teri, do you want to update us. We actually had some good news about CMS. A meeting about ICD-10 testing.
MS. DEUTSCH: Yesterday, you were informed that there was testing going on, and we didn’t know if the results were going to be coming in sooner or during the NCVHS meeting. Apparently, it has. It has been posted. The testing that CMS has done has been very successful.
I do have the website links, if anyone is interested, so that you can look at the actual results. It describes the number of providers and the industry that has submitted claims for testing, what the submissions were and the results of them. It was very successful. Medicare and CMS is ready for ICD-10, if there was any question.
MR. DAVIS: I am not familiar with what it takes to test Medicare for ICD-10. I don’t know if everybody else already knows what goes into that test.
MS. DEUTSCH: There are many things that can be tested. There was specific criteria that was going to be looked at to see if the ICD-10 transmissions were able to be successfully transmitted and the results. That is described in these links that I have for you. I don’t know. I have to go and read it, exactly which components were tested. The email will give you the links.
MR. SCANLON: We have a project going on. Denise and I were talking yesterday with our NCHS, the physician ambulatory care survey, where we are looking at the transition from ICD9 to 10, and what that means for statistics and indicators. I don’t know if we are that far along, Denise, but maybe at a future meeting we can have a update. It is a major change, and you wonder does it change the rates?
MS. LOVE: It will change the rates, but more than that, they have been involved with restructuring their dropdown menu. They limit what they get from the abstraction. It really has also created some efforts to review that dropdown list of 5000 codes.
DR. COHEN, B: I don’t want to prolong this anymore, but I think the issue of comparability ratios, the way NCHS did for mortality, needs to be on the table for guidance. Maybe in the fall — for getting the data in for generating statistics, not for claims and payment, but for statistical analyses. I think this would be a longer conversation that we should engage in.
MR. SCANLON: No, I think the question will come up right away. Otherwise, it looks like causes of hospitalization of changing.
MS. LOVE: I am happy to offer my team to update with Pat Romano. This has really been, at least the project that I have been involved with, quite detailed.
DR. SUAREZ: We did postpone actually a hearing that we are going to have with ICD-10 to the fall. That might be a time to think about doing it, both the perspective of the transition to ICD-10 and the codes of maintenance, as well as the public health window.
MS. LOVE: There is a structural issue. If you have a dropdown menu of 5000 codes, now you have to decide what trimester you want, not just pregnancy. You have to decide laterality and all of that.
DR. SUAREZ: So if people go to CMS.gov/ICD-10, they will find a countdown to ICD-10. We are 217 days, 10 hours, 41 minutes and 49 seconds away.
DR. MAYS: I just want to make sure in terms of the workgroup, in terms of the May meeting. Depending upon what our progress is, we may have something to bring in May. I just want to make sure time wise. Then, we can tell you ahead of time.
MR. SCANLON: We wanted to have some time for discussion and reflection and some issues to tee up, to help us guide the planning down the road. We will be at the Humphrey Building for the May meeting.
DR. SUAREZ: There are a lot of things coming up in May. I wonder if we should not at least consider to extend the meeting to a half a day on the third day. Right now, we have Wednesday and Thursday full day. We could consider having Friday half a day, Friday morning. Then people can then depart on Friday, if they haven’t made reservations. Just an idea.
DR. COHEN, B: We are running a little late. Are we going to start the data workgroup at 1:00?
MR. SCANLON: I believe we are officially adjourned.
(Whereupon, the meeting was adjourned at 12:20 pm)