[This Transcript is Unedited]
Double Tree Hotel
300 Army Navy Drive
Arlington, VA 22202
Agenda Item: Call to Order, Review Agenda
DR. CARR: Day two of the National Committee On Vital and Health Statistics is now open. And, I want to welcome everyone back. And I want to, just before we get started, make sure that we take a moment to acknowledge the superb support of our staff from NCHS including Debbie Jackson, Marietta Squire, Katherine Jones, and Jimmy Christiansen. Are they here?
Typically, they're out working here. But, they really are behind the scenes. They are incredibly patient and just a very can-do mentality. It's not easy to get it all organized here. So, thank them.
(Introductions)
DR. CARR: So with that, we'll go around the room. I'm Justine Carr, Steward Health Care, Chair of the Committee, no conflicts.
MS. GREENBERG: Marjorie Greenberg, National Center for Health Statistics, CDC Executive Secretary to the Committee.
DR. SUAREZ: Walter Suarez with Kaiser Permanente, member of the Committee, and Co-Chair of the Standards Subcommittee, and no conflict.
DR. WARREN: Judy Warren, University of Kansas School of Nursing, member of the Committee, Co-Chair of Standards Committee, no conflicts.
MS. MILAM: Sallie Milam, member of the Committee, no conflicts.
DR. SCANLON: Bill Scanlon, National Health Policy Forum, member of the Committee, no conflicts.
DR. MAYS: Vickie Mays, incoming member, no conflicts.
DR. COHEN: Bruce Cohen, new member, no conflicts.
DR. CARR: Also, I think customarily also where we're from, our affiliation. And, where did we leave off? Let's go back to Raj.
DR. CHANDERRAJ: Raj Chanderraj from Las Vegas, member of the Full Committee, no conflicts.
DR. CARR: Okay, whatever your affiliation is in Las Vegas.
DR. CHANDERRAJ: I'm a practicing cardiologist.
DR. CARR: Okay, good. This is clarifying.
DR. MAYS: Vickie Mays, University of California, Los Angeles.
DR. COHEN: Bruce Cohen, Massachusetts Department of Public Health.
MS. KLOSS: Linda Kloss, Health Information Management Consultant from Chicago. I guess day two and not any longer a new member.
MR. QUINN: Matt Quinn, NIST.
DR. CARR: Liaison.
DR. NICHOLS: Len Nichols, George Mason University, also day two no longer new. (Laughter)
DR. FITZMAURICE: Michael Fitzmaurice, PhD for Healthcare Research and Quality, liaison to the National Committee.
MR. BURKE: Jack Burke, Harvard Pilgrim Health Care, former new member, no conflicts.
DR. MIDDLETON: Blackford Middleton, Partners Healthcare, Harvard Medical School, member of the Full Committee and Co-Chair of the Quality Subcommittee, no conflicts.
DR. HORNBROOK: Mark Hornbrook, Kaiser Permanente, member of the Committee, no conflicts.
DR. FRANCIS: Leslie Francis, Law and Philosophy, University of Utah, member of the Full Committee, and Co-Chair of Privacy, Confidentiality, and Security. And, I've been around for a while, and I don't have any conflicts.
DR. GREEN: Larry Green, University of Colorado which is in Colorado, and member of the Committee, and no conflicts.
DR. CARR: Can you get everyone around the table?
MS. JONES: Katherine Jones, NCHS, staff to the Committee.
MS. LOWE: Gwen Lowe, CAQH.
MR. BARTOW: Nate Bartow, Price Waterhouse Coopers.
MS. JACKSON: Debbie Jackson, NCVHS, Committee staff.
MS. WILLIAMSON: Michelle Williamson, staff to the Standards Subcommittee.
MR. ALBRIGHT: Matthew Albright, Center for Medicare and Medicaid, Office of E-Health Standards and Services.
MS. ROURKE: Cynthia Rourke, America's Health Insurance Plans.
MS. WEIKER: Margaret Weiker with HP, also representing ASC X12, NCPDP, and the DSMO.
MS. KHAN: Hetty Khan, CDC's National Center for Health Statistics, staff to the Subcommittee on Privacy, Confidentiality, and Security.
MR. ALM: Frank Alm, American Dental Association.
MR. CHAPPELL: I'm Dave Chappell, ASPE.
MR. BIZARRO: Tom Bizzaro, First DataBank.
MS. ESQUALA: Mariette Esquala, staff to the Committee.
MS. GREENBERG: Well if I could just say, and you know, you know how I love this Committee so and its history, but I'm also concerned about its future too so that it should always, we should have new people all the time, new members, and new people staffing and understanding it. So, we do have two interns today whom I'm mentoring this summer from my alma mater. And, I wanted to introduce them. They're both there now. So, I just want you to know, I am thinking about your future.
DR. CARR: I also want to acknowledge the staff from the various agencies that do yeoman's work for the Committee. It makes it a tremendously effective group effort. And, one more thing that should not go unacknowledged, in Boston we have a hockey team. I'd just like to point out winning three game sevens is not nothing.
MS. GREENBERG: Well, we were celebrating that potentially last night at the dinner, and I wanted to specifically thank Jeanine for organizing our dinner last night which was very successful. And, I think Mark and Don both, enjoyed it very much. And, we welcomed them into the alumni association of the Committee and then also welcomed the new members.
DR. CARR: Members on the phone or participants on the phone, could you identify yourselves please? There are none? Thanks. Well, if you hear them come on, just let me know.
Agenda Item: PCAST Letter -- ACTION
DR. CARR: All right. If we will move straight into the PCAST letter, and if you could bring it up there, that would be great. You have copies at your desks.
We took all of the comments from yesterday. I'm going to give you just an overview. I think the comments were generally to ensure that the discussion met, you know, that the recommendations and the discussions and observations lined up. And we, let's see, we you know, some of the language, we broke down a couple, in a couple of cases, we broke recommendations into two parts.
We also, there was concern about what was yesterday's section three that it was disproportionately long and complex. We have done a couple of things. First, because it speaks about metadata tagging, we moved it up to be the second topic.
And then, we collapsed the three recommendations down to two since two of them were quite similar. And, we've tried to organize observation and recommendation so that the right discussion is under the right heading. It's been an exercise, I want to thank Matt Quinn and Linda Clos for their contributions to this because I think it was very helpful and especially I think Linda, you know, getting an unbiased perspective on just reading for content was very helpful.
So with that, I guess I think in this situation, we probably do need to just read a whole lot of, okay. So, the topic then is NCVHS recommendations to achieve the goals of the PCAST Report on health information technology. Are we good with that as the headline? Okay.
(Reading) Dear Madam Secretary, the NCVHS is the statutory public advisory body on health, statistics, national health information policy to the Department of HHS. Our standard text.
Next paragraph. The December 2010 "Report to the President, Realizing the Full Potential of Health Information Technology to Improve Healthcare For Americans: The Path Forward" by the President's Council of Advisors on Science and Technology ("PCAST Report") presents a bold vision for interoperable health information system for our nation. NCVHS reviewed the report and applauds PCAST support for concepts that have been priorities for the Committee.
Some of the areas, let's see. Some of the areas include, are we missing, oh yes, include, right, emphasis on improving clinical care efficiency and decision making through an effective health IT infrastructure. Second bullet: enabling universal data exchange, thereby fostering "network effects" that increase the utility of information both for individual patients and in aggregate.
Third bullet: structuring the information in Electronic Health Records for real-time re-use, e.g., public health, comparative effectiveness research, treatment quality, adverse event detection. Fourth bullet: enhancing possibilities for using information in ways that protect against unintended linkages to individuals.
Let's see, is there more? Fifth bullet: strengthening privacy and security protections as information is transferred, aggregated and used. Next bullet: providing information in a manner that enables patients to become more involved in their care. And last bullet: enhancing patient identification methodologies.
MS. GREENBERG: Can I recommend that the first bullet be changed to a gerund instead of a noun just for parallelism on here?
DR. CARR: Oh, yes. Yes.
DR. FITZMAURICE: How about emphasizing improved clinical care efficiency?
DR. WARREN: Either improved or improvements. I wasn't sure which one should be used.
DR. FITZMAURICE: And, if I could suggest as just a, some of the areas, change that to some of the concepts because it refers to the concept you talked about in the previous sentence.
DR. CARR: Okay. Some of the concepts. Where is that, Mike?
DR. FITZMAURICE: Third paragraph.
DR. CARR: Okay. Some of the concepts.
PARTICIPANT: Shouldn't the first one be "emphasizing"?
DR. CARR: Yes, that's what he just said.
DR. FRANCIS: Why not just "improvement"?
DR. CARR: Yes, yes. That's right. That's what it was. Improving clinical care efficiency, right?
DR. WARREN: Well, I don't know. I like "emphasizing" because that's stronger than "including".
DR. FITZMAURICE: "Improving".
DR. WARREN: Oh. "Improving". Sorry. Okay. "Improving". All right.
MS. GREENBERG: Now, I guess you'd say "some of these areas include".
DR. CARR: Okay. So, we're saying some of these "concepts".
MS. GREENBERG: Oh, these "concepts". I mean, the two sentences didn't go together.
DR. WARREN: So, I would make a suggestion on that. These really aren't concepts; these are ideas because a lot of the bullets contain more than one concept. So just, I think it reads better.
MS. GREENBERG: As a terminologist speaking, yes. You're right.
DR. FITZMAURICE: So change "concepts" to "ideas", and change "areas" to "ideas"?
MS. GREENBERG: Yes. Ideas.
DR. HORNBROOK: I am having trouble understanding the fourth bullet now. Enhancing possibilities for using information in ways that protect against -- It sounds like the goal is to use information that protects against.
DR. CARR: I think the goal is to protect privacy. The way it was previously written, we said enhancing possibilities for using information in ways that cannot be linked. But, we specifically talk about, the whole point of this is that you can link. But, we want to link and protect unintended consequences.
DR. HORNBROOK: Is that bullet necessary given the next bullet?
DR. CARR: Yes. Enhancing possibilities we can leave. Yes. All right. So, let's just take it out?
MS. MILAM: Add link to the next bullet. We could add transferred, aggregated, linked an used because linking is different.
DR. CARR: So, is the recommendation to blend the bullets or delete "enhancing possibilities"?
MS. MILAM: I would delete "enhancing" and add linked to the third from the end.
DR. CARR: So we are deleting "enhancing" and strengthening privacy and security protections as information is transferred, aggregated, linked, and used.
PARTICIPANT: Yes, I think that says it perfect.
DR. WARREN: So, I just have a question. As I think about linking records and what we're asking for are for people to look at new methodologies to provide these links. If you just put links up in the other one, then it really gets tied into privacy and security.
And, the linking is more than that because I do think we need more research on how to link records for whatever the purpose is. Am I not understanding?
DR. SUAREZ: I think the bullet where editing is, against unintended linkages, I think we're talking about intended linkages which is what I think you were referring to. You know, we weren't saying two different things.
We're saying we need to find better ways to improve linkages whenever they are well intended. But, we're also enhancing the possibility of using information in ways that protect unintended linkages.
DR. WARREN: So, if that bullet then means that we're also enhancing possibilities for use, are we also enhancing new methodologies for linking the records?
DR. CARR: Okay, and so we did have a hearing on linkages, right? And, the inclusion of that was enhancing linkages. So in the interim, could I ask you to just think about some words, and we'll come back to that Judy? Try to state that?
DR. WARREN: Well, I guess I was happy with it being a separate bullet. So that I could try to edit that other bullet --
DR. CARR: So that the concept, we don't want to lose the concept that enhancing possibilities –-
PARTICIPANT: I think it's redundant. I mean, if you add linkage to it --
DR. CARR: Well, yes. I think this bullet isn't saying what you were saying. You're saying that this is promoting the fact that linking data enhances learning. So, there's not a bullet that says that.
DR. SUAREZ: I don't want to mix up, but the last bullet says enhancing patient identification methodologies. Maybe we can weave in the concept.
DR. CARR: That's an excellent point. I would say enabling relevant linkage while still protecting against unintended consequences or something like that.
DR. WARREN: So, when we say "enabling linkage", are we implying then that people will look for ways to do that? Okay, so there are different possibilities right now. One possibility, which is we can't do, is a patient identifier.
Another possibility is using an algorithm. And, we've heard testimony on several algorithms that only need about five items in order to have a 95 percent accuracy rate of linkage. If we add in one more item, we can get it up to 97, 98 percent.
Well, what about encouraging research on methodology to get that linkage up higher because while that linkage may be good for analyzing data on a secondary perspective, it's not good enough for clinical decisions.
DR. CARR: So, what we are talking about here are the areas of alignment of what we're seeing in the report and previous work of NCVHS. So, I think that what we might want to do here would be, I like the recommendation of enhancing patient identification methodologies, you know, enabling linkage. And then, if we want to discuss how we do it, we can do that in the recommendations.
MS. GREENBERG: But, there are linkages that don't involve patient identified information, certainly the support of the letter that came from the Population Subcommittee. I'm looking for Don. I've picked him already, but that includes just linkages --
DR. CARR: In the interest, I want someone to go online and look at that linkage letter and come back with what did we say. We'll start with what we said at that letter and see how it fits with this. And then, I'd like to move forward.
MS. GREENBERG: I don't have, but maybe you can do that?
DR. CARR: Yes.
DR. HORNBROOK: There is at least two kinds of linkages here. One is the business linkage. I mean, when you plan on your bank account, it is a positive definite linkage. There could be mistakes very rarely.
And so, if you're trying to link my pharmacies together back to my medical record, that ought to be done in a hundred percent certainty environment, not to be probabilistic, versus those of us who are doing research and we've got to link, we have no way of knowing to confirm the validity of that linkage. So, which linkage are we talking about here?
DR. CARR: I mean, I'd wait for Judy or Walter to come back with what we said in the letter. We have two different letters. One is saying the importance of enhancing patient identification methodology. The second is about linkage.
And so, here we're simply presenting the concepts that align with NCVHS. So, they'll come back with that. When we get to that section, let's talk about how prescriptive we want to be in the recommendation.
DR. SCANLON: I could say that, I mean the population hearing on linkages was all about research. It wasn't about health conditions.
MS. GREENBERG: And, does the PCAST Report address the research? I mean, I've read it in the past. So, that is relevant. But, it's different than the patient matching issues.
DR. SCANLON: And in that context, there's this issue that there are identifiers on a lot of records. I mean you know, for Medicare beneficiary, there is a Medicare beneficiary ID that can be used for linkages. The issue of when you start to combine records that may not have, you know, that I'm paying for this out of pocket and this wasn't covered by this insurance or something like that, that's where the problem comes in terms of lack of a single identifier.
DR. CARR: Are you making a recommendation?
DR. SCANLON: No. (Laughter)
DR. SUAREZ: Let me suggest because I think that the last bullet, you really should emphasize the main point which is enhancing patient identification methodologies to improve patients' care and enable appropriate linkages of health information. I mean, I think it goes to the point of the reason for improving or by enhancing patient identification methodologies, we want to improve patient care whether it is directed patient care or whether it's to research or to other population management mechanisms.
DR. CARR: Okay. So, what I heard you say is enhancing patient identification methodologies to improve patient care and enable appropriate linkage.
DR. SUAREZ: Yes. Exactly.
MS. GREENBERG: But, that doesn't actually reference back. I mean, I think that is in the functional or a different letter. But, that doesn't reference back to the linkages letter.
DR. CARR: So, Judy is looking that up, and we'll come back to that in a bit. But I think, so let's move to the next paragraph, the PCAST, again in some of these, this does not show, oh no, yes it does. In some cases, we simply rearranged paragraphs. And, I think you'll see that as we go along.
But, on the next paragraph, the PCAST --
MS. GREENBERG: Can I ask a question?
DR. CARR: Yes.
MS. GREENBERG: I thought we were going to possibly start this letter, or did you decide to end it that way with what you took us through yesterday?
DR. CARR: Yes. Those main bullet points. So, Matt and I got a little out of sync here. And, Matt incorporated it in a more narrative way. I think we've covered all the points, but it wasn't the exact language. But so, I'm honored that you want, we could go back to that.
Let's read this through, see what's missing, we can go back and do it. (Reading) The PCAST Report intentionally and importantly calls for transformative (not incremental) change in our nation's approach to building an interoperable, usable health information infrastructure. The Committee, however, recognizes the significant gap between the infrastructure envisioned in the PCAST Report and the current state of health IT.
So that, in fact, was one of our points yesterday. One is we applaud the bold vision. Two is there's a line-up of things. Three is that there are gaps.
Moving on to the next paragraph. The Committee observes that sweeping change carries a notable level of execution risk. And such change, especially in the area of data exchange standards may have unintended consequences, such as, impeding current provider adoption of electronic health records, slowing progress toward care models reliant on this infrastructure, and sacrificing progress to date on standards harmonization.
The Committee recommends an approach to this transition that is accelerated but also deliberate and judicious. The approach should continue to build upon work and progress to-date while initiating necessary foundational research to support the new architecture envisioned in the PCAST Report. In other words, in the words of one Committee member, "before we commit to exchanging everything, we should focus on exchanging something on a national basis."
PARTICIPANT: Could we change that to nationwide?
DR. CARR: Yes. Are you good with that? Okay. Nationwide.
The vision of the PCAST Report can be achieved. Achieving the potential of health IT's role in improving the nation's health and health care depends on achieving the vision in this letter. The Committee recommends specific actions to address critical steps needed to bring this vision into reality while maintaining current progress.
Is that properly stated, Mike?
DR. FITZMAURICE: In reading these two paragraphs, I go up to the PCAST Report intentionally and importantly calls for. I would suggest deleting intentionally and importantly because we're like putting them up on a pedestal so we can knock them down.
DR. CARR: Right, yes. Okay. (Laughter)
PARTICIPANT: Where are we?
DR. CARR: So that, after the bullets, the opening sentence, the PCAST Report calls for transformative. So, we take out intentionally and importantly. Just "calls for transformative (not incremental) change in our nation's approach".
DR. MAYS: Justine?
DR. CARR: Yes?
DR. MAYS: The next one, the Committee observes. It's a really long sentence.
DR. CARR: It is, yes.
DR. MAYS: So, the Committee observes that sweeping change carries a notable level of execution risk. Period. These changes especially, I mean, that would break it up at least a little bit.
DR. FITZMAURICE: Or these risks include.
DR. MAYS: Oh, okay.
DR. FITZMAURICE: Which is kind of what you were suggesting anyway.
DR. MAYS: Yes.
DR. FRANCIS: I would also, the last sentence in the preceding paragraph, I would take out "the Committee, however", and I would just say "there are, however, significant gaps".
DR. CARR: Okay. Is that, are we good with that? Are you getting this, Matt? However, significant gaps between the infrastructure and vision. Okay. That's great.
PARTICIPANT: Do we leave the "however"?
DR. CARR: Yes. Good. So, we've covered the gaps. Okay so, let's see. We read the vision. Okay, now we're on to observation number one.
DR. KLOSS: I think we maybe need, excuse me, I think we need a little rework on that paragraph before the observations.
DR. CARR: Okay. Do you want to look at that, Linda? Or, do you have a recommendation?
DR. KLOSS: Sure. I just think we could just frame the observation and recommendation format that's to follow but cut the more gratuitous comments about the vision can be achieved and the achieving.
DR. FRANCIS: Do we really think the vision can be achieved?
PARTICIPANT: Well, that's what I wondered myself.
DR. COHEN: Yes.
DR. FRANCIS: Okay, so we need a more straightforward introduction into what follows.
DR. CARR: So, the recommendation is delete the sentence of "the vision can achieve" and just begin with "achieving the potential of health IT's role in improving the nation's health and health care depends on achieving this vision". So, I agree. I don't think that we're saying that.
PARTICIPANT: No.
DR. FRANCIS: I would just start in this letter that may recommend specific actions --
DR. CARR: The first two things, yes.
MS. GREENBERG: Well I think, though, what you're suggesting is you introduce the observation reaction. So I think, makes a number of observations and recommends specific actions. So, it introduces the observation/recommendation format.
DR. CARR: Okay, good. Thank you. Okay, perfect.
MS. GREENBERG: So, you don't want to say the, or you're just starting with that.
DR. CARR: No. We're just starting with "in this letter, the Committee makes observations and recommendations".
MS. GREENBERG: Okay.
DR. CARR: Observation number one, the PCAST Report calls for development of a new universal data exchange language to support interoperable exchange of tagged data elements and Data Element Access Services, a mechanism to index, search, and access health information across the universe of systems where such information resides. The PCAST Report recommends that metadata tags would include information about the provenance and consumer privacy preferences of each tagged data element. In a PCAST vision, do I need to stop?
Okay. In a PCAST vision, offerers of DEAS will then be able to allow seekers of data (providers, researchers, and others) to make queries across multiple systems in a federated approach to identify where data about a patient resides and, with appropriate access controls, allow the retrieval of such data.
Okay. Continue in the observations. The metadata tagging in DEAS mechanism describe the PCAST Report exists today in commercial internet environments (search engines). However, given the depth, complexity, and sensitivity of health information, implementing these models pose significant challenges both for privacy and technology. Further, the PCAST Report recommended approach would require new and complex technical capabilities and information infrastructure components not currently in place in health care. Mark?
DR. HORNBROOK: Just a quick question. Is there any reason to make reference to public health surveillance use of these data? I mean, CDC looking at reportable diseases.
DR. CARR: Right. So, where would you put that under, in that offerers --
DR. HORNBROOK: In the parentheses, you have providers, researchers, but you leave out public health surveillance.
DR. CARR: Okay.
DR. HORNBROOK: I would think, given that we're NCHS, we should emphasize that.
DR. CARR: Yes. That makes sense. Mike?
DR. FITZMAURICE: In the middle of paragraph, of the observational in paragraph, it reads recommends that metadata tags would include. I would strike the "would" because it makes it sound like it's conditional.
DR. CARR: Yes, good. Just leave out "would include", makes a recommendation that tags include information about. Leslie, did you have? No? Any other comments?
All right. So the recommendation then is, before embarking on the deployment of the new universal data exchange language, metadata tagging, and DEAS infrastructure in health care envisioned by the PCAST Report, HHS should work with all stakeholders to do two things: 1) explore the feasibility of such a system, both technically and operationally, and 2) develop new policies, including those for governance, privacy, and security, that would be needed in order to implement this new approach.
And, we broke out a second recommendation, 1.2, identification of the near-term opportunities to exchange some set of metrics or data across the entire population in a limited time period as a way to demonstrate value and establish linkages among providers, population health departments, and other stakeholders.
PARTICIPANT: It doesn't read like a recommendation.
DR. CARR: No, it doesn't.
PARTICIPANT: It should do something.
MS. GREENBERG: It should be identify near-term opportunities.
DR. CARR: Yes.
DR. SUAREZ: It should be public health department, not population health department.
PARTICIPANT: Yes, that's what they're called.
DR. CARR: Okay. So change, put a verb at the beginning and change population to public. Any other suggestions?
PARTICIPANT: If you're talking about departments, yes.
DR. CARR: Yes. Okey doke. And then, a little more elaboration on recommendation 1.2 begins, health care organizations and providers are confronted with rapidly escalating demands to support quality measurement activities and to coordinate care using health IT systems. It is crucial that, while we envision and work together toward a fully interoperable national health information system, we achieve in connection with communities and care teams and in delivering near-term value to clinicians and patients.
DR. SUAREZ: Can I ask
DR. CARR: Yes, we've lost our format here. Yes?
DR. SUAREZ: One question on recommendation 1.2. I'm trying to understand how that relates to the PCAST Report.
DR. CARR: Yes. And, I think this is, let's go back to the question yesterday. This centers around the NCVHS letter on meaningful measurement that talked about developing data elements that build measures. So, it's similar. And so it's, I guess, an example of, you know, a near-term win.
DR. SUAREZ: Yes, but I mean I'm trying to connect it back to universal data exchange language, metadata tagging, and DEAS. I mean, those are the three key elements in the observation one and the recommendation 1.1. So, it seems like we're drifting to a different area.
And, I think the recommendation is fine, although it's hard to connect back to this question of is this instead of implementing the universal data exchange language, metadata tagging? Or is this as a way to test the universal data exchange language, metadata tagging, and DEAS?
DR. CARR: Matt, did you want to say something?
MR. QUINN: Yes. I think that this goes back to the vein of doing something, instead of waiting for the new world of metadata tagging, doing something now. And, this is more of an incremental plan to be able to do something now and increase our capabilities instead of waiting for the future.
DR. COHEN: Observation two and the discussion of metadata tagging actually follows the first paragraph above here much more closely. I mean, maybe you should integrate that into part of the recommendations in one and then keep the, as a separate recommendation focusing on doing something in the short-term.
DR. SUAREZ: Yes, I think that should be put out in a separate --
DR. CARR: Excuse me. Let's go down the line. Bruce, did you want to have something --
DR. COHEN: I'm sorry.
DR. CARR: That's okay. It was Walter who was out of line. (Laughter) You were fine. Just kidding. Vickie?
DR. MAYS: It is almost like, let's see, this 1.2 which is talking about metrics, etcetera might also then get moved when we begin to talk about development of measures that are meaningful to patients. So, I might take and put that as part of three because there's a recommendation three, there's an observation, then the recommendation, and then maybe it could go under that as a 3.1.
DR. CARR: Okay. So, I'm hearing pretty consistently the 1.2 doesn't belong here. Let's continue down the row. Bill, were you next?
DR. SCANLON: My concern about it was the issue of the entire population because I was saying that the short-term thing has to be universal across the country. And then secondly, I think that if you take that out, then go back to Walter's point, what are we talking about demonstrating?
Because I mean, I think we have already demonstrated about the exchange of information. I mean, Mark Overhage certainly showed that operation demonstrates that. So, we need to identify what's new here.
MS. MILAM: In recommendation 1.1, there are really two parts. And, the first part suggests that HHS explore the feasibility of a system. In the second part, we're telling them to go ahead and develop new policies.
And, we've heard from Joy repeatedly the level of effort required for the current Privacy and Security Tiger Teams. This is a huge level of effort. And, I'm wondering if we should have some sort of condition precedent for jumping into that such that
DR. CARR: Once feasible.
MS. MILAM: Yes.
DR. CARR: Once feasibility demonstrated? Are folks okay with that? Okay.
DR. WARREN: So, my question is recommendation 1.1. You're basically, by the time I read through this, if I'm an administrator, I haven't a clue what I'm doing because I have this enormous prepositional phrase in front that I've got to wade through before I get to the actual action. And so, I would delete the "before" and start with "HHS should".
DR. CARR: Okay.
DR. WARREN: And, if the stuff before is important, push it back up into the observation on what we're doing. But, by the time I get through the observation, I should read I've got the following things I need to do. And, that may be something that we need to split apart. I need to explore the feasibility for number one, I need to develop new policies for number two --
DR. CARR: Once feasibility is –-
DR. WARREN: And then, you know, I'm not sure about 1.2. I've got to think about that more on whether it needs to be moved.
DR. SUAREZ: Very quickly, I'd like to follow up on this point. It is important, the part "before embarking on the deployment of". I think what it is is it's convoluted because it adds a bunch of other things.
So let me just, one suggestion would be the start would still be "before embarking on the deployment of the new universal data exchange language, metadata tagging, and DEAS infrastructure, HHS should work with all stakeholders to" and then drop the "do two things". I don't think we need the "do two things".
So, it simplifies the statement and shortens the "before embarking". But I think it's important to highlight that before we really embark, I mean, we in the observation we say, you know, there's a problem with all these things. And then, in the recommendation we just --
DR. WARREN: That's true. So then, put the prepositional phrase at the end because you want to highlight the action. And, you can say because or, you know, some statement like that.
DR. CARR: So, I know we had some, yes Marjorie?
MS. GREENBERG: Okay. Well, I was going to say, since you were talking about staging these, you don't want to do them concurrently unless you want to determine, you could say, before embarking on the deployment of the new universal data exchange language, metadata tagging, and DEAS infrastructure", I tend to agree that that needs to be in here, "in the health care envisioned by the PCAST Report, HHS should work with all stakeholders to explore the feasibility of such a system, both technically and operationally. Based on that finding, HHS should then develop new policies including those for governance or whatever.
So that, you know, you've got first explore the feasibility and then, depending on what you find, which you might find that some things are more feasible than others --
DR. WARREN: But, what I am saying is first you've got to think of three things first, and then you get to what we want them to do. So, let's tell them what we want them to do, and it's because of those three things.
MS. GREENBERG: Well then, how about HHS should work with all stakeholders to explore the feasibility of the new universal data exchange language, metadata tagging, and DEAS infrastructure before embarking on the deployment of these features or something.
DR. SUAREZ: That's good.
DR. CARR: Okay. So, we'll add that HHS should work with all stakeholders before embarking on the deployment, or should work with all stakeholders to explore the feasibility of such a system technically and operationally before embarking on yada ya. Okay, and then --
DR. SUAREZ: The feasibility of deploying the new universal data exchange language, metadata tagging, and DEAS infrastructure. If that's what you were saying, right?
MS. GREENBERG: Right.
DR. CARR: HHS should work with all stakeholders to explore the feasibility of deploying and then --
MS. GREENBERG: Or, of the new universal data exchange language, metadata tagging, and DEAS infrastructure envisioned by the PCAST Report. The technical and operational feasibility, so you could bring that in rather than both technically and operationally.
DR. CARR: Okay. And then, the new sentence --
MS. GREENBERG: Based on these findings –
DR. CARR: Based on, yes, based on the finding, HHS should develop new policies including those for governance.
MS. GREENBERG: Right.
DR. WARREN: I would put HHS should develop new policies yada yada yada based on the findings.
DR. CARR: Okay.
MS. GREENBERG: Okay, you're tough.
DR. CARR: Okay, let's hold on 1.2 and move onto two and come back to 1.2 to see whether it adds value or not. So, where are we? Observation two, which is at the bottom of page five. Actually, do you have some additional language on this one? Matt?
MR. QUINN: We had some more discussion here. We had some more here on underneath the
DR. CARR: Observation two. Can you make that bigger so we can read it?
DR. FRANCIS: It would be great if you could also do it in final rather than final mark-up. Okay, and then make it a lot bigger, a whole lot bigger.
DR. CARR: Okay, great. Observation two. Now, let's have a reading. Matt, can you read what's up there?
MR. QUINN: The metatagging approach to data liquidity described in the PCAST Report raises a variety of privacy and standards concerns.
DR. KLOSS: Excuse me. The notation that I put in there calling for a little more detail was to make sure that that description, metatagging, was in the letter and kind of parallel to the description of the universal exchange language that was under the first.
DR. CARR: Right. So, is that what you added, Matt?
MR. QUINN: Yes.
DR. CARR: Okay. So Linda, see if this helps you.
MR. QUINN: The PCAST Report describes an approach to privacy and security based on tagging individual pieces of data with privacy-related information as part of their metadata, DEAS, and their components –-
DR. CARR: Is it privacy-related? Or, is it privacy preferences?
MR. QUINN: Privacy-related information is their words.
DR. CARR: Right.
MR. QUINN: DEAS and their components would act much like today's web search engines but with additional levels of responsibility for exposing only those data elements authorized by applicable privacy rules and policy, including a patient's preference to privacy choices and only to authorize authenticated users. Now this, pay attention to this one because I don't want to go too far on it.
While the approach envisioned in the PCAST Report could allow for greater granularity in reflecting individual privacy preferences and other benefits it, as the Report notes, assumes the existence of certain national infrastructure for finding health data and for controlling access to it. Without additional details and validation of this assumed infrastructure, NCVHS does not support the use of metadata tagging, of privacy preferences at the data element level.
DR. CARR: No, that's in, no. We've had a lot of discussions about it, and I think it should stand as written. NCVHS, you know, at least in where we are today, our history of what we've published, NCVHS does not support the use of metadata tagging or privacy preferences at the data element level. Is that --
MR. QUINN: I wanted to write because or to have something there because we talk about how, the Report talks about the potential benefits of it. And so, we should have something there.
DR. CARR: Well, do we say has not supported? Or, no.
DR. FITZMAURICE: Or, maybe at this time. It seems like you haven't done the research to make the assertion.
MR. QUINN: Yes that's, I was trying to get at without validating it or without knowing what the assumed infrastructure is that you can't know.
DR. FRANCIS: We don't particularly care whether somebody actually tags it that way. What we don't support is the implementation of privacy through that methodology. We're also, I mean, we're very worried about the effects on clinical care of allowing patients at the data element level to mask --
PARTICIPANT: Patient safety.
DR. FRANCIS: Yes, clinical care and patient safety were also, whatever else anybody does, I mean, you can tag it however you want and just not use the tags, whatever else anybody does, we have emphasized the importance of categorical ways of managing information for certain specified legal purposes. For example, if you're going to release a report with respect to gene protections, you have to have separate handling of genetic information.
DR. CARR: So we had extensive exchange of discussion and e-mail. So, I'm very respectful of the amount of work that went into the development of the language here. So, I really --
MR. QUINN: So, just leave it as NCVHS does not --
DR. COHEN: Is there anything that we would find out that would change our opinion about whether we would support the metadata tagging of privacy preferences?
DR. SUAREZ: At the data element level.
DR. COHEN: At the data element level.
DR. CARR: Would we say has not supported the use?
DR. COHEN: That was the question. I don't know.
MS. GREENBERG: I think that Leslie makes a really important point, and it's been bothering me since the caves of Turkey and that is that I don't really think there is a direct link back to a letter in which the Committee has said, we're against metadata tagging at the data element level. But, what the Committee has said is that it does not support privacy policies based on identifying patient preferences or based on patient preferences at the data element level. I don't think it mentioned metadata tagging.
DR. SUAREZ: Well I mean, metadata tagging is the technical way of doing that. So by inference, the metadata tagging is not.
MS. GREENBERG: But, it doesn't say you couldn't identify at the data element level but that the policies for use and for sequestration, etcetera shouldn't be based at that level.
DR. FRANCIS: I mean, somebody could be interested in metadata tagging at the data element level. I mean, we didn't say you shouldn't have that interest. We just said that patient preferences with respect to interoperable exchange should be categorized rather than at the data element level.
And, we also said that there needs to be, as interoperable EHRs move forward, there need to be ways of, whether it's through a metadata tagging or some other mechanism. I mean, it could be metadata tags that categorize, right?
MR. QUINN: Are we confounding the policy and the technology?
DR. FRANCIS: Yes. I mean, I think the policy has to be at the categorical level. That's our point.
MS. GREENBERG: Yes. You could tag at the element level, but the policy doesn't --
MR. QUINN: That doesn't follow our prior policy.
MS. GREENBERG: Right.
DR. CARR: Okay. Let's get, I think we just need to get back to --
DR. FRANCIS: It actually says it right after you get past the first sentence.
MR. QUINN: So, take out the first sentence?
DR. FRANCIS: So, if you just took out that first sentence –-
MR. QUINN: Agreed.
DR. FRANCIS: And say that we see problems with respect to data de-identification, and we see problems with respect to privacy management strategy. I mean, there are two points there. One is that the PCAST strategy doesn't deal with de-identification in a way we support, right?
And, that's our HIPAA now. Take it up to the beginning of the observation please. Can you make it just a little bit bigger please?
DR. CARR: The metadata tagging approach to data liquidity described in the PCAST Report raises a variety of privacy and standards concerns.
DR. FRANCIS: Can I actually suggest that it should say "raises important privacy and standards concerns"?
DR. CARR: Okay, good. The PCAST Report describes an approach to privacy and security based on tagging individual pieces of data with privacy information as part of their metadata.
MR. QUINN: There should be a period there.
DR. CARR: Period. Okay. DEAS and their components. Who's they?
MR. QUINN: It's –-
DR. CARR: And it's components would act much like today's web search engines but with additional levels of responsibility before exposing only those data elements authorized by applicable privacy rules and policies including a person's persistent privacy choices and only to authorize the predicated users. Okay. So, that's the description.
So Linda, does that make sense? Understandable? All right. So, we've achieved that. Two paragraphs.
All right. Now, on to the next paragraph. While the report showed vision, the PCAST Report could allow greater granularity in reflecting individual privacy preferences and other benefits it, as the report notes, assumes the existence of certain national infrastructure to finding health data and for controlling access to it.
So Matt, this is your language. Is this, do we want to say this everyone?
MR. QUINN: It was, I added that because I wanted that dependency on the next sentence.
DR. CARR: So, do we still need this?
MR. QUINN: I don't think so.
DR. CARR: Okay. Take it out.
MR. QUINN: I'll let Leslie --
DR. FRANCIS: Or, you could take the while clause, and after the comma, say something like it fails to take account of concerns about de-identification and the implementation of privacy preferences that this Committee has emphasized or something like that. Because the problem isn't the lack of a national infrastructure; it's that it's just misguided.
DR. CARR: Do we want to just take that out?
DR. FRANCIS: Oh so, you want to just take the whole thing out.
DR. CARR: Just take that paragraph out.
DR. FRANCIS: Yes, take the paragraph out.
MR. QUINN: It is a huge assumption that, you know, this all relies on an assumed infrastructure that doesn't exist, it's not validated, etcetera.
DR. CARR: Okay. Assume you took out NCVHS does not support the use of metadata tagging, privacy preferences at the element level. Are we, is that being true to the discussions we've had previously?
MR. QUINN: We've removed that.
DR. CARR: We've removed that. So, it then just goes to NCVHS as emphasizing the importance of developing appropriate privacy, confidentiality, and security policies and practices with respect to the sharing of information in interoperable health records. In several respects, it's unclear whether the PCAST strategy for de-identifying data will meet these goals. For many forms of data and use and re-use, it will be necessary to link data points in a record. One important policy question is how to realize the PCAST Report's vision of privacy when data linkages across organizations over time are needed for other components of its visions.
DR. FRANCIS: So I would say, instead of in several respects, to sharpen up that paragraph I would say "a first concern about the PCAST Report is that in several respects, it is unclear whether the PCAST strategy for de-identifying data is appropriate". So, make it clear that what we're doing is, our first concern is about the PCAST de-identification strategy. And then, there's several sub-issues about that: linking data points in a record, dealing with genetic information --
DR. CARR: So, if you can write that down so that he'll have it, is there any objection to what was just added? I'd like to move on to the next paragraph that we tried to shrink down taking out Dr. Garcia and all, the example that they gave, and simply go to the PCAST Report proposes an automatic real-time de-identification process for protected health information in response to a query. De-identification, let me see, for certain queries in situations and under certain laws may require more than following a template algorithm. Researchers require different data sets, and so on.
Is this more information than we need? I mean, we talked about the length of this. Is it important to keep all of this detail?
DR. FRANCIS: I really think, and this is something I've been thinking of, that we need to put in earlier on, we didn't put in earlier, we just say that the strategy, we need to look at feasibility of the strategy. Somebody from Google is going to say metatagging is totally feasible.
What we need to do is say feasibility for health information and clarify that at the, you know, back in those earlier paragraphs and then say, here I think it's really important to have this, why is de-identification of health information in this strategic way problematic? I really think we have to have that explanation or people won't understand or Google isn't going to figure it out. Right?
DR. CARR: So, I want to make sure that we get a consensus. What I would like to do is probably jump ahead to the recommendations, ask Leslie to maybe take a look at how you'd like that written.
DR. SUAREZ: Okay, because I do want to point out a couple of things. I was trying to go back and back and forth, and there's even some places under observation one where we, at the end of page four, we again talk about the fact that NCHS does not support the use of metadata tagging of privacy information.
So here again, so I started thinking, well where is the recommendation standards and where is, I think at the end of the day, there's five areas of recommendations we're making. Number one, we're making recommendations about the whole issue of embarking on universal data exchange, language metadata tagging, DEAS infrastructure. That's number one. So, it's more of a technology issue.
Number two, we're making recommendations about privacy. I mean, privacy we're making at least two if not three recommendations. Number one, we're concerned about metadata tagging at the data element level. Number two, de-identification. And number three, there's a whole host of implications about population health and health services research and secondary users.
And, I think when you read the whole observation two and then the patient to have sort of observation, you get all confused now. So, we need to really be identifying the specific steps. So, number two is observation, recommendation, privacy which includes those three that I mentioned. Number three, we're talking about recommendations on quality.
And number four, we deleted but I want to bring back the recommendation on the standards because this is a very important one. If we want to go down the path of metadata, we need to really emphasize the fact that there needs to be a process to define and establish the standard for the metadata, totally independent from privacy and security and all those things. So, that's number five.
And, number six is the final one about patient identifiers. So, if we pick any chunks smaller like that and just do very shorter descriptions of those and don't mix up, you know, sentences that appear in multiple places, I think it's –-
DR. CARR: Yes, I agree with that.
DR. WARREN: So, I have a concern too. One of the things that I've been struggling with is that on observations one and two, we have observations, we have recommendations, and then we have a bunch of paragraphs.
And then, on observations three and four, we have observations and recommendations. And, I think that's the model we need to follow because I don't know what to do with the paragraphs.
DR. CARR: No, I agree. And, there was a model in the beginning, but it unraveled. So this is, I'm going to make a recommendation. I think, Walter, your framing is very, very helpful.
And, what I think we'd like to do is maybe Matt, if you can take some of what Walter has recommended and try to move things around, then at the lunch break, I'll work with you or anyone else if I can get input. And, we will work to achieve that with the clarity and the consistency in terms of the formatting.
And then, bring it back right after lunch hopefully in a form that is acceptable. It's important that we get this right. It always seems that we run short of money, I mean time and money. But, I think that --
DR. FRANCIS: Justine, I think there's some really important stuff this afternoon. And, I would hate to have us cut into any of that. I will volunteer to go to lunch with you and Matt and others and then have something we could send out to everybody. I think everybody knows what we want.
MS. GREENBERG: I think if you can maybe at lunch agree on the recommendations or at least get those laid out, and if you can get agreement right after lunch on that, it could then be approved and referred to the Executive Subcommittee. Otherwise, I mean we're talking about a pretty substantial re-write here. And, as I know in the past, people have been uncomfortable giving that much authority to the Executive Subcommittee.
But, I don't think this can all be done in a way that could be approved after lunch. And, I agree with Leslie. So, I think we may just have to send it back to a small team of writers and do a call.
DR. HORNBROOK: I think that we need to make certain all of us are clear which I wasn't until after here. The issue is you can metatag any variable or element with all the rest of the variables in the file. So, I could say I've deleted a person's name, date of birth, address, phone number. But, if I put all of that as a metatag, it's still there. Is that what we're trying to address?
MS. GREENBERG: No. I don't think so.
DR. HORNBROOK: I mean, are you trying to address the technology of metatagging? Or, are you trying to address what people put into the metatag?
PARTICIPANT: Both of those.
MS. GREENBERG: And then what they do with it.
DR. HORNBROOK: So, it's in a sense a parallel as what you put into the list of variables in the traditional flat file that most of us think about. I mean, this is sort of hyperspace information processing.
DR. CARR: Well, yes. So I think, you know, what we'll do is come back with the recommendations in the format that Walter recommended. And, we'll see where we are. Then, I think what you need to think about is do we need to convene the Full Committee to finalize this? Or, do we hand it off to the Executive Subcommittee. And, I think that really is a function of how much re-work there is and, you know, what folks comfort level is.
You know, it's unfortunate that it's taken this long to do this. But, I think the input and the growing appreciation for the implications of it, in part, is what is contributing to this. And given that we've really not had face-to-face time to meet about this, it's understandable. Unfortunate, but understandable.
DR. SUAREZ: Can I say real quickly, I was just writing, typing what I said because I will forget it. But, as I was typing it, I realized each of the sections is one of our subcommittees. There's the recommendations and standards and privacy and population and quality.
MS. GREENBERG: That's why we did it this way.
DR. CARR: Yes. So we've kind of rearranged and lost that. So, we'll get back to that. So, I'll conclude this right now, and yes, thank you all for doing that. Judy?
DR. WARREN: Just to report on my task, so we did talk about patient matching methodologies in our letter on the NHI Framework. So, that's where we had the hearings and all of that. So, we can refer back to that letter.
MS. GREENBERG: Yes, which we have actually. But, what about the linkage letter?
DR. CARR: Yes, the population linkage.
DR. WARREN: But, it doesn't talk about the methods to link the data. It talks about the policy. And, my concern is the methodology and the research behind the different strategies to link data.
DR. SUAREZ: So, the record linkage that are in review, it is a workshop that does point out data linkages to improve health outcomes. There were a number of policy recommendations. The ones specific to data linkages said standardized data linkage and sharing agreements should be developed to facilitate sharing of data among HHS agencies.
MS. GREENBERG: Right. So, is that relevant to this letter?
DR. CARR: No, I don't think so.
MS. GREENBERG: Probably not.
DR. CARR: Okay. All right, so we have to then be cognizant of that when we put our alignment things in.
DR. WARREN: So, what I will do is I'll e-mail the exact place in this to both you and Matt, and then you'll know.
DR. CARR: Okay. Next up, I believe, is Walter. Is it you?
DR. SUAREZ: Is it me?
DR. CARR: Yes. It's all about you today.
DR. SUAREZ: It's all about me, huh? (Laughter) So, Standards. All right. this will be very short. It won't probably take the whole hour or 50 minutes that we have to meet.
So, we wanted to bring back to the Committee a couple of points that we mentioned yesterday. One is the review of the findings and directions on the acknowledgments recommendations out of the hearing. And, the other one is a review of the HIPAA Report to Congress outline and the progress to-date.
So, on tab six of your book, well let me start with the Report to Congress is as in the order in which it appears in the book. So, I don't know that the slides are on the presentation. Do you have them? Can I have the book? Or, should we use the slides? Are the slides in the, or the slides are in there, but some people might not have the book.
DR. CARR: Everybody, oh.
MS. GREENBERG: Well, people in the audience won't have it. This is the overview of the Tenth HIPAA Report to Congress.
DR. SUAREZ: Yes, so while they're pulling it up there, we'll start with the Tenth HIPAA Report to Congress presentation. So, the opportunity really was to take advantage of this tenth anniversary I guess or not anniversary but Tenth Report to make it a little special report and provide a more extensive review of where things are, talk about where we started with Administrative Simplification, where we are today, and where things are going.
It also kind of brings up the fact that we're about to, you know, engage in the transition to the next version of the HIPAA standards starting January 1st and to the new codes at ICD-10 starting in October of 2013. So, it gives more weight to the fact that we're now moving into that new step in the journey. But, it is all about the journey, and so, that's what we wanted to point out in the Report.
We're building on all the previous work, reports, HIPAA hearings, use of existing industry information. We're not doing really a new data collection on where things are. There is, in fact, a hearing tomorrow where we will hear more about where things are. There are surveys coming out that have been done in the past and new ones coming out later this year.
So, we're really building on a lot of it. And, when you put it into a report, and we started really drafting this, and it's amazing the amount of work over the last, of course, ten years or so that have been happening just with the Committee.
So, the report outline is as presented in the slide. We started with a transmittal letter. We will have an executive summary of the Report. And then, we divided the Report into three sections.
The first section is about the journey and the evolution, if you will, of Administrative Simplification. What is Administrative Simplification, the description of the various components of Administrative Simplification, you know, standards, causes, identifiers, privacy and security. And then very importantly, how has NCVHS served as a catalyst in this journey? And, there is significant, of course, work that NCVHS has done in convening the industry, making recommendations, all those steps that we've taken.
Section two is the implementation of HIPAA Administrative Simplification, so where are we today? How well have HIPAA transactions, code sets, identifiers been implemented? How well have HIPAA Privacy and Security Standards been implemented? What has been the effect of HIPAA Administrative Simplification implementation? What's the evidence about, you know, are we better off than before? And, what are the gaps and issues that remain? So, all of those are part of the section two.
And then section three, on the next slide, is the future of Administrative Simplification including some of the things that we know about now like the next version of the standards, the next set of transactions, the next identifiers, all those things under the Affordable Care Act as well as other things beyond what we know today, like this concept of the linkage of Administrative Simplification with the overall effort in the industry to move toward Electronic Health Records and meaningful use of those records and health information exchange adoptions and things like that. So, that's what's going to be the chapter three or section three of the report.
We will have a number of attachments to support all of that information including some of the reference materials that we have published with NCVHS. Next slide is the process.
So, primarily we have done work as we have done in other activities of the Subcommittee, engaging a subject matter expert to help us with the research behind this and the drafting of the report. So all of it, it really is a job of researching, identifying resources, collecting the information, and then beginning to draft the various sections. And, that's where we are right now, in the process of completing that drafting.
Again, the resources that we have used as a sample are NCVHS previous letters, reports, hearings, industry published letters, data, reports, analysis of implementation like WEDI reports, Gartner surveys, association surveys, reports from CMS and OCR and other agencies on the implementation of HIPAA. So, all of that is information that's out there, and all of that is what we're bringing together into this section.
So where we are today in terms of timeline, we started this process back in actually more like February with planning, and between March and now, we've been again drafting various sections. We have been having periodic meetings with the lead staff that is helping us with the drafting. And, we are now at the point where we're going to be bringing to the Subcommittee on Standards a draft of the report in the next couple of weeks.
The Subcommittee on Standards is going to be meeting, working on the draft, expanding it, and you know, providing input. We are also going to be, of course, working with the Privacy, Security, and Confidentiality Subcommittee on the section related to privacy and security on the report. And over the next couple of months, July and August, we will be kind of finishing out the first draft.
We'll present, hopefully, the draft by the end of August sometime to the Executive Committee as a draft. And, we expect to be able to bring a draft to the Full Committee in September for the review, be ahead of the September meeting so that we can bring it up to the Committee. I don't expect that we will, of course, do a full review and wordsmithing of the Report at that point. We will be taking in comments, you know, general direction comments and high-level comments. Any wordsmithing, we can take the e-mail and offline.
But, we wanted to try to get the Report finalized by September, but of course, I'm pretty sure with the review process we always are engaging and are still engaging. We are expected to have to do two passes to the Committee. So, one will be in September, and then, we will go back with both the Standards Subcommittee and the Privacy Subcommittee and work at changes that we need and then bring it back as a final report to the Full Committee in November. And we expect that then, by November, the Full Committee will hopefully approve it, and we'll be able to file it or submit it to the Congress this year, 2011 still.
So, that is the timeline. And, let me stop there and see if there are any questions about it.
DR. KLOSS: Has the Subcommittee or the Full Committee considered whether some of the high points of this report might be prepared in a way that could be more broadly communicated to the industry as a way of reflecting on actually how much has been accomplished and what an impediment it would be if, in fact, we hadn't had the beloved HIPAA legislation? So, it may be an opportunity for some broader industry talking points, communication plan.
DR. CARR: Yes. I mean, I would envision that we would put this together as some of our other publications but really anticipate distributing it widely electronically, beyond the paper. Okay, Larry?
DR. GREEN: Building off of your opportunity slide phrase "where things are going" and then skipping to section three of the report outline, "the future of Administrative Simplification", what is your thinking? And, is it proper in this letter to recognize that all roads to health care reform go through payment reform and that fee per service is a failed strategy? And, that so much of Administrative Simplification is designed to enable the adjudication of item by item, piece by piece charging?
DR. SUAREZ: That sounds like section four of the report.
DR. GREEN: There was a question here, Walter. I mean, this may not be the right place, the right time, the right way. But, it troubles me greatly that ten years out of HIPAA we're going to supply a report that is silent on the fact that the fundamental way we pay for health care must change. We know this.
DR. SUAREZ: Well, I can give you my reaction, and then, I'm sure others will have. I am very much in agreement with that idea. I mean, clearly the industry is moving towards, and certainly, you know, Federal efforts and other private sector efforts are looking at that and moving towards a new model, a transformation of the way we do business in health care.
And, it's a transformation that includes those points of saying we're not anymore doing business as in you do, we pay, you do, we pay. It's more about you take care of a population, and we, you know, look for insuring that the taking care of the population happens.
So still, you know, we do have a very strong and significant role and a very extensive set of regulations that precede these. And so, we have to be mindful of that. That's why the Report's written, it's in recognition, of course, of the fact that we have a law, and we have significant regulations.
I do think that we should be moving into, writing something into the report that points to that. So that's, I mean, I wasn't facetious about maybe it's in section four.
DR. WARREN: Let me give a very concrete answer to Larry's question. This is a report required of us by Congress on the status of HIPAA implementation period. So, that's what the report is for.
We have noted other efforts that are going, and but all of the stuff that we've put in there is based on hearings that this Committee has had. To my knowledge, since I've been on the Committee, we've never had any hearings on what you're talking about.
MS. GREENBERG: Well, nor would it be appropriate.
DR. WARREN: Well, and so the assignment is we can then discuss whether or not to have that. But in this report, I don't --
DR. CARR: Marjorie?
MS. GREENBERG: Okay. This, as you know, is an Advisory Committee on National Health Information Policy. Now, the whole issue of how we pay for health care, etcetera, obviously influences the type of data that we need and the type of data that we use, etcetera.
So, it's not that the Committee cannot address, certainly needs to address the type of information needed in these different health care or health care insurance models one of which is fee-for-service. I think, actually, HIPAA does have implications beyond fee-for-service because you're still going to need to collect information in order to do all the other things besides just paying the claims. But, I think going into which is the best way for us to pay, for the country to pay for care is basically out of scope for this Committee.
DR. GREEN: For the Subcommittee?
MS. GREENBERG: No, for the Committee. That doesn't mean, the information implications are very important. But actually, how we should reform our payment systems specifically, I don't think is within the scope of this Committee.
DR. CARR: Okay. I'm going then, I'm sorry, Blackford? And then, okay, Blackford.
DR. MIDDLETON: You know, I think Larry's point is well taken and well understood and appreciated by all. And, I agree with Marjorie at the same time. It's not the role of the Committee to make those level of policy recommendations per se.
But, what I think is entirely appropriate is that we reflect upon HIPAA at ten years, and we are entirely cognizant of what's coming down the road in a variety of potential scenarios that we have to suggest how HIPAA has to be viewed in the context of these potential futures. And, we don't make recommendations about one or the other.
But, if we're moving toward X, Y, or Z, then HIPAA has to be mindful of what X, Y, and Z might be. I think that's in the gap analysis of the future part where we have basically liberty I would suggest, not to make recommendations but to say, if this were to happen, blah, blah, blah.
The second thing, though, just while I have the mic, I think the, you know, the review of the initial draft process July through August, I'd love to volunteer Paul and myself to contribute to the quality dimension of this and at the same time, volunteer the Populations Committee. I think in our considerations of impact that we have to look at what we think are the future sources of data quality which has been directly relevant to HIPAA's existence or history to-date and potential future scenarios and similarly Populations.
So, I just want to volunteer that. I think we have to have a similar analysis on the population and quality front.
DR. WARREN: So Blackford, what you are suggesting, I mean, traditionally this report has been the responsibility of the Standards Subcommittee to initiate and then run through Privacy because there is privacy in HIPAA, as we all know. And, once we get through those subcommittees, to go to the Executive Subcommittee and then to the Full Committee.
What you're suggesting, though, is maybe what we need to do is at the point that we're running it through all the subcommittees, we should just run it through the full Committee.
DR. SUAREZ: I heard something different actually. I heard that maybe there is in section three a literal, a smaller section about quality and a smaller section about population than what we just read.
DR. MIDDLETON: I am assuming no responsibility for the report.
DR. CARR: Yes, so I think this is just an incredibly value, great suggestions because we've had this responsibility report, and we've taken it in a certain way. But it is certainly, at a ten year point, a time to look at the totality and that it's a report from the Committee. So, it should have input from the Committee. So, Len?
DR. NICHOLS: So, as much as I love Larry's passion and share your vision of where we're headed, I'd offer two thoughts. One, even if we started tomorrow and everyone in this country agreed which, of course, they don't, it would take ten years to get rid of fee-for-service. It's going to be around a long time. So, the stuff that's been done is actually going to be valuable.
Second, when I look at the best I can imagine us being able to bundle, package, lump, however you want to call it, there's going to be 15, 20 percent fee-for-service past ten years because it's just too complicated. Everything else that's going to go on, which is in my view the real core here, is the claims adjudication standardization. That's what's going to reduce the administrative cost for providers. That's the future.
And, that's going to be the same whether there's 8,000 CPT codes or 40. That is what really is going to affect cost in the system. And that's, I hope, I guess that's what section, the opportunities are that's not yet in your timelines anywhere, right? Although ACA calls for it if I understand, but I guess that's my question, Walter. Where does that fit into this report?
DR. SUAREZ: Well, I mean it is part of the future really. It's part of the statements that can be made in this section on the future.
Now as a process, one thing is writing it in the report, the other thing is actually executing it is what maybe you're asking if where in the process into the future would the Committee or Subcommittee actually engage in those points? But, I think that is an agenda that we can look into particularly if we insert some of those statements into the report as part of the chapter three, the future.
DR. NICHOLS: That's a good reason to do it, in my opinion. I would just say if I could, while it may be out of scope, it's also our responsibility to acknowledge that is the way we're going whether it's ACO or --
MS. GREENBERG: I agree completely. It's just how we frame it. I mean, I think we need to frame it within scope, to not ignore it.
DR. NICHOLS: ACOs are here to stay, but we do need to say fee-for-service is probably going to be --
DR. WARREN: So one of the challenges we have is that the letter is about HIPAA period. And, we've also done things with the NHI Framework, we've done things, what?
DR. NICHOLS: But, we are going to mention the ACA right here, right?
DR. WARREN: I know. And, what I'm saying is in the future, we're going to mention the other bills that this relates to. And, it's in that future section, that challenges ahead that we can start linking up to some of this.
But, it's not going to be, I think, to the extent the Committee would like to frame things. It's going to be framed in relationship to HIPAA. So I just want to keep bringing people back. This letter is to Congress solely about HIPAA. It's a requirement in the legislation. And so, we need to frame that that way and the less we're going to have to change our minds.
DR. CARR: So, we're going to have Bill, Bruce, Linda, and Marjorie. Bill?
DR. SCANLON: Most of the time I've been on this Committee, I've spent time also on a committee that was charged with payment policy at MedPAC. And actually, I thought the two were complements because there is, in my mind, a very strong link between wanting to make payment reform and having the data that's going to support it.
DR. CARR: Of course.
DR. SCANLON: And to be honest with you, we've moved from fee-for-service in home health to a bundle payment in home health, and it's unclear which one is the bigger disaster. Okay? And so, I think what we need and what I've been pushing for in this Committee, and I used to take some of the message from this Committee back to MedPAC was the issue of we've got to get more data.
And, this is why I think, you know, being positive about the PCAST Report is very important because we've had a whole history here of excuses about why we can't have data. And, the reality is we can have data.
There are certain conditions we have to meet to get those data, but we've got to sort of get to the point where we can meet them and then we can have the access to information so we can develop the bundle, bundles and other sort of forms of payment reform because the other interesting reality about sort of payment in health care has been that when we have a good idea and we implement it poorly, it becomes a dead idea. And, unless you kind of find some clever language to resurrect it under a new name, you can't do it, okay? I mean, because people will tell you, well, we tried that, and it doesn't work. So, I think we've got to worry about this from a strategic perspective.
In terms of thinking about the future, I mean I think, and I'm not sure how many members of the Committee would go along with this, there was in my mind when we did the quality sort of data letter, okay, this issue of that what we really want to be able to get out of EHRs is a set of common data elements that we can work to build measures, I mean measures that we haven't envisioned yet, but we're going to work in the future to be able to enhance our quality measurement by having some data that we can work with. When do we reach the point where those data are starting to flow routinely in the form of claims? And this, you know, this would be a great simplification.
MedPAC has come out and said we should have live values on claims. All right? This would be a fundamental change in terms of being able to do risk adjustment, to be able to bundle things together. I mean, it is even a fundamental change in terms of some of the bundles we've got now like inpatient hospitalization.
So, I think we've got to maybe lay out some of this at least sort of preliminary language in terms of what are the possibilities? What are the desirable things to be thinking about for the future?
DR. CARR: Excellent. Bruce?
DR. COHEN: I agree with the general sentiment that this is a real opportunity both to focus on the short-term future and the longer term future particularly in the context of data needs and data quality and information exchange. There's nothing wrong with making shorter term recommendations like Len was talking about and envisioning longer term data requirements and needs if there were different approaches to the delivery of medical services.
So, I feel that this is an opportunity to reflect on that as well. And, I would encourage us to include that as part of this report.
DR. CARR: Linda?
DR. KLOSS: I think we are on a theme here because of my question and I guess suggestion was whether there is a consensus on the part of the Committee as to what the longer term opportunities are with regard to HIPAA, even the discussion we had earlier this morning in terms of expediting updates to standards and issues like that. And if not, then we need some dialogue and identification of what the four, five, or six key areas are that need to be highlighted in the future section.
DR. CARR: Did you want to respond on that, Marjorie?
MS. GREENBERG: Yes. I think this has been a really good discussion. And, I thank Larry even though I told him he was out of scope. (Laughter) I thank him, as always, for stimulating us and getting us to talk about, you know, things that are really important.
And so, I think in that context, I think that, I mean I understand and I agree with Judy that this is part of a mandate that we've had every year since HIPAA was passed. And, we have to be mindful of meeting that mandate.
But, I think we were challenged by Justine, maybe six months or so ago, to take this opportunity that it is the ten year, the anniversary tenth report that we're doing. It's not the ten year anniversary so much, but the tenth report and to make it more than just a reporting on what was done this year and, you know, what needs to be done.
So, I think we have a few options. One is not to worry too much about the fact that this is the tenth in a line of reports that have been much more sort of factual and to really make it an omnibus study that, you know, captures these other areas and brings in other pieces of related legislation.
I certainly couldn't agree more with Blackford because as we were talking yesterday, Walter and I, maybe it was in the Population Health Committee, but HIPAA actually stimulated tremendous bringing together of communities that used to be one-offs, I mean, used to hear if you've seen one Medicaid agency, you've seen one Medicaid agency. And yet, they were all kind of doing the same thing although they're different obviously across states. Well, they had to come together to agree on their data needs because of HIPAA.
Same thing happened with public health. And, that's where we developed the Public Health Data Standards Consortium. I think all of that actually belongs in a report like this because it was HIPAA that stimulated those things to happen. You know, it didn't all work out the way you might want.
But so, there's that possibility of this omnibus. We could also, you know, after we get the whole thing done, extract something that is a more traditional HIPAA report to Congress if you wanted to do that. And then, make this other report something beyond that.
I mean, I don't think we have to worry about that right now whether we're going beyond what our specific mandate is because we aren't going beyond the mandate of the Committee. So you know, I think we just see where this comes out, if it's too much forward thinking for what we think is appropriate in this particular report to Congress.
I also think what Linda has suggested is excellent where we not only have an executive summary, but we can do other smaller types of documents that are web-based or whatever to try to push out the report. In fact, after going to the Todd Park conference last week, I started thinking of really ways that we could make our website a resource center. But in any event, we won't go there now.
DR. CARR: And so, I have Judy, Vickie, and Larry. But, I just want to interject that I think this new approach is not just a tenth annual –
MS. GREENBERG: I agree.
DR. CARR: That we can perform an enormous service by bringing all of this together and owning the progress that is or is not being made and continuing to push forward. This is why we set deadlines, this is why we have deadlines, and here is the price. We heard some enormous prices need at EFT, ERA discussions that were mind boggling when you think about balancing the budget. So, Judy?
DR. WARREN: So, I have a recommendation in listening to all the perspectives. One, we need to get this report out. It needs to come out as a HIPAA framed report. Now, we also have a contractor that contracted to do certain things. And, we're talking way beyond what's in her contract.
Now, so saying that, I agree with everything that has been said, but, what I'd like to propose is, you know, we've talked about these big cross-cutting things. One of the things that Walter and Lorraine and Margaret and I keep coming back to, and we have to re-think about what's our purpose here is we are now having a set of legislations and regulations that are coming out that are all about data.
I think we need to look at a second report that really pulls all of these together. So, here's what's happening with HIPAA. And then, we followed that by e-prescribing, and we followed that by the functional requirements for NHIN. And then, HITECH came into place, and it had requirements. And then, ACA came in, and they've got requirements.
And so, all of these have basic data requirements. And, I think we could do a really good impact in laying what all those out. And then, I think we'd meet some of Larry's, you know, original comments of we need data to support what Larry is talking about, we need data to support what Bill is talking about in reform.
And so, if we meet our mandate from Congress and give them the HIPAA report, and then we generate this other larger let's look at the landscape and what's happening and then project out the five or ten years as you were talking about, the economics, I think we'd be much better poised to do that kind of work with each of the four Committees adding a chunk to that report.
DR. CARR: I think that's very good. And, I actually also think as we do our report out on the work ahead for the Committees, we might want to now incorporate this into priority, as you say, for the resources that we need to get this first statement out. If we haven't embarked on some immediate thing and particular Subcommittee, for example Quality, we might think about just making this a priority to work with this because we can look at what claims data is the fundamental element that exists today.
DR. WARREN: The concern I have for the HIPAA Report is we keep adding stuff. We're always going to add stuff because the landscape is changing so rapidly. And, we will never get it out.
DR. CARR: Right. No, we have to have something. But, I think it teases out a feed. You know, as we look at the other FACAs, we look at all the work that's being done, all the thing about Data-Palooza, EHRs, ONC, all the FACAs, you know, we don't need to be in the me too space. We need to be, first of all, in the space that is legislatively assigned to us.
And secondly, are we doing all we can and should do in that assignment? And I think, in fact, this opens the door to so many of the other things and even the Learning Health System that we talked about. So, I think this is fantastic. Okay. So, I have Vickie and then Larry and then Mark.
DR. MAYS: Marjorie was starting to go in a direction of something that I wanted to suggest and that is thinking about what's been going on in terms of the community data health initiative. One of the things that we have to do when we have NIH whatever grant or anything you get from them, they make us do this one or two page public health relevance.
It might be great if what we started thinking about for this report because, believe me, the community has a range of notions about HIPAA, be it good or bad. And the minute it comes up, it sometimes is a barrier.
So, if we could draw something out, and then again, I don't know what the resources are, but if we could do a one or two page, you know, finding that's good for the community about HIPAA and its state and be able to start doing this I think in general for some of our reports, I think we'll bring the community with us, we'll get better input from them in general. But, just like a two pager.
DR. CARR: Okay great. And so, I have Larry and then Mark, and Walter really has, what?
DR. SUAREZ: We have a second report.
DR. CARR: Has a second report.
DR. GREEN: In the letter, slide four, section three, the future Administrative Simplification, the last bullet "challenges ahead". I think Lynn and Bill really expressed important sentiments that Judy came very close to summarizing just a moment ago. And, I want to seriously agitate that they be put into this HIPAA letter there. And, the main reason I want that done is to set up what we just discussed for future work.
DR. CARR: Excellent. Mark?
DR. HORNBROOK: I don't know whether we want to say this, but we were all, in research, very much afraid of HIPAA that it was going to really stop research. And yet, over the years, we finally learned how to use Data Use Agreements and Business Use Agreements to accomplish a lot of, in fact, most of our research.
So, I'm finally doing a study where we're sending PHI to a survey research company under a business use agreement as a vendor to the research center so they can do a recruitment on a national study. And, that was never envisioned when we started. So, the horrors of HIPAA have eased quite a bit.
DR. CARR: So I think, you know, back to Blackford's point, I think each of the Subcommittees needs to think about what is the story, the piece of the story that we need to say. So, I'm going to cut off further conversation here, although it's rich, and Walter, you're back on.
DR. SUAREZ: Well, thank you. Of course, I thought we were going to take five minutes. (Laughter) So the next one, I promise will be only five minutes.
This is absolutely terrific. I mean, I've been thinking about it too. And, I think it's important to enhance this report with other perspectives without creating multiple reports inside the report out of it.
I mean, we're not talking about a, you know, 25 page chapter on population health and HIPAA and a 15 page chapter on quality. I think it's important to highlight it in a section appropriately and tease it out as an opportunity. So, that's great.
MS. GREENBERG: Can I just say too that if, I'm just, I am mindful certainly of the engagement we have with the consultant and the expectations. And if we need to, you know, then do a different type of engagement or an extended engagement or something in order to meet these needs, then we can certainly write about it.
DR. SUAREZ: All right. I appreciate that. All right. The next topic is about acknowledgment. So, I'm going to take you way down back into a very specific activity. So I mentioned, we mentioned yesterday this is one area of recommendations since we had a hearing.
We had earlier this morning a meeting with our Subcommittee where we started to look at the letter of, observation of recommendations coming out of the hearing. We had already a number of important updates and adjustments. So, this presentation reflects what we presented at this point to the Subcommittee, so there are some changes. And, we'll talk about the process at the end.
But, the important thing for everybody to understand is some of the key elements, concepts around acknowledgments. This is a brand new transaction under HIPAA. There was no regulation in 1996 that said you should do claims, claim payment, enrollment, eligibility, referrals, and acknowledgments. So, acknowledgments was something that came out new this year.
What are acknowledgments? A standard, a use to communicate the status of a transaction between trading parties. So, I send an eligibility transaction in a package, an electronic package, and the recipient has to open that package. And the idea is that they will send the acknowledgments about, as they receive the package, that they receive it and as they open the envelope and start looking inside, acknowledgment about any issues that they might have found and then more detailed acknowledgments in terms of a response about all sorts of issues and errors and mistakes.
So, acknowledgments are really ways to communicate data electronically efficiently, that a transaction has been accepted with some errors and is beginning to be processed or has been rejected. Those are sort of the three main outcomes.
They answer the three critical questions really. Did you receive my request or my submission, my transaction? Is it okay? Can you process by request? And if not, why not? And, we'll talk about some of the value benefits, but they do save a lot of time and money and resources. So, that's part of the business case for acknowledgments.
Usually, in the current mechanisms for those that don't use acknowledgments, providers have to do things like knowing if they have received the transaction and if the transaction was accepted. And if they can't do it electronically, they have to call and follow up and do a lot of this manual processing to try to get to whether the transaction was indeed okay.
Or, they might just forget about it, they don't know about it, and six weeks later or two months later, they wonder about what happened to that. The transactions are not straightforward, many parties are involved and trading partners are involved in them. Providers don't usually send transactions directly to the health plan, or the plan does not often send directly the response to the provider. There's a claim house in between or other trading partners.
There is a use of existing acknowledgment transactions, but it is inconsistent and usually has a lot of nonstandard formats. And, the voluntary nature of use inhibits really the widespread adoption of this in the absence of a mandate. So, next.
So when we had the hearing, we clearly identified there's three, the most critical three types of acknowledgments. And, these are sort of technical terms within the transactions. There is something called a TA1 acknowledgment which is primarily to acknowledge the receipt and acceptance of a particular transaction. It's sort of like when you send a FedEx package, the acknowledgment that the FedEx was received. Nothing tells about that yet.
There is another one called 999 implementation acknowledgment which is a notification that a transaction has been accepted or accepted with errors or is going to be rejected. So, that part of the transaction can be used. And, this too applies to all the nine plus HIPAA transactions, whether you're sending an eligibility inquiry or a referral or an enrollment transaction or any of those transactions, you can send a TA1 or a 999 to recognize the fact that you received it.
And then, there's a third one which is a specific acknowledgment for the health care claim. Of course, the health care claim is the most significant, important transaction. There's a really big transaction, of course, in health care.
And so, there's a specific transaction that acknowledges the receipt of the health care claim and provides a lot of the detail of whether the claim was rejected or not with some codification of a claim rejection codes and things like that. It's called a 277 claim acknowledgment or the health care claim acknowledgment. Those were the three types of acknowledgments that we focus on. Next slide.
Now, there is a question about so can we go, you know, assume all HHS has been using through authority given to it by the Affordable Care Act entering final rule, a survey expedite rule-making process for all the things that we have seen in the last year basically and that we will be seeing in the coming months with some of the things like the health plan identifier. Well, maybe that one might not be a good example. The EST transaction or the operating rules for eligibility and claim status. Those aren't going to come out as entering final rule because we want to a survey of options of those very quickly.
But, with acknowledgments we cannot see, HHS cannot use that dropped in directly to writing and entering a final rule and publishing it. They'd have to go through the full rule-making process. And, that implies a couple things. They have to write a notice of proposed rule so people can comment on, and then, they have to go back and review the comments and write a final rule. That's the first step.
And then the second step, because in the HIPAA law in 1996, it said that the Secretary must give the industry two years to adopt a new standard. And, this is a new transaction, new standard. Then, this acknowledgment will have to go through that as well.
So, as soon as the final rule on acknowledgments is published, two years after that is when we all will be required to comply with it. So, those are some of the aspects of why can't HHS just adopt acknowledgment standards right now. So, we're going through that process right now.
We are also going to be looking at, and we're recommending in the letter, that the Secretary look at ways to assimilate that process. Examples would be like, you know, there's a lot of consensus in industry about this. So, we might not need 60 day commentary, or we might need only 30 day for the comment on the NPRM. And then, the comment might not be that extensive, and so we can accelerate the process in other ways. So, that's the expectation that we will go there. Mark?
DR. HORNBROOK: Just a point of clarification, Walter. This has nothing to do with the handshake required for electronic transmissions. This actually is talking about business practices of saying denying all claims to keep my cash flow at zero while everybody else suffers. Is that what we're really trying to get at?
DR. SUAREZ: No, this is not about that.
DR. HORNBROOK: Is it somewhere in between that?
DR. WARREN: I send you a claim. You send me back electronically, all automated that you received it. Or, you can send me back I received it, I can use part of it but not other parts. Or, I can also send you back an acknowledgment, here are the things I need you to clarify. All of that is electronic. Currently, we do all that through telephone calls and faxes.
DR. HORNBROOK: And, what if I set up a system that tells you for every single claim that it has an error? Just as an automatic response.
DR. SUAREZ: I don't know. I mean, that sounds like a practice that I don't who would use it because, you know, there has to be a reason for rejecting a claim. You cannot just invent a reason.
DR. WARREN: Let me try another way to respond. There are X12 standards for electronic transactions that capture business that do a variety of things. This is another X12 standard. It was not specified in HIPAA that it be mandated.
However, industry has found it to be very useful, and they would like it to be part of this suite of standards. So, we're going back and pulling that together. This comes out of a recommendation from a committee that has been designated to coordinate all of this. And then, they bring their recommendations to us.
So, we're doing a cleanup. So, it's a regular standard. All it does is say, yes, I got it. I mean, that's one of the, that's the TA whatever. There's a 999 that says I've got it, I've peeked in, it looks pretty good, you might have a couple problems.
And then, there's the final one on the billing that says, you know, you need to fix these particular data elements because there's a problem with them. Or, everything is wonderful. The check is in the mail. Or, actually we're hoping the check is assured. (Laughter)
DR. HORNBROOK: But, this has nothing to do with assuring consumers that all our health insurance transactions will be done in 90 days.
DR. WARREN: No.
DR. SUAREZ: No, no.
DR. MIDDLETON: I wonder, if I heard correctly, the Chair of the ASC Program is in the room.
DR. WEIKER: Yes, I am.
DR. MIDDLETON: Would you care to comment? It might be useful.
DR. WEIKER: Oh look, she's already timing me. I haven't even spoken, and she's timing me. (Laughter) No, what Walter and Judy said are absolutely correct. This is not a way for a health plan to deny all the claims. They could do that today. You don't need acknowledgments to do that today.
This is a way of electronically doing, as Judy said, what a lot of time is it's a phone call, it's a go to a website and look something out, it's a proprietary report. And for a lot of the providers, in the health plans they deal with, you know, they could get 20 to 30 proprietary reports that somebody's got to look at and understand and then go back and say, oh, I forgot to put this procedure code here. Now, I've got to go to, it just takes out a lot of the manual effort.
And, these transactions, the TA1, the 999, and the 277CA are only applicable to the ASC X12 transactions adopted under HIPAA. I wanted to make that point. These are not related to any of the NCPT transactions adopted under HIPAA. So, I wanted to make that point as well.
DR. SUAREZ: The pharmacy transactions.
DR. WEIKER: Yes, yes.
DR. MIDDLETON: Does an endorsement of this recommendation require any regulatory action? Or, is this sufficient?
DR. WARREN: No, CMS will write regulations based on our letter, what our letter will say.
DR. SUAREZ: That is what this slide is all about. It's basically that, yes, we will, NCHS will make the recommendation to the Secretary to adopt the standard. CMS writes an NPRN justified with all the other HIPAA documents.
So, the next slide is a highlight of the acknowledgments hearing. And again, very briefly to summarize, there was strong consensus underneath to adopt the standards, the strong consensus on which transactions, acknowledgment transactions, strong consensus on which standards of X12. There was some question about which version.
And then, there's some urgency of the interest of trying to have this in place very quickly because already there is a lot of benefits on usage in it, and we want to make sure that the standard is what is being used. So, I think this is all going to be reflected in the letter.
Next slide is about the key recommendations. And again, we're adjusting some of those based on the discussion from this morning. But, the read out basically is that we would recommend to adopt the three recommended acknowledgment transactions to become HIPAA transactions. So, that's the first step is saying the Secretary should adopt the following three transactions as HIPAA transactions.
The second one is adopt a standard. So, the recommendation is to adopt the X12 standard. And, you see there we say version 5010. We're going to rewrite the letter more in this way of saying that the version that is appropriate at the time of implementation which might be 6020. Who knows, depending on where we are at the time of implementation.
Then, we are also saying that adopt this transaction standard using an expedited notice of proposal making process, so really accelerate the process. We're also pointing out the exclusion of the pharmacy transactions that are conducted in real-time. So, we have those kinds of recommendations.
And then, the next one is the very important one which is making clear the conditions for which an acknowledgment transaction must be submitted. So for example, the TA only would be required when the submitter has elected to receive the TA. So I send a transaction, and there is a place in that transaction that I say, oh please, there's sort of the acknowledge receipt when you send an e-mail perhaps.
Or, it's the acknowledge receipt, I mark it up and say, please send me back a TA1. If you don't mark it up, you don't want a TA1, there's no implication that you will be still required to send it back or something like that. The same with the 999 and the 277CA. The expectations of that will be a very defined set of circumstances in which you will be expected to submit those.
The next one, the next steps that we are finalizing as I mentioned, with the Subcommittee, the recommendations we're going to be preparing the letter of the recommendations over the next month and a half, present that to the Executive Committee, Subcommittee. And then, present it to the Full Subcommittee in September. And before the September meeting, we'll send it to the Full Committee for their review.
And, we hope that we will have it approved by then. And, that is my report. That's it. I'll turn it back.
DR. CARR: Thank you, Walter. Outstanding work. Really, the Standards Committee has been incredibly vigorous and prolific in their work, and thank you.
So, I think right now, we're going to break for lunch. We're due back here at 12:45 for a presentation. Thank you.
(Whereupon, a luncheon recess was taken.)
Agenda Item: All-Payer Claims Database (APCDs) Department's Initiative
DR. CARR: I would like to convene the afternoon session of day two of the National Committee on Vital and Health Statistics. Do we have anybody on the phone? No. Okay. A little bit of housekeeping. We're not going to address PCAST any further today. We will do that offline and keep everyone in the loop.
So today, we're very happy to welcome Andre Chappel and Denise Love to hear about the All-Payer Claims Database. So, thank you for coming. And, we look forward to your presentation.
MR. CHAPPEL: Thank you for inviting us to come and speak today. I know I'm really looking forward to receiving some of the advice. I've got my notebook out already because I'm sure you're going to have some great input on what we're trying to do.
So, the genesis for this project came out of funding from the American Recovery and Reinvestment Act of 2009 which provided $1.1 billion to improve comparative effectiveness research. $400 million of this was allocated to the Office of the Secretary for infrastructure investment. And, the FCC for CER advised the Secretary on how to allocate these funds.
One of the projects that came out of these discussions was the Multi-Payer Claims Database. The objective for the project is to build a Multi-Payer Claims Database to support comparative effectiveness research on a foundation of public and private payer claims data. There's a number of objectives that we have, goals that we have.
Probably our highest goal for sure would be the privacy and protection of the patient data that's included within the database. However, we also would like to build a comprehensive database that can be used to look at conditions that are rare, subpopulations, different geographic areas, just to broaden the amount of comparative effectiveness research that can be done over and above what can be done today with single payer databases.
And, we are engaging the private sector in this effort. We view it as a public, private partnership. In the future, I'll talk a little bit later towards the end of the presentation about some of the future enhancements that we envision for the database.
Again, this is just covering some of the value-added that we see from a Multi-Payer Claims Database including making it easier for comparative effectiveness researchers to be able to identify and acquire the data necessary to produce comparative effectiveness research. We also are looking at expanding geographic coverage and density over what's available today and studying the less common conditions and potentially linking the data across payers and across time to allow for longitudinal profiles.
And, this slide just is a brief overview over the various pieces of the project as its developed. In Phase 1, we contracted with Avalere Health who provided a strategic framework for how to build the database through contacting multiple stakeholders who would be involved in data, who are involved in data collection and private payers.
We had two design pilots thus far that have looked at different ways to combine data as well as building a user interface which would be used to be able to access, to select data elements and access the necessary data. In this case, it was performed as an analysis tool. And, these projects are either complete or nearing completion in the next, and then in the next few months, we will be testing both of these sources.
And for Phase 2, which is really the major portion of the project, is the implementation of the MPCD. And, this was awarded to Ingenix which is now called OptumInsight. And, this started rapidly in January, and we've been making pretty good progress so far.
And again, this is a collaborative effort between ASPE, CMS, and Ingenix in terms of implementing and building the database. However, we're also consulting with a number of other leaders in the health care sector as well as states with All-Payer Claims Databases.
And, the databases we currently envision would have two separate components. It would be a hybrid design with both a centralized data repository as well as a distributed data network. And, I'll describe the distributed data network in a little bit. There's a few different variations on this, and this one's a little bit unique.
And, the reason for having hybrid design is feedback from potential data partners about how they would prefer to participate in the database. So, this allows some flexibility. And, initially the two primary sources of data for the MPCD would be data from the Chronic Condition Warehouse from CMS and data from Ingenix, their Normative Health Information Database which is a national database of different payers, private payers.
And, we also for Phase 1 are looking at including an additional two partners which might include all payer states as well as the data, they call them data aggregators or insurance plans. We can't, unfortunately right now, list who these are because we're still in the negotiation processes with them.
In terms of our advisory structure, we have a number of different entities that are helping to guide the projects. We have a governance board which is being recruited by, which has been recruited by AcademyHealth. And, David Helms is the Chair of the Governance Board.
We've had one meeting with the Governance Board which was terrific, actually got a lot of great feedback on the project. And, this is a little bit different than a typical governance board in that it's providing advice to the contractor so their recommendations are not binding.
We also include a, we also have an HHS Leadership Council which includes various people who are involved with data release at CMS as well as Todd Park and others at HHS that would be influential in terms of providing guidance on various implementation issues. And, these are the full list of members of the Governance Board and probably a few familiar names on there to you including Jim Scanlon.
And, this is the membership on the HHS Leadership Council. So, it's mainly composed of ASPE, ONC, CMS, but we also have membership from AHRQ and FDA.
And, other entities that will be providing guidance on the project are what we're terming a Data Stewardship Board. This Board will be made up of contributors from the various data partners within the distributed network.
And, this Board will be used as, in a couple of different ways. For one, it's going to be providing guidance on how the data distribution process will take place. And, it will be an arena for data partners to voice concerns that they may have. It will also be providing a first-level check on data requests that come into the MPCD to make sure that they meet the basic criteria and that the researchers meet the basic qualifications for making a data request.
And finally here's, we have a HHS Expert Panel. The Expert Panel will be testing the database. We have a total of 100 queries in the contract that HHS can make to both test and provide feedback on the usability of the database. And, this testing will begin in February of 2012.
In terms of privacy and security, we have a number of different efforts to make sure that the database is private and secure but also usable. We are including DUAs with all of our data partners. And then with researchers, we'll be signing DUAs with, depending on which file they request, with either the MPCD or the data partners themselves.
We're also including encrypted, we're encrypting the PII for matching across partners. But, all data that would actually be sent, will be transferred across the distribution network will be de-identified before this happens. And, I can discuss that a little bit in further detail if you'd like.
We're making the database compliant with FISMA, HIPAA, ARRA provisions, and CMS IT policies. We've been working very closely with the CMS team on making sure that their participation in the database is suitable to their needs and to the security and protection provisions that they generally provide. We're also looking at various ways to protect against linkage attacks through masking small cells and changing dates of service as well as providing a statistical review of various other ways to de-identify, make sure that the data cannot be inferentially identified.
And we're including for, these last two bullets are actually, well the second to last bullet is actually more focused on the protection for data partners in the network. And, this is when the data will be distributed, there will be an auditing trail so that data partners will be able to identify what's happened with their data at every step along of the process so that they can feel that they will be able to identify if there's anything suspicious that's happened to their data along the way.
And, we're also looking at testing various data enclave models. There are a number of these models available. We're working with CMS because they're going to be testing some data enclave approaches to their own data release strategies as well. And, we want to be able to use those testings in collaboration so we don't have overlapping efforts.
Various data enclave models would be probably the most restrictive as a researcher going to a research data, a research access center and using a computer that's been secured so that they can't copy any data back onto a USB drive or any other means. But, there are also other approaches in terms of sending technology like a laptop that's been secured that doesn't have any means of copying data off of it as well to researchers and then retrieving that laptop when the study is then completed.
In terms of the technical design for the project, we've been learning from some of the insights that CMS has had in their project on de-identification. We are looking at both, as I mentioned, a distributed versus a centralized database for different partners depending on what their needs are.
In terms of what we mean by a distributed data network, it's a little bit different than some of the other models that are currently being used, such as, the FDA Sentinel Model. In those models, the data is actually never transferred from the contributor's site where they store the data. Instead, researchers would be sending, in this case it would be the FDA looking at medical events, would be sending SAS query analyses to the data contributors that they would run on their own databases and then just send the results back.
In our case, we've received a lot of feedback from comparative effectiveness researchers that they would not feel comfortable with the data contributor for transparency reasons performing all of the analysis on the data themselves, that they would much prefer to be able to have access to the data, especially for academic purposes.
In addition, it's a little bit different in the need for comparative effectiveness research. Part of the, one of the selling points of the MPCD is that we'd be able to combine data to increase statistical power. And finally, given the volume of data requests that we'd expect to receive through the MPCD, generally research is an iterative process, and it would be very time consuming for data partners if they had to constantly be re-doing analyses as a project developed.
In terms of our data access model, we have three tiers of data access. The first tier would be providing a public use file which would be similar to if you go to MAPS. It's just aggregated forms of data that you'd be able to generate. There would be no person-level data available at that tier.
For tier 2, we'd be looking at a standard analytic file. And, this would be a file that data partners would agree to provide their data for under a less intense review so that the MPCD could approve these data requests without it going all the way back to the data distributor, the original data distributor to reduce the burden on the system and also to sort of reduce the burden on the system. And so, at this level, it would be more of a limited data set. It wouldn't be the full list of data elements, and it would be only data that the contributors are willing to provide.
In tier 3, we're looking at the full claims files. And so in the state, in tier 2, it would be something like a five to ten percent extract of data. In tier 3, it would be the full claims that the data contributors would be willing to provide as long as they are able to review requests.
In terms of the data sources, as I mentioned, we are going to include the Chronic Condition Warehouse and the Normative Health Information Database from Ingenix. And so, I've discussed most of this, and so I'll just skip over this slide.
And moving into later stages of the project, we are going to be looking at where we have gaps in data and pursuing other states that are in the process of, currently in the process of building All-Payer Claims Databases and those that will be building them in the future, as well as, including other insurers and other entities that are interested in participating but need some more time to consider how they would like to participate.
These are some of the criteria that we use to evaluate the initial partners that will be included in the MPCD. We're looking at what kinds of restrictions on use that they would require for their data.
We're also looking at the quantity and coverage of the data that they provide on a national basis as well as the quality and format of the data and the cost for their participation. So, we include a number of incentives for the MPCD, which I'll be discussing in a moment, and how much their data would be overlapping with data being received from other partners.
These are some of the incentives that we're looking at pursuing with different data partners. Each data partner, depending on the type of partner, has a very different criteria of what they would like to see by participating in the database. For instance, a lot of the states would need some kind of funding, subsidization for setting up their link into the MPCD. Some of them are also looking at, are interested in receiving data from CMS, Medicare data that they don't already have in their databases.
We're also looking at, when I say royalties and revenue sharing, so this is a little bit similar to what you might see with HCUP where they have the state and patient databases, and each of the states retain ownership of those databases. So, they also set up a price for researchers to obtain their data.
And, some of the other data partners have expressed that really what they would like is to be able to be given a certain number of free queries into the database. And there's a few other, such as, public recognition, a few other incentives as well.
We're working with a number of key partners. As I mentioned, AcademyHealth has been supporting us with, in terms of recruiting for a Governance Board and also providing outreach to the research community. We're working with NAHDO on facilitating access to the state All-Payer Claims Databases as well as being a partner in helping us to sort out some of the technical issues and other implementation issues in building the database since their partner states have already been through much of the process.
Buccaneer is where the data is being hosted. So, this is the same database manager that manages the CMS data warehouse. And, the University of Washington has been contracted to both provide testing on the database after HHS has used the database as well as to provide information back on the usability of the database and provide some demonstrations in the literature of how the database can be used.
We have a number of outreach activities including presentations at various organizations. We just had a panel session at the recent AcademyHealth ARM meeting. We've been going out and presenting at various committees, such as, the NCVHS. And, we're looking to publish in the literature on different applications for the database.
Some of our future plans for the database are to look at including episode groupers, risk adjustment, various quality metrics, as well as, as was discussed a little bit earlier today some of, in the future, claims data may include lab values as well as some more information coming from electronic medical records. This isn't relevant to the implementation phase of the database, but it's something that we're preparing for nonetheless, the contractors including fields for these different sources of data that we might receive in the future so we're prepared if these sources become readily available and feasible to include in the database.
And, this is the rest of the ASPE team. And, I just want to thank you very much for allowing me to come here and present here today and for all the feedback that I'm sure we'll be receiving shortly. (Laughter)
DR. CARR: Okay. Let's start with Mark.
DR. HORNBROOK: Has the FDA expressed any interest in using these data for the Mini-Sentinel Surveillance?
MR. CHAPPEL: Not at the moment. We are working with the FDA at the moment to look at how these two efforts might be able to collaborate in the future. But currently, our database would be used mainly for comparative effectiveness research whereas theirs is to look at identifying medical events. Now, there's overlap in how that is seen. But, the FDA at this point has wanted to keep a certain barrier between the two projects for their own legislative purposes that they have built the Mini-Sentinel project for.
DR. HORNBROOK: Okay. It's interesting because the Mini-Sentinel is not research. It's surveillance, so there's no IRB involvement in the Mini-Sentinel. So, that's one way to get these data into a whole different pathway for public utility as opposed to just going down the research pathway which is good too.
DR. CARR: Walter?
DR. SUAREZ: Yes, thank you. I have a couple of very quick questions. The first one is about the quote unquote owner of the data. So, there's ASPE, there's CMS, there's AHRQ. Who owns the data?
MR. CHAPPEL: Right. So, that is a very good question and that actually has been something that's been evolving over time as our understanding has been evolving. So, we've been working with CMS, and we're still in the works with this.
But, it looks like it will make most sense for the ownership of their data be transferred to ASPE and, of course, Ingenix would be managing their database through Buccaneer. But like I said, I don't want to make a definitive statement about that yet because we're still in the middle of discussing how that might make the most sense.
In terms of some of the other data partners, most of the other data partners will want to retain ownership of their own data. And, they'll be able to do so. And at any point, they'll be able to halt the process of processing their data if they don't feel comfortable in participating in the database any longer.
So, in terms of the project, ownership of the infrastructure will reside with ASPE. However, the data will reside with the data owners unless they would like to, unless they find it would be easier for them to transfer ownership to the MPCD.
DR. SUAREZ: And, the data owners will have some direct oversight over what happens to the data?
MR. CHAPPEL: Yes, through the data stewardship.
DR. SUAREZ: The second question is, I'm a consumer, and I am in some health plan enrollment file. And now, my data somehow ended up in Ingenix database, right? And, now Ingenix is sharing that data with ASPE who is going to be making the data available to researchers.
And I've gone in, this is important because this is the important part of really being able to, from a consumer privacy relief perspective, understand how did I get to that place when I started here and it went through all these steps because in reality, you know, and I know HIPAA created all these opportunities too and also some restrictions on if you have data, you are the core entity. And, if you are going to share that data, you have to have a reason to share without consent. And, nowadays consent is expanded or expected more and more. And, there's data use agreements and all these things.
So, if you follow a patient that is enrolled in a health plan that somehow end up participating in an All-Claims Database or in Ingenix for some reason, the health plan data can come to there, and Ingenix or APCD participation in this database, and then this data gets to the researcher, is there a clear mapping of that process from a privacy perspective?
MR. CHAPPEL: Right, in terms of how a mapping in --
DR. SUAREZ: Well, mapping in the sense of helping everybody understand, you know, the source, the origin of the data and the process by which the data ended up from the original point into point Z.
MR. CHAPPEL: Yes, the whole process will be transparent and will be actually sharing information on how that, not for any particular plan per se, but how in general the data will be, how it will be originating and then being transferred through the system so everybody understands how that will operate.
DR. SUAREZ: And then, consumers at this point would not have any opt out possibilities? Or, would they?
MR. CHAPPEL: So, that's something that we're still in the process of figuring out because consumers, as it is right now, they have some rights in how their data is shared at a plan level. So, this is something that we're kind of working through right now and figuring out what that will look like. That's a great question.
DR. SUAREZ: Okay, and just my last very quick question. I think on your slide number, page thirteen, the data access model, the three tiers of data access, tier one is a public usage file which is aggregated, I suppose, unidentified data I presume.
MR. CHAPPEL: Yes.
DR. SUAREZ: So, it's totally de-identified data.
MR. CHAPPEL: Yes.
DR. SUAREZ: And then, tiers two and three are the limited data set, are much richer. It wasn't there, but I assume that for tier two and tier three releases, you are requiring a DUA.
MR. CHAPPEL: For tier two and three, yes.
DR. SUAREZ: There is an expectation that there will be a DUA?
MR. CHAPPEL: Yes, there is. Yes. Absolutely.
DR. CARR: Okay. Bruce?
DR. COHEN: Yes, fascinating, very interesting. I know your focus is on comparative effectiveness research. But, these All-Payer Claims Databases can make a substantial impact on evaluating population health.
So, I would strongly encourage the team in ASPE to work with University of Washington or other researchers thinking about the potential use for surveillance of population health and research on population health issues. I think this is, in general, is a theme around APCDs that might not have been there initial focus but can add enormous value to public health.
MR. CHAPPEL: Certainly, and I think Denise will be able to talk a little bit about what the states, what they're using their own data for in just a moment. But I think, you know, in terms of the initial implementation phase we're looking at just comparative effectiveness research, but it's yet to be determined what will happen with the database moving into the future.
DR. COHEN: I think it's important to begin the planning now especially around the content and potential use particularly for access for tier two and tier three and implications for using to evaluate population-based morbidity and community health surveillance.
MR. CHAPPEL: Okay.
DR. COHEN: My second thought is, will tier three data be linkable to other databases?
MR. CHAPPEL: Ah, that's an interesting question. And, as we're looking at it right now, you mean outside of the NPCD I'm guessing?
DR. COHEN: Yes.
MR. CHAPPEL: I think this is something that we're also working through as well, but it's not, I have to say, it's not something that I've given a lot of thought to up to this point. I was mostly thinking in terms of the NPCD itself. But, I think that it –-
DR. COHEN: Well again, subject to the privacy issues and security issues, it's got enormous potential value, for instance, linking it to NDI ultimately or some surveillance systems.
MR. CHAPPEL: Oh, well we're actually looking at bringing in some of the NDI data as well.
DR. COHEN: And again, this has huge implications for data ownership and access, but it's got enormous potential value for, you know, the first thought I had about population health.
MR. CHAPPEL: Sure.
DR. COHEN: To begin linking the claims data within AHRQ system surveillance systems –-
MR. CHAPPEL: I was thinking in a little different terms, but yes, we are actually. We're actually looking into the NDI data right now.
MS. GREENBERG: Everyone may not know that that's a national database.
DR. CARR: That would be a good outcome measure.
PARTICIPANT: We all agree on that outcome?
DR. COHEN: Yes, but ongoing surveillance databases too for the states that have APCDs linking them to, you know, their cancer registries or also their vital statistics. There's enormous potential. And, I'd like to see that, you know, when this is further down the road, addressed in the details.
MR. CHAPPEL: That's a really good recommendation. I've written it down, and we're going to look into that.
DR. FITZMAURICE: Thanks for a nice presentation and a really good project. I can't wait to see the lessons learned at the end of it.
States can get the private insurance claims data, the Medicaid data, the scant people who don't pay claims, the uninsured I'm not sure what the states will do about that, and it's difficult and impossible for them to get the Medicare data. You know, maybe only one state has gotten it so far.
PARTICIPANT: Four.
DR. FITZMAURICE: Four states? Great. That's a good increase. My question is, do you have any thoughts of supporting states in collecting their own APCDs and being able to submit research and policy questions for them to answer?
There can be a federated approach where maybe we give them the Medicare data to link to all of this, they then can also pull from the hospital discharge data to fill in some variables for those who don't pay claims to get a really robust database. The states who are considering ACOs are very likely to need just that kind of a database in order to assess risk and to assess where in the population they need to have better provision of care.
So, have you thought about a federated approach where it's out there, but you have the ability to ask questions? And, if the states don't have the resources and don't want to do this, then you have a contractor place where they can come and store their data like a cloud and still control the use of their data.
MR. CHAPPEL: Right. So, and part of that is the centralized versus distributed network and how that will operate. We don't have specific funding in the contract to build state All-Patient, you know, to help fund the building of state All-Patient Claims Databases per se.
But we do, if they're able to build them already have some funding for them in terms of their participation in the NPCD. In terms of the sharing CMS data and other sources of data, this is an ongoing discussion that we're having at the moment. And you know, it will depend on what CMS is comfortable with at the end of the day. But, there is, I hope to be able to do that to some extent.
DR. CARR: Len?
DR. FITZMAURICE: At the end of eight years from now, do you envision that this data will be inside of ASPE?
MR. CHAPPEL: Inside of ASPE in terms of --
DR. FITZMAURICE: Maybe a private contractor, but under the control of ASPE as opposed to being the federated approach.
MR. CHAPPEL: I can't say at the moment. We're still working on looking at assisting ability plan. Unfortunately, I can't provide more information now.
DR. NICHOLS: So, I am really curious. And, I applaud the effort. But, I guess I have two kind of big questions, and maybe both are stupid. So, I'll start with the one I think is the dumbest first.
This seems like a really odd way to do comparative effectiveness research. We've got claims. And, the last time I heard, docs are pretty nervous about using claims for anything like this. So, I'll just let you deal with that how you want.
But the second thing is, all versus multi. I think I understand all because all is about finding all in a region so you can actually say something fairly definitive about spending patterns and use patterns, use variation to guide all sorts of things.
With multi, as big as it is and exciting as it is and interesting as it is, I don't know what you've got. Here you have a representative sample of what? And if not, how do you fix that?
MR. CHAPPEL: So, we certainly won't have a representative sample at the national level in, you know, the early phases of the project. However, I think it's a matter of not letting the good be the enemy of the better. At the moment, I think there is a lot of use for claims databases, for comparative effectiveness research, especially in like hypothesis generation. And moving forward, we are looking at including some of those other types of databases.
So, it's sort of a first step in a direction of building a more comprehensive database. For some of the states, we will have, hopefully we'll have All-Payer Claims Databases to get started. But, it's sort of an incremental process as I see it.
DR. CARR: Leslie?
DR. FRANCIS: This is a question about the relationship between what you're doing and the role of our Committee. Particularly, the interest I have is in privacy, confidentiality, and security because that's the Subcommittee I Co-Chair.
You're getting data from different HIPAA categories in a way. Some of it will have come in from the public health exception piece of HIPAA, some of it may be data at least it sounded to me as you were describing the database that didn't originally go through state public health but perhaps came some other way.
Then also, you're getting data from very different states or data sources that may have different privacy concerns. And so, my question really is, this is a Federal advisory committee that can write letters that make recommendations about good policy. Are there places where you're experiencing tensions, either with respect to HIPAA category or with respect to data source, that we can really help you with?
MR. CHAPPEL: I think this is a good point. It was actually a discussion at lunch today. Absolutely. I think the more standardization over the process to the extent possible. States are always going to have their own standards as well, but to the extent that we can push greater uniformity of standards, it will certainly help the process.
In terms of the MPCD, we'll have a certain level of standards that will be established within the MPCD. And then, if states have additional standards, we'll be incorporating them into their data use agreements. So, does that --
DR. FRANCIS: Well so, I mean my question really is could we be of help in, for example, considering recommendations about what should be in data use agreements? Or, could we be of help with respect to the question of how to think about data enclaves? Or, could we be of help with respect to the problem of de-identification?
Those are just three issues that seem to me to surface in your presentation. But, I'd like specifically to say to you I know Mia, who is right here next to me, is at ASPE and she's lead staff to the Privacy Subcommittee. So, it could make a great deal of sense to surface some things that you'd like us to work on.
DR. CARR: Okay. Let me just do a little time check. I have Larry, Mark, Vickie, and Marjorie. But I, Denise do you have a, you want to do a presentation as well. I'm going to ask that we just hold those questions until Denise has had a chance to do her presentation.
Agenda Item: National Association of Health Data Organizations
MS. LOVE: I have slides and quite a few of them, but I won't present them all because we don't have time. And, I'll make them available.
So, thank you for this opportunity to be here today to talk about a new data system in states and one that's emerging fairly quickly. I wanted to acknowledge that the All-Payer Claims Database movement in states that began in the Northeast is expanding rapidly across other states. And, we do this work as a team.
And, I'd like some of my Board members and data stewards of the All-Payer Claims Databases and team members, we've got Patrick Miller, University of New Hampshire, we've got Al Prysunka from the Maine Health Data Organization who kind of started all of this in the states, we've got Keely Cofrin Allen from Utah, Craig Schneider from Massachusetts Health Data Consortium, and we've got Brad Finnegan from National Governors Association, and Emily Sullivan from NAHDO. And, there's others who aren't here today.
And, we really do pull as a team to get traction in the states to help them implement their data systems. So, much of the content that we produce, the papers and the resources for states and the latest information we found at the APCDCouncil.org site, and this is a site that's maintained at the University of New Hampshire. Some of this is funded through the Commonwealth Fund in part and others to put up information as states develop it about the latest and greatest in all-payer claims.
What are APCDs for the purposes of my talk? Well, these are large-scale databases typically created by state mandate. And, I'll talk about that in a few. It includes the claims data from the medical, pharmacy, and often dental along with eligibility and provider files from both public and private payers in a state. This includes the carriers and PBMs and TPAs in some states. Public payers, Medicare, and Medicaid are, if not in the database, they're an active planned part of the state data system.
We want to emphasize that these All-Payer Claims Databases are enhancing and supplementing but not replacing hospital discharge data, Medicare, registries, and other data assets in the state. That we've had a few states say, well, then we can let our hospital district database go, right? Wrong, because they really work in synergy and tandem together.
The databases, a big driver is transparency. And, I won't go through all the questions and uses because we really, frankly, don't have time. But, we can come back to some of these.
So, what is in a state All-Payer Claims Database? Well, the outer circle represents where states start. And, that's the commercial and Medicare Parts C and D, the claims.
And so, a state will typically start with the large payers defining the threshold in that state and begin with the commercial claims. Once they aggregate and solve the collection of the commercial claims, the next stage in an APCD evolution is to add Medicaid Fee-For-Service and Managed Care and SCHIP in states where that is available and normalize that to the commercial format.
We talked about Medicare Parts A and B, that's down the road. We have four states who are gathering that data, and that is not an easy process nor is it timely. But, states need Medicare data because this whole concept is to build a population-based dataset to look at the health care delivery performance and the health of the state's population.
The innermost circle is the uninsured, TRICARE, Indian Health Service, and FEHB. We have not solve that, and we go into this knowing that that's not going to be part of the initial mix. We do have a pilot in Maine that one of the insurer's had issued an ID card for the uninsured and runs that through the system. So, it is technically possible, but we are not there yet. We start, again, with the outermost circle of commercial and move inward.
Typically, I won't go through each data element, but it's what's on a claim. State laws differ in what identifiers they collect, and we can go back to that because that's another ask. I think at the end, we can talk about what states need to make the All-Payer Claims Databases work. Many of the states are getting encrypted or de-identified social security along with the other patient demographics.
Excluded, we also cannot capture everything in these databases initially. And so, an addition to the uninsured typically the denied claims, workers' compensation, referrals, lab test results, provider affiliation, premium information, capitation fees, administrative fees, and such are missing.
That doesn't mean we can't go back in the future and backfill. There is some thinking about these pieces of information could find their way into an All-Payer Claims Database. But again, that's for future work.
So, the early APCDs began in the early 2000s. And since that time, we've seen a proliferation of initiatives that make the APCD more urgent in many states. Again, transparency, coalitions, payment reform, HITECH, and health care reform. And so, states are seeking a dashboard by which they can manage some of these new initiatives. And, APCD is one that's based on available data, and it's a feasible data system to build in most states.
So, where are we? I think we're here. It changes almost daily. But, we're filling in the gap. The darker states are states with active implementations and existing databases. The medium-dark states are states that have laws and have early implementation but have not fully ramped up their data systems. And then, the light blue states or the lighter blue states are states that we have been on the ground either helping form some stakeholder activity or early planning legislation to discuss the possibilities.
Now, I want to say typically in this map we have state mandates. So, the states are collecting the data under a Health Data Authority Act or a Public Health Authority. We do have two exceptions: the light blue states, Wisconsin and Washington. And, these are employer coalitions that started with commercial claims and aggregating voluntarily and are trying to go the other way where they add some of the public data and some of the other payers going forward.
There is a different challenge to the voluntary. I mean, it's not as onerous, it's maybe a little more flexible. But, they can't compel a reporting. And so, there may be some gaps too like paid them out, the adjudicated claim because sometimes, I mean that's a pretty sensitive data element for payers. And so, on the voluntary initiative, that may be a hard nut to crack. So, we're watching these voluntary initiatives, and again, they are providing useful information.
But, where the APCD Council and NAHDO spend a great deal of their time are helping states put the data policies, the regulations, and the laws together for an APCD. Well, we're involved with a whole range of activities including USHIK, national standards, issue briefs, working with the National Governors Association and others to document some of the lessons learned and disseminate that information throughout the state data world.
I just want to acknowledge our Technical Advisory Panel to help develop core standards that was funded by the Agency for Healthcare Research and Quality. And, I think Cynthia is in the room from AHIP. And, we've spent quite a bit of time getting the core data elements that states collect and recommending a core that states can use as a starter collection set.
This is important to national payers because, as you can imagine, if you have a different state format in every state that's bringing up an APCD, your workload expands geometrically as they add data elements. So, we're really trying to think about the burden on the payers, educate the states to adopt a core standard, and then work collectively to update that standard and enhance it going forward. And, we can talk more about that if you'd like, but I really feel we'd like to get to dialog.
Now these databases, as we've heard have something, I think, for everyone. And, it's kind of exciting. It's not just the data geeks. We really can get at consumer pricing and information.
At the end of the day, quality is great. But, with more out-of-pocket, more burdens on the households, we really need tools so that some of these elective and shoppable conditions can be put out there for consumers. And, they can look at the variation within what they have to pay in their mix.
Employers are using it to benchmark. Employers have their databases and have long been advocates for the use of claims data, but you get to pretty small numbers pretty quick when you look at the physician or procedure level. So, this large aggregation really gives some useful benchmarks.
Health plans, payers, again, larger benchmarks. Providers, researchers are eager for this. State government and policymakers can really get a dashboard to look at what's happening as they start making these cuts and, you know, the impact on the populations. And, I could go on and on.
We haven't really decided how, I mean I think this new initiative is intriguing. We'll learn what that Federal engagement has been. Up to this point, it's really been a state-driven initiative.
And, I have a whole host of use cases that I could share. They're looking at, and especially New Hampshire is a power user. And, Patrick is back there to answer any of those questions. But, looking at variation in outpatient services and patient migration, prevalence of populations, Medicaid to the commercial, and again, looking at quality and per member per month pricing.
And, episodes of care are a big deal, a really big deal to legislators who aren't typically data geeks, but they get it. They get that we can tell them what a total hip costs in a hospital and what the, not even costs, but what the variation is.
But, you start putting the care 30 days out, 60 days out, the whole episode, and with or without these scans and other ancillary services, and you start really thinking about pricing and how benefits packages can be structured and evaluated. Medicaid benchmarking, and I don't want to belabor all of this, but we can come back. And, some of the experts are sitting back there. And so, I want to be very mindful of the time.
This was one of my favorite slides in New Hampshire, and it's really looking at patient migration for inpatient and outpatient services. But interestingly, it was to prioritize. If I made Patrick, you know, where are the HIE nodes should be connected? Where are the patients going in and out of first? Because you can't connect the whole state at once.
So, it gave them a sense of where the migratory patterns were and where is the starting place to start talking to each other? And, I thought that was a huge use of some of these data sets and against prevalence and populations. And, this is significant because this is a Tri-State report.
Because the Northeast states have been at this for a while and have historical data and have used the same format, it's possible to do some comparative analysis across states and start getting a snapshot of a region. And when Massachusetts comes online, you've got a huge regional snapshot here. And, that's pretty exciting. So, you can start doing Tri-State reports in variation on more than Medicare claims.
So, we work with states. A big part of NAHDO's work and the Council's work is helping states with this roadmap that's pretty much been played out in most states. We work up in the front-end to engage the stakeholders, make sure the right people are at the table, it's inclusive, and that the process is transparent.
We guide states on governance models. States vary in their laws and what they're required. But, the governance models, there are some options for states. And, they need to weigh those out in light of what they want their APCD to do and then helping with those rules and collections of policies.
Funding is the number one big roadblock. We could have more of these states in that map, but states are kind of broke right now. And it's to me, it's just this conundrum because at the time they need the data the most to make these decisions, they have no money. It's like they're spending money on a Bentley, but they're just foregoing the dashboard and the speedometer and they just are going to steer this big health care delivery system somewhere without some benchmarks and pre- and post-.
And so, we really are hoping that we can help states cobble together the right kind of funding to begin, even if it's not a perfect solution. We do help with technical build, and Al Prysunka has helped so many states go through how to build an APCD that works. And then, analysis and application, UNH has been a leader in helping states with their reporting and their applications. And I see that as more states bring on APCDs, we're going to see better tools, better benchmarks, and better open source applications.
So transparency, relationships, all the things that a good data system has, the APCDs have to put in place. But, we really want to talk a minute here about linkages. Again, we missed some populations. We don't get everything, but we're working hard to make consistency across the states happen.
I wanted to take a minute, and this is Patrick Miller's slide, and I feel like I shouldn't be talking from it. But, it does depict some of the questions that Bruce brought up. So, the bottom bar with the data release, Data Use Agreements, and the Boards are where states are pretty much now. I mean, you know, the requests come in for de-identified data, they get reviewed given the policies put in place for that state, and those releases happen.
But as we have HIEs develop, and as we have registries that are eager for their own information, for instance, cancer registries would like to really look at an episode of care as well. So, we have these silos of data up there with pretty good data. And so, with the right identifiers, the right design with an APCD, we can start bringing in just-in-time linkages or strategic linkages.
And then, we suggest that perhaps a linkage review board in the states would review the request for the linkage, actually maybe conduct it, and then let the analytic file go back out that fits the criteria legally and technically that the user needs. But, I do think that APCD is just another important database of a whole spectrum of data that states need to manage that.
So given that, we need an APCD, what we call a version 2.0. So, we need complete data, we need standards across the states so that they're capturing the proper unique identifiers. And, this can't be understated.
But, the states have a particular problem. They have people like me that go out and talk to legislators and stakeholders and say that you need identifiers and given their law permits that. But in this era, it's really tough sell to say that the government wants me to address and all of those things that we know are important to public health and population health. So, some messaging or from a national organization or a series of national organizations that help frame that message so that states have some political backup and support for collecting direct patient identifiers.
Again, we need to figure out how to collect premium information. Some of the newer states, really their stakeholders absolutely want that into the mix. I can't say enough about a Master Provider Index, MPI, something that identifies a unique physician because we have a couple states spending a fortune just cleaning up the provider databases.
And then, we need some standardizing and collection of benefit information. We will have a core data set. We hope to in the future have funding enough to go make an implementation guide and take that to X12N.
We have standardized Maine, New Hampshire, Vermont, Minnesota, Tennessee, and Massachusetts. And again, USHIK has been putting that up on the USHIK site and the metadata and listing these files of the states.
And again, what determines, we can't say to a state today how much this is going to cost to bring up the state of Virginia. I can't give you a number, but we can give some considerations. What is your population, and what's the market like? And, what coverage are you willing to start with and then grow to? How many carrier feeds and platforms are involved because each platform is the cost driver?
The more platforms, the more normalization, the more work it takes. And, it takes a lot of work especially at the start-up. Again, that provider database, depending on how you're going to use it, is not trivial. And then, you've got to think through who's going to be the users? What are your laws? And, how are they going to access? And, who are you going to release to?
Now, the analytics, reporting, and applications can be limitless. But we're hoping to, in the next year, to have a starter set of some open source measures that states can generate common benchmarks. How much do they spend on analytics? Or what partnerships, academic or public health, is up to the state and their laws?
I won't talk through funding models per se. One thing that I tell a state up front is, don't count on data sales because you don't have a database yet. You've got to build it, and then you can get data sales. And even then, users and researchers aren't going to foot the cost initially.
We hope that there's public good and public source investment, there is public funding. Because again, the state is, and the population and the citizens are the beneficiary once this thing is put in place. But, we encourage states to look at grants, Medicaid match, HIE, partnerships, whatever they can garner on the grounds, Beacon Community funds.
Unresolved issues, and I'll quit. And then, we have experts in the back for the dialogue. But again, I can't stress enough that it's difficult to get money for standards development. But, it is critical.
So, standards development and maintenance going forward will be huge. As we get to version 2.0, what should that look like? And so, we don't have each state doing its own thing, state fields. We remember those days from hospital discharge data.
We look forward to increased Federal and state collaborative opportunities including CMS and HHS and analytic tool development. I can't underscore how we need the message on direct identifiers, why they're important, why states would need to build these into their data sets.
We're working with states on the ground to get those policies in place. But, my fear is we have a few states that even have direct identifiers. With the new legislative class that comes in, the question could be raised, why are you needing name in these data sets?
So, we have to help those people on the ground who are interacting with legislators to say because the population's health depends on it. But you know, what are those pithy messages that are so critical? And again, we will see an evolution of analytic tool development.
I encourage dialogue, and I encourage you to contact us at NAHDO and Patrick Miller at UNH. We're all at the APCDCouncil.org. And, you can find these slides and others there.
MS. GREENBERG: So we'll have, we have a copy of the slides, right? And then, we'll provide them to everybody.
DR. CARR: So, I want to just start by thanking you. This is tremendous. The work that you're doing is tremendous. And, we're really honored that your whole team is here. That means a lot, and we'll be calling them, especially all three back there. So, thank you.
And, we'll be a little off on our timing, but we do want to take the time now to get our questions in. So, where to begin. I think maybe we're just going to go around the room because I think everybody has a question. Len?
DR. NICHOLS: So Denise, that was great. And, I wanted to ask two questions. It sounds like the first message is national messaging would be helpful. I think we understand that, and that's probably our stock and trade. Although, this is my second day, so I'm not positive about that.
But, the second question is actually kind of more subtle maybe. And that is, I understand there are a number of conversations going on around standardizing these core elements. And probably, your work has spawned a fair bit of that now that I see various names of people I know who have been working on these things. Do you need help with that at the same level? Or, is that kind of taking care of itself?
I mean I just, what I hear, serious people, some of whom are on Andre's list, from real plans have gotten the memo that this is coming. And, they understand quite clearly they want standardization for their own safety and sanity. So, is that process in need of Federal help?
MS. LOVE: Absolutely.
DR. NICHOLS: Or, is that process working?
MS. LOVE: The, and Patrick, I don't know if you want to come up here and join me for a minute, but we cobbled together some funding. And the Agency for Healthcare, Research, and Quality has funded the initial discussions with the NGA, AHIP, and the states to come to agreement or consensus on a common core.
What we don't have is actual support going forward as states want to add data elements or update that core. And, this is Patrick.
MR. MILLER: Good afternoon. Thank you for the opportunity. So, we've been working for over 18 months now on the pharmacy side with NCPDP directly. Medco shepherded us through that process, and we expect that by July of this year possibly as late as August, we'll actually have a workbook out with the standards that's been vetted through the Technical Assistance Panel that was described earlier.
The harder push for us has been with ANSI X12, to be perfectly honest, on the claims and eligibility side. I think there's an issue just internally they have a lot on their plate with 5010, with ICD-10, and other things that are in the works. But, to really get those standards fully formed and into place, again, the drafts are posted on our website, and they've been submitted through the process.
It's going to take external stakeholders from the plans themselves to folks like yourselves to help, I think, push and prod and support positively the standards movement there. We've had phenomenal support directly from a lot of the individual carriers and meeting attendants and promotion, etcetera which has been great and from several of the Federal agencies and groups like NGA and others. But, we can definitely use more on the bandwagon.
DR. FITZMAURICE: Just, maybe three statements. It's been a pleasure, AHRQ supports this. It's been a pleasure to work with Denise and Emily Sullivan, Patrick Miller, Al Prysunka, Amy Costello, Ashley Peters, and the others in the state government who supply the data.
USHIK, AHRQ's metadata registry has data from seven states including the core data set from APCD. So, each state can compare its data with the core data set to see if they define it the same way.
Recently within the past month, X12 has given us the data elements from 5010 which will be mandatory this coming January. So, we can put in as another column that states can compare their own data with let's say the reference terminology, the data that everybody is shooting for. So, this is great work by X12.
DR. SUAREZ: Just real quick, I'm sorry, I don't mean to interrupt you, but you said you have the data from seven states. You don't mean the data. You mean the data elements.
DR. FITZMAURICE: Excuse me. You're absolutely right. We have the metadata: the names, definitions, and how they're represented. You're right. We don't have the data. Andre would like to get the data.
We have been working with CDC, NCHS on a X12 producing and implementation guide for the states to make their data more uniform and to be able to compare it. So, it's a great partnership all the way around with the Federal government and the states and with sister agencies.
DR. BURKE: I am Jack Burke from Harvard Pilgrim Health Care. So, I'm familiar with how we supported the effort in the past. And, to underscore what Len said, I think the future success is going to rely on getting input from more stakeholders, particularly health plans. And, I know AHIP is one of the entities that you sought that from. But, that's going to be a key channel to get other large payers online.
DR. HORNBROOK: Can you follow people across payers, say into Medicare? Across Medicare Fee-For-Service and the Medicare Advantage?
MR. MILLER: Yes. We can. Different states are doing it with different methodologies, both the states that are doing direct identifier collection and those that are not. And so, someone switches from let's say Cigna to Harvard, etcetera, we can track that.
MS. LOVE: So, if you can get a national patient identifier. (Laughter)
DR. FRANCIS: So, same question as for Andre about what might be a roll for this Committee. And in particular, would you be interested in having us work on the question of how to protect privacy appropriately with the use of identifiers and linkage?
MR. MILLER: We would absolutely like to work with you on that. I think that we've made it, we've created models, and there are existing things in some states and to be created in others. I think there's huge opportunity, especially in states like West Virginia who are just coming online, who have also had robust efforts on the HIE for example. Vermont is another one.
I mean, there's a series, Colorado, Utah, and others where there's already the need and desire and some linkage going on, and I think that those are the states to make a test pilot of some of this work. And, we would love to work with you on that.
MS. LOVE: And, I would like to elevate this even a little more across public health data sets. We've been working with the cancer registries, and we cross-walked the hospital discharge data with the cancer registries. And, it surprised both data stewards to find out that we're collecting things kind of similar, the same, but not.
And so, if we're going to do these linkages that are in version 2.0, I think what should a demographic data set look like in public health data? And I mean that broadly, including all-payer claims, hospital registries. And so, it would help all the data sets including cancer and others if we had a common demographic direct identifier sort of suite and if it came with a national message that these are critical to build into your data sets with those standards.
DR. FRANCIS: But, I'd say on the other side of that, how do you make sure people can trust it?
MR. MILLER: Yes, that's absolutely critical. And, the other sort of silo that we had on that slide was the health benefit exchanges because there's been incredible interest from states like California and others in using this data, not just the eligibility files but also the claims data, you know, within the exchanges.
So, that was one of the silos on that 2.0 slide. But, we need to move from silo to integration. And, we need to do it, we're not going to solve the health care problems if we're not able to do that.
The CR question came up earlier about how do you do CR with claims alone? Well, you can't. You have got to be able to link it to the clinical data sets. And so, it's paramount in our view.
MS. LOVE: And, also as we talked at my Subcommittee hearing, you know, some guidance on smart de-identification.
MS. GREENBERG: Well, thank you. This is very exciting. As Mike said, NCHS is my agency, has also been very involved with this. And of course, you're singing many of my songs and many of our songs as a Committee.
So, I appreciate Denise actually bringing this to me and saying, don't you think the National Committee would like to hear about this? And, it was great that you all could come and that brought a team. So this, I would say this is the beginning of a dialogue. And of course, Andre and then Jim said well, you need to hear about what's going on in the Department too. So, I appreciate your coming as well.
And the question I had was actually for you. It's kind of around the question Len asked. But, I think I understand these All-Payer Claims Databases, at least what your goal is. This is just a simple question that I don't understand and that is about the, I know about the CMS Chronic Condition Warehouse.
But, what is this Ingenix? Maybe everybody else knows. But, what is that database? Where did it come from? How did claims data get into it? And, does the tier three actually include direct identifiers?
PARTICIPANT: Two is called the limited data.
MS. GREENBERG: The limited data. But, there's also a tier three which is a complete file. So, what is that? Where did that come from? And, how does it compare to these All-Payer Claims Databases and the content they have? I just don't understand that.
MR. CHAPPEL: So, Ingenix is owned by UnitedHealth Group. So, it's primarily UnitedHealth Group data in this database. There are a few other payers as well that are in the database as well. But it's, for the most part, it's UnitedHealth Group data.
MS. GREENBERG: Okay. So, when it says multi-payer, there aren't a lot of different payers who put their data into this Ingenix.
MR. CHAPPEL: Into the Ingenix, no. It's primarily UnitedHealth Group.
MS. GREENBERG: Okay, thank you. And, does it have direct patient identifiers in it?
MR. CHAPPEL: Yes, they do.
MS. GREENBERG: Okay, okay.
DR. CARR: Walter? Or, you can pass. Do you have a question?
DR. SUAREZ: Oh, yes. (Laughter) I mean I think it's, I have been around data, I have been working on discharge data for many years. And so, what is scary about all this, of course, is the fact that we're now building a lot of databases out there because now we have the data, and we have the data electronically. And, there is a lot of value, of course, and benefit in doing that.
But, I think everything really boils down in my mind to trust and this whole concept of linkages of data and someone like a, you know, you mentioned a linkage review board, one would think of like a privacy board under HIPAA or things like that. But you know, I can't but think of the, all the public reactions that we have heard on other committees like Standards Committee and Policy Committee on all these things from a privacy perspective and the protections of data and the fact that my data is going to be on this database and all this.
I'm really thinking that that needs to be elevated to a point where, you know, it is the most critical aspect of it. There is all these benefits and benefits for population health or public health professionals that are looking to access and link this data, for researchers that want to have a new toy and play with this data, and all these things.
But, it's fundamental that we have a very strong privacy and confidentiality framework and that we can answer the five or six questions that consumers will have all the time about this. Because without that, we are risking, you know, the fact that the data is going to be there and consumers are going to want to know where it is.
DR. CARR: So, I want to jump in just before you answer that because I think that we, you know, we've talked consistently over the last year and before that about a learning health system. And, I think what we're seeing here is that we're now approaching a useable learning health system. We've also talked about stewardship and trust and privacy.
So, I think that it's important to recognize that the process is moving forward and that the pieces are all present, you know, just from Leslie's comments and from what we've heard here that it is a bold, new environment. But, it is the one that we believe is going to bring us to greater, you know, individual and population health.
So I just, I want to be weary of the adjectives that we use because I think this, you know, this is the reality that we're looking at. And, I think we're well-positioned to articulate, understand, and partner so that trust is built, the community is educated. That was one of our topics yesterday.
DR. SUAREZ: Well, I hope that by your statement, you're agreeing with this point.
DR. CARR: I'm agreeing with the point, but I'm saying that we, NCVHS, are on the leading edge of the importance of a learning health system. And, what we're hearing here are the building blocks of a learning health system. And you know, there isn't a meeting that we have or even a subcommittee meeting that we have where we don't talk about the importance of trust, of education, of stewardship, protections, and privacy.
So, I just want to make those comments but without any adjectives. It is, you know, it's a new set that we're going into. And yes, when you link and when you have identifiers, it's risky. But, we have to be, demonstrate our confidence that partnering and working together, we're going to manage through this so that privacy is protected and trust is built.
DR. SUAREZ: Yes, I mean, partnering and working together is important. We need to go beyond that and develop the written, stated framework that describes not just partnering and --
MS. GREENBERG: Let me comment on this too because we, Denise just fortuitously was able to come to our May 12th hearing when we, the second part of the Workshop, the Community Health Data Workshop. And, I think we had a very interesting discussion around these very issues of course.
And, I think at least what we have heard from those two sessions was that trying to fly under the radar so that nobody really exactly knows what you're doing, just so you can do all these good things, is probably not a good strategy. You know, that's what we've done in the past.
I mean, and I don't think we've done, I mean some people have done bad things. But for the most part, I mean public health has a very good record, etcetera. There's some notable exceptions. But, that you don't build trust ultimately down the road and end up being able to have this learning health care system without being pretty transparent and not trying to fly under the radar.
So, I'm not disagreeing with anybody. But, I think we have now several subcommittees, probably the whole Committee wants to help this effort.
MS. LOVE: Well, and there's something that I'm excited, I have been at this game for 20 years. And, I've had these conversations for 20 years.
But, the reason I'm more excited about APCDs is because for the first time, we are building, Patrick and others are building tools that directly benefit the consumer. And, I think overlaid with that trust conversation with the consumer is showing the value that they can shop for a colonoscopy and find out how much it varies and how much they have out-of-pocket a priori is huge. And, that might help that messaging with the consumer.
MS. GREENBERG: Right. Including the consumers obviously by serving them and being very transparent so --
MR. MILLER: I think having the consumers and the employers as customers for the first time is really important. You know, they really haven't been users. They have the discharge data sets, for example.
But I'd also say, I served on the HISPIC project and also on our state's HIE efforts. And, I think the HIE and HISPIC and also the coming benefits exchanges, you know, these issues of privacy and trust have been hashed out to a large degree in public dialogue around some of these efforts, will continue to be hashed out as this rolls forward.
And, I think if we can roll this effort into those dialogues that are already going on, it will only benefit the work that we're doing in terms of the transparency of, you know, models such as the one that's on the screen. Again, just a model.
But I would agree, Walter, that developing the formal framework and documentation is absolutely straightforward and something that we should be doing. But, I think we have folks that have gone before us and some good models to look at and active discussions that we shouldn't try to necessarily re-create if that makes sense.
DR. SUAREZ: And if I may just say, absolutely. I think a lot of the work has been already developed, discussed. There are a lot of principles and things like that. It's just that it's almost like on every implementation that you do, the first things should be, and these are the guiding privacy framework, that shows that you have done that.
MR. MILLER: That's a very helpful comment. Thank you.
DR. CARR: Okay. Judy?
DR. WARREN: So, mine is about one of our favorite dialogues, patient identifiers. And, I noticed that I think both of you said you could follow the patients across health plans, and so aggregate the data as they transition from say UnitedHealthcare into Medicare and whatnot.
So, I would be very interested in a continuing dialogue with both projects with this Committee on the strategies, the problems, and the methods you're using to identify those patients, both to identify them so you can follow the data, to de-identify them, and then you link them back up together again because I think that's one of the big hoops that we have. And then, how do we protect that process? That from my perspective, it's more of the technology on how do we put that all together.
MR. MILLER: And, if you want to, I think our contact information will be on the slide. So, if you'd like to contact me, I do have some information on that process in a little bit more detail that I could send to you as well.
DR. WARREN: Great. Because we may be in a position, well I'm sure we're going to be in the position of writing more recommendation letters and stuff like that to the Secretary. And so, we may be asking you to come back and give a very targeted testimony and stuff, on some of this as well as a dialogue.
MS. LOVE: I just also want to acknowledge that the state laws that many of these are developed under have pretty state-specific --
DR. WARREN: Yes, and we are aware of that, and that's part of the problem.
MS. LOVE: And, we can provide information about those to you as well.
MR. MILLER: All of the state legislation and release rules and collection rules are also on the APCD Council website. So, if you want to look at anybody's legislation or collection, release rules are there. And then, there's links on the map on the front page to every state contact and every state website what they're collecting, what reports they're issuing. It's kind of a one place shopping kind of mart for APCD information.
And, if you can't find something, there's an info@apcdcouncil, and Ashley Peters at ten o'clock at night will respond to you with what you need. She's really great. So, she's not here, I wish she was because she really supports a lot of our backbone efforts.
DR. CARR: So, I want to give the opportunity for a couple other folks to ask their questions. We are at the end of our originally scheduled time. I'd like to, I hope we can go over 15, 20, even 30 minutes to just hear a brief summary of next steps for each Committee. Actually, we've heard, Mark, from your Committee. But, all right. Larry?
DR. GREEN: Justine, it is a very good day when long held visions are seen are seen to becoming reality and whole teams show up, and you can't manage the meeting because everyone wants to talk. It's quite clear that this is a sweet spot.
And, in the spirit of thinking about next steps, that's really where I want to go, I want to make the direct request for information from both Andre and Denise, not that exchange as you just had with Judy. From Denise's slides, starting with APCD 2.0 on, every line on every slide had an implicit policy, law, or regulatory issue in it.
And, I want to ask you guys to produce a list based on experience, not what you think might happen but what's already happened, what you're already grappling with that would help us identify the legal, policy, and regulatory issues that you've already learned about and you now are on the table. And, Sally had the good fortune to leave to go on vacation.
But, as the Co-Chair of the Populations Committee, one of our next steps is we want to recommend that we have a Committee-wide workshop that goes back to the community as a learning health system working on the delta. And, I believe that we should have that request for information seed that workshop.
DR. CARR: Agreed.
DR. CHANDERRAJ: I am really concerned, and I'm going on the sentiments of Len and what Walter was saying. This data, hunger for data, is congenial of the Federal agencies and unfortunately, nobody is paying attention to the provider in the equation. Every emphasis is being paid on the consumer.
But, what's happening as a provider is this data is being used for economic credentialing and affecting adversely some other providers in the market. I think what's happening here is we have taken this field of medicine and converted it into a business and everything is being modeled on statistics rather than the science of it. The science of medicine is being converted to a statistic model, and we're trying to learn from statistics rather than the science of medicine.
DR. CARR: I think you are introducing a larger conversation that we're not going to be able to get into today. So, I'm going to ask Vickie to give her question.
DR. MAYS: One of the things that is not in the data set that's difficult to get at is the issue of race and ethnicity. You knew I was going to ask it, didn't you? (Laughter)
PARTICIPANT: How much time do we have?
DR. MAYS: No, no, no, no, no. Again, it is because, I mean we all know this, so it's not an asking as much as it's a direction towards problem solving. And that is, I think that if we try and link this with the issue of trust, consumers, etcetera, and try to get you out before certain people, certain groups to really talk about the usefulness of this, that then the request, using things like what happened when Edna asked the questions and what happens when certain states like, you know, California is out there asking us all the time and can tell you we have very good results, Massachusetts does, that maybe the Committee can help you build a business case for being able to do this.
But, I think that that has to have the linkage with the community. There are just certain places that you need to go and do this, you know, presentation in order to get the dialogue at a place where those leaders like the NMA, etcetera will really back you and make it a lot easier to get it because we have to have it.
MS. LOVE: The places where that, I am hoping for the EHR someday to have these things reside in that, and we can link to that. It's a really tough sell on the claim on the individual encounter. But, my dream long has been the eligibility, the 834, because as when I enroll in a health plan, they ask me a lot of questions. And, it's not usually a stressful time other than, you know.
But, that is another place where it could live. But, employers are really reluctant apparently to ask that information of their employees. So, it's a larger conversation as well. But, we welcome that messaging as well.
DR. COHEN: This is probably another longer conversation, and it has to do with CMS's reluctance to provide data. Is it substantive? Or, is it administrative? I mean, again --
MR. MILLER: The problem is there's not one CMS. There's not one set of data. There's not one process. I mean, I think that's part of the problem. I mean, there have been states that have successfully acquired through the ResDAC process data. But, the limitations on re-release –-
DR. COHEN: Yes, those make it untenable. The data are not useable.
MR. MILLER: Exactly. You know, the dual eligibility data that we're supposed to have as of April 30th. The form was still, as of last week, not available. And now, we're told there's no retrospective data, only data moving forward should we provide it.
And then, timeliness is another issue. As of September, I should be able to get 2009 data through ResDAC. Well, I've got a six month lag right now on my All-Payer Claims Database state in the commercial and Medicaid. So, there's a whole host of issues.
And, I think what we need to do as states and perhaps with your assistance is really put together the laundry list and the ask and the business case for why states need this, and go right to the top and say, this is what we need. Because I think right now, it's being done piece meal for various projects and various entities, and it's not being done in a reasonable way if that makes sense.
MS. LOVE: Yes, states have the ability maybe to give back to CMS the under 65 --
DR. COHEN: That's right. It would seem like it could be a win-win.
MR. MILLER: You might want to know what Medicare is going to cost in 12 years based upon who is 45 today.
DR. COHEN: And, I think that's certainly in the purview of this Committee.
DR. CARR: Linda, did you have a question?
DR. KLOSS: Just a comment, and I just am really happy we've opened this dialogue. It seems that in some of the comments, Andre, that there are just a lot of very vexing policy questions that are maybe a little bit behind the building of the actual database.
And so, there does seem to be some urgency in understanding the scope of the kind of governance that's needed particularly because this is individually identifiable data. And so, you know, build the database first and then finish up the policy framework. And, I guess there's something we can do to help accelerate that part which I think is probably the basis for the unease.
DR. CARR: Okay well, that's terrific. You had your hand up. Are you, can we --
DR. HORNBROOK: Well, I just want to remind us all that there are three important concepts here: de-identification which is never 100 percent perfect, limited data sets which of course are a little bit more richer than de-identified files, and then re-identifiability. So, just because you think it is de-identifiable doesn't mean that it's de-identified.
DR. CARR: Thank you. As you can see, we have great interest in this initiative, and we look forward to ongoing collaboration. And, thank you again for bringing the whole team down. It really is tremendous to have everyone here. So, thank you very much.
MS. LOVE: Thank you for having us.
Agenda Item: Subcommittee Report Outs, Strategic Plans and Next Steps
DR. CARR: Okay, Marjorie just reminded me that we didn't hear the CMS update. So, what I would like to do is just get a brief read from each of the Subcommittees on where we are. Walter, do you have, did we hear --
DR. SUAREZ: Sure. The Standards Subcommittee already expressed a lot of the next steps. We have about seven core items that we need to finish between now and December, including the HIPAA Report, the two letters, the review of the ERA and EFT proposed operating rules, the 10109 provisions in the Affordable Care Act which are basically what's the next set of standardization activities, claim attachments, and then a primer on standards and operating rules.
So, we're going to lay out this into a work plan over the next five, six months, present it to our subcommittee, discuss it, and then bring it back to the Full Committee.
DR. CARR: Perfect, excellent. Okay, Larry? Populations.
DR. GREEN: We want to consolidate the Susan Kanaan reports that you saw at this meeting that were in the agenda into a foundational document that we can reference and build from the two hearings. We want to secure lead staff for the Populations Subcommittee. We want to participate and continue to work with the Privacy Subcommittee, and like on that call that they're planning on scheduling to identify next steps.
And most importantly, we want to organize a what over the last 24 hours has gone from being maybe a Privacy and Populations Subcommittee workshop to maybe a Full Committee workshop that we, maybe the Executive Subcommittee could help formulate the deliverables for that workshop together in some way. But, we think that that is a timely and very important next step to get ourselves organized for, you've asked us to think in terms of 24 months. And, we think that we're five or six months into that.
And, we think for the back 12 months of that window that you're asking us to plan for that we need this workshop in anticipation of that and that what the products will help us know where to go next. We want to stay focused on the community as a learning health system. And, we want to patent or copyright the phrase "working on the delta".
DR. CARR: Sounds Egyptian. (Laughter) So, that's great. And you know, again, I think that the themes that are going around are, you know, the HIPAA Report pulling all of those things together, the Community Data Initiative as a piece of the Learning Health initiative.
And, I think the discipline that we'll need in deciding next steps will be to secure deliverables that add value obviously and move, you know, capture the low hanging opportunities.
DR. GREEN: Though I think from our Subcommittee meeting, that was made by three or four folks, particularly some of our new players, that we want to not occupy space that's already well-occupied. However, our confidence intervals are very, very tight, that there is unoccupied space begging for attention. (Laughter)
And, we want to get there, but we want to get there in a thoughtful, prudent, targeted, judicious manner. And, we want to do it coordinating across the whole NCVHS. We are convinced that the Committee, as a whole, has work to do around this issue.
DR. CARR: And so, will the next step be that you'll put something framing together, and we'll have an Executive Subcommittee call to work from that framework? Or, how are we going to get to these deliverables or next steps?
DR. GREEN: Well you know, I don't know how to do that because I just don't know what's the most expeditious way to do that because I need Leslie, I look around the room and I'd almost need everybody in the conversation to get that right.
DR. CARR: Well, I think even if you can just say, you know, pulling out the themes in Susan's report I think is a good next step, working with Susan to just sort of say what did we hear? What would be immediate action items that would impact and make a difference in communities?
I think that's something that we want to keep an eye, keep track of. And then, looking beyond that, I mean, we've heard a number of powerful needs today. I think even if we just had a list, just unprioritized list, here are all the things that are on the table. I want to start with something and then think about how we might group it to get something out there.
DR. GREEN: Yes, right. We can put together what I would, why don't we call that issues that have surfaced so far this year or something like that, you know, what we've learned or the issues that need, that are out there. Not trying to prioritize them or be clear that that's what we ought to work on, but just to try to get an issues list.
DR. CARR: Right. Leslie?
DR. FRANCIS: So for Privacy, the next steps strategy that we had arrived at yesterday was a conference call which, what we will do is we will attempt to outline the privacy challenges and some of the strategies for dealing with those challenges that emerged from the community use of data. That would be the basis for our part in the planned workshop.
I believe there are two other things that are on our plate in the near-term that have surfaced today. The first of those is we've got to help Walter and Judy with the tenth report. There is a big place for privacy in that which we will be doing.
And the second one, I really do think de-identified data, this is my own personal view, that de-identification is neither necessary nor sufficient for protecting privacy. And, we heard from Denise and from Andre some of the kinds of concerns with All-Payer Claims Databases, and I would like to have us take on, I'm going to propose to the privacy group, people could say no, but that we take on quite directly the question of how to deal with the privacy challenges that we heard about in the discussion a few minutes ago.
DR. CARR: Thank you. Blackford?
DR. MIDDLETON: Well the, I guess the opening remark is we still need to focus on the relationship between costs and expenditure for health care in this country and quality. It's not clear that spending more is bringing us to higher quality. HIT has a variable effect on quality, so we need to understand the data and needs on those data or analyses with those data to demonstrate improving, methods to improve quality in this country.
So you know, the past two days, I think what we focused on has been three new topic areas for inquiry, consumer source data, the quality of quality data, and knowledge management as it relates to quality, information, and data. And, what I'm going to propose is that the Co-Chairs and staff meet in conference calls and bring a work plan to the Executive Subcommittee for review and discussion.
Requests of the Chair and the Executive Secretary, I think we do want to look at the staff complement for the Quality Committee and try to flesh that out a little bit. And, we also see us, the Quality Committee, being on the hook for the tenth HIPAA Report, and hopefully, we can be helpful there.
DR. CARR: Okay. So, I would add that you have a number of directions under consideration. I think, I'd encourage you and the Committee to think about the tie-ins to the learning health system, the Community Data Initiative and sort of really in the learning health system.
So, in some ways, some of the topics might not fit into that. But, I think as we think about the resources and the work product coming out of the Committee, we might want to prioritize to contribute to both of these initiatives.
DR. MIDDLETON: Yes, I think the third one, knowledge management, is actually central to the Learning Health System Initiative.
DR. CARR: Yes, okay. Are there any other comments? I have one. What a fantastic meeting, and, what a fantastic membership. I'm really ecstatic about all the work that's ongoing and the opportunities for working together going forward. So, thank you everyone and safe travels. The meeting is adjourned.
(Meeting adjourned)