[This Transcript is Unedited]
Hubert H. Humphrey Building
Room 705A
200 Independence Avenue, S.W.
Washington , D.C. 20201
Agenda Item: Introduction - Review Workgroup Charge - Mr. Hungate
MR. HUNGATE: All right, I’ve got the microphone calibrated and I think it’s about 9:00 and I think we’re about all here. So given that I don’t know how to specifically call a retreat to order but I guess I won’t worry about that too much. Welcome to you all, we’re here to work on a work plan for the next year to 18 months for the Quality Workgroup. Best to start with introductions and I would ask that anyone that has any conflicts of financial interest as it relates to the content of the discussion identify that conflict.
I’m Bob Hungate, I chair the Quality Workgroup here at the National Committee for Vital and Health Statistics, and I see two more people, one more person arriving, welcome, Don, and we will go around to my right. I chair the work committee, I’m principal of Physician Patient Partnerships for Health and chair the Group Insurance Commission in Massachusetts as well. And to my right --
MS. POKER: Hi, my name is Anna Poker, I’m from AHRQ, staff lead to the Quality Subcommittee and staff to the NHII.
MS. MCCALL: Carol McCall, I’m with Humana, I am vice president and director of our Center for Health Metrics, no known conflicts.
DR. SCANLON: Hi, I’m Bill Scanlon, I’m with Health Policy R&D and a member of this workgroup and no known conflicts.
MS. GREENBERG: I’m Marjorie Greenberg from the National Center for Health Statistics, CDC, and executive secretary to the committee.
DR. COHN: I’m Simon Cohn, chairman of the committee and the associate executive director for health information policy for Kaiser Permanente and I guess a guest for this workgroup. I have no conflicts of interest.
DR. JENCKS: I’m Steve Jencks from CMS, I’m a guest, I have the usual conflicts that I work for CMS and therefore I have a whole collection of biases but they don’t seem to effect other people so they shouldn’t effect me.
DR. DETMER: Don Detmer, sorry, the METRO ran slower then I thought it would today. I’m president and CEO of AMIA and had the honor at one point of chairing NCVHS.
DR. EDINGER: I’m Stan Edinger, I’m with AHRQ and also staff to the Quality Committee.
DR. HALAMKA: John Halamka, I’m a guest to the committee, I’m the chief information officer of Harvard Medical School and the CIO of CareGroup. And most recently I’ve been named the RIO CEO in Massachusetts so I’m responsible for all our data sharing. No conflicts.
DR. CARR: I’m Justine Carr, member of the committee, I’m in health care quality at Beth Israel Deaconess Medical Center and I have no conflicts.
MS. SQUIRE: I’m Marietta Squire, I’m with NCHS and I’m staff to the subcommittee.
MS. KANAAN: Susan Kanaan, writer for the committee and subcommittee.
MS. JONES: Katherine Jones, CDC, National Center for Health Statistics and staff to the committee.
MS. JACKSON: Debbie Jackson, National Center for Health Statistics, CDC, committee staff.
MS. ALBAVA(?): I’m Christine Albava, I’m a fellow at AHRQ.
MR. RODEY: Dan Rodey, I’m with the American Health Information Management Association.
MR. HUNGATE: Okay, thank you everyone. Don Steinwacks is also a member of the Quality Workgroup, will be joining us tomorrow, he was unable to join us today. I’d like to comment just a little bit on what we’re going to try to do. I’m indebted to Simon Cohn for suggesting that we might well spend a time in contemplation of what it is we should do and I’m also indebted to you, Simon, for the admonition to say what are the questions that you want to answer and I think it’s a helpful discipline. I sometimes have trouble coming up with all the questions, they seem to occur and I can’t always anticipate them. I think that’s part of what the context of this discussion is meant to be. We all come to this discussion from different paths, different experiences, different exposures to things, we have different conceptions built from that, and the point here is to share those and draw them into a broader discussion of performance measurements in health associated with quality.
The flow that we have set up for the session itself starts with a little background and where we’re really going to draw basically on both Simon and Marjorie for the things that have preceded what we’re doing here. Then we’ll move into a focus on looking at the customer perceptions as it relates to quality. We’ve set this up so that the members of the workgroup are chairing different portions of the discussion, sometimes I’m a little lazy and like to share the workload so we shifted it around so different ones of us will chair different parts of the discussions and that leaves it with a balance for others to contribute.
We’ll move from current performance measures to look forward into health informatics, an important evolving area that we need to understand as well as we can in order to know what to do today. Then the next day we’ll spend some time coming back and looking at what we’ve done today and then Carolyn Clancy will join us and Jerod Leob from the Joint Commission will join us and that blink is probably Brent James joining the discussion. Brent, why don’t you say hello and introduce yourself?
DR. JAMES: This is Brent James, pleased to be with you at least in voice if not in physical presence.
MR. HUNGATE: Well, we’re sorry you couldn’t join us in presence but we’re glad to have your voice and it sounds like the same voice I remember --
DR. JAMES: That must be you, Bob.
MR. HUNGATE: Yeah, it’s me, so we’ve got an authentication anyhow. We were just going over the agenda roughly.
The work plan then will get discussed after Carolyn and Dr. Loeb have brought us into their perspectives on what’s going on in their shops.
Now the way we’re trying to go about this is to look for, identify the issues and gaps in data and information, or unused information which has obstacles impeding its use. Part of I think our challenge is to understand what we can unique do as opposed to what happens in the rest of the system. My three years on the committee is about my learning curve for understanding what both limits are and what opportunities are and I now see that there are two specific kinds of work products that come out the NCVHS, there are letters which advise specific actions and there are reports which are more general and outline an area of activity. We don’t know yet whether our work plan will be characterized by a report at the end of it or whether it will be characterized by individual things. That’s part of what we’re trying to get at in this discussion but understanding what it is that will happen by itself based on the tensions in the health care system and what needs a nudge that we might be able to provide is another important aspect of that.
After some initial background by primarily Marjorie and Simon we’ve asked Brent to comment from the Intermountain Health and then we’ll move into the other discussion.
Now this is intended to be a discussion, not didactic presentation, this is an interactive, if you see something that you don’t think sounds right you’ve got to challenge it, this is an intellectual exercise, it’s actually being broadcast by the VA, so it’s a very public retreat which is just fine, I think what we do is in the public interest and needs to be in the public eye.
So with that said let me turn it over to Marjorie first and you can talk a little bit about what the Quality Workgroup has done and Simon will have some comments and then we’ll turn to you, Brent.
Agenda Item: Summary of Hearings/Work to Date - Ms. Greenberg, Ms. Kanaan, Dr. Cohn
MS. GREENBERG: Okay, thank you, and thanks to everyone whose joined us today and tomorrow for what I really have been looking forward to this and I think this is really an important kind of milestone or turning point for the workgroup.
Susan Kanaan, who introduced herself here, prepared a short summary of the two hearings that the workgroup had in 2004 which you all received in advance so I won’t go into great detail about them. Just to say sort of to provide some context and say that I guess Susan and I are sort of the committee historians at this point. We haven’t really been with the committee since 1949 but sometimes it feels that way. But when I saw Don Detmer walk in too of course part of the history, an important part of the history of this workgroup is that it was under Don’s leadership that the workgroup was formed. It might have been in this very room, over in kind of that corner I can actually picture it but this is either the blessing or the curse of having that kind of situational memory.
But I started thinking about kind of how we’ve gotten to today and I guess its partly my role to tell the new members that the committee didn’t start the day that they joined, that it did exist before them, and of course now we have Bill who can validate that because on his second round.
But not going back to the beginning but there is a long history of this committee, not having specifically a Quality Workgroup but struggling with the issues of the uses of administrative, statistical and to a lesser extent in the early history but certainly very much now clinical data for a variety of purposes to at the end improve the health of the population. In fact the very first, I was thinking the first subcommittee that I was lead staff to was called the Statistical Aspects of Physician Payment Systems, seems kind of quaint now but that was in the ‘80s and that was when we first really engaged with HCVA, when the committee really began to engage with HCVA to look at their payment systems and how they could contribute to help information beyond their basic purpose of paying bills.
And that led to a number of activities but to get to the more recent history, which may not seem that recent, but in 1998 when Don was the chair there was a commission as some of you remember, the President’s Advisory Commission on Consumer Protection and Quality in the Health Care Industry. And I believe that was actually before we had formed the workgroup but we did have a group of members and led by Cathy Coltin whom many of you know I’m sure, a very knowledgeable and experienced person in trying to use whatever data she and health plans could get their hands on to look at health care quality. And they produced this report, Quality First, Better Health Care for All Americans, and the committee commented on that report, this is on the website actually, August 31, 1998.
And the bottom line of their comments were that it gave, the committee was particularly interested obviously in the issue of investing “in information systems” which was one of the chapters, and kind of committee itself to contribute to furthering some of the aims identified by the commission in this particular area, in investment and information systems. And particularly looking at what kind of information is needed to measure quality of health care, primarily I would say the focus was on quality of health care and yet if you look at the title, Quality First, Better Health Care for All Americans, I mean I think there was some of that populations in there as well.
So it was shortly after that that we formed the workgroup, as I said under Don’s leadership and Cathy Coltin was the first chair, and the workgroup began to hold a number of hearings, or really more organized panels, in front of the full committee to look at what were the information gaps in measuring the quality of health care. And they held a number of really excellent hearings over a period of a few years and many of us felt that we needed to kind of bring all that together and it was about that point that Cathy rotated off the committee and Bob came on and we dropped it all in his lap and said okay, now what do you want to do with it --
MR. HUNGATE: It was certainly full emersion if not total submersion.
MS. GREENBERG: So this report that Susan has prepared reflects what came out of that. First there was, and you mentioned the two ways of operating, committee reports or committee letters, and both of those have come out of this. We had a report which pulled together all of the findings and the themes but they had been from hearings over a period of several years so there was a decision that we really needed to kind of refresh that knowledge and so the committee put out candidate recommendations rather then direct recommendations. And then we held two further hearings last year to talk about those candidate recommendations.
And I think as I said this summary really captures the high points and just to look at what major obstacles and challenges I think the purchaser and provider disconnect was very obvious in the hearings that we heard, both sides had very cogent arguments and pretty much saw things rather differently from the point of view of the need for certain information and measures. At the same time everyone agreed on the need for standardization, and agreed on the need to look at the tradeoff between looking at administrative data and on electronic health records. And this as I said goes back to many years of the committee’s work looking at what are the uses of administrative data and how can they be used for statistical or in this case quality assessment purposes, and trying to kind of balance that now with the new emerging electronic world and electronic health records.
So I think the bottom line was exactly what Bob said, that it was important, the workgroup agreed, to try to decide what its unique contributions could be to this debate. I personally don’t see it that much as a debate, I think Ken Kiser, who was at the second hearing, expressed at least my views and I don’t know if I’m supposed to be expressing my views here but since it’s a retreat I guess we can, that we really have to do both. We can’t just say well we’ll wait until everybody has electronic health record, on the other hand we have to recognize the burden associated with adding any additional information to administrative data, and not only the burden but the quality of the data and so we have to look for targets of opportunity and that’s where the committee did bring forward one of its candidate recommendations as an actual recommendation to the department recommending that the indicator for secondary diagnoses, whether or not they were present on admission, should be added to the new uniform bill for hospitals as well as the HIPAA transaction, institutional transaction.
And I think that really did have some influence because I in another hat sit on the National Uniform Billing Committee and they were leaning towards that way anyway because of recommendations of the Consumer Disclosure Project but the fact that the committee did recommend this to the Secretary I think just gave it that extra bump and it reminds me of maybe 15 years ago when we had the same impact on external cause of injury codes and recommended that those be added to the UB-02 which was the current version of the uniform bill and that also had an impact to then greatly improve the collection of external cause of injury data.
So I think that’s all I wanted to say, that I think the committee can make a difference, it has over its 55 year history but it has to be strategic obviously and recognize the different viewpoints and needs and I think that’s why we invited everyone in, to kind of help us think through some of these issues and I’m looking forward to the next two days.
MR. HUNGATE: Thank you. Anybody have any questions as we go through here in this dialogue stimulation process? Don, would you like to make any comment as it relates to having been referred to?
DR. DETMER: No, it’s great to be here obviously and obviously I really enjoyed working with Marjorie and Simon obviously have been at this too for a very long time with John Lumpkin and others. No, it was an excellent summary although a lot has happened obviously since then that I was learning from so I do look forward to the discussions as we move forward today so thank you.
MR. HUNGATE: Thank you. Simon?
DR. COHN: Obviously I have some remarks, one of the wonderful things of airplane flights is that you can sort of put your thoughts together on a long flight. However even before I start that I just wanted to make a comment based on what Marjorie was saying and this is I think just an observation for those of us who have dealt with information systems for many years is that even though I think we all think wonderful things of the electronic health record, and I speak for obviously Kaiser Permanente which is aggressively implementing a record, the Veterans Administration who has obviously is going on their next generation, Intermountain Healthcare and anyone else, Brent, you’ll probably comment on this also, but that the electronic health record is actually not a panacea for all data collection issues and indeed there probably is not a day that not one of the organizations that I’ve referenced and others aren’t making decisions about well, geez, yes we have an electronic health record but still what sort of information do we really need to be capturing here.
And so just the observation that those who look towards the electronic health record as the solution, we all just need to realize that the solution needs a lot of guidance and input from all effected parties because clinical care is the number one purpose of the electronic health record and everything else is sort of a subsidiary and exactly what collects above and beyond all that and which is really important for quality sometimes is a political process and certainly is a workflow issue because it does take additional time to collect additional data that isn’t necessarily only important for that one episode of care or for the next episode or whatever. And I think we just need to be aware of that as we move forward, once again I do think there’s a valuable purpose for that dialogue and actually for this workgroup as it begins to ponder some of those issues.
Now obviously I want to thank Bob for the opportunity to both be invited to this as well as to talk with you. This is sort of my opportunity to make a couple of comments since I’m actually not going to be here tomorrow and one of the sort of, I think Bob has done a good job of laying the groundwork but I wanted to reemphasize a couple of things.
First of all I think as we commented and I mentioned in March this is really from my view a time of change and transition for the workgroup. Up until now, and you’ve spent, really spent a number of years working on the Information and Quality Report which was started by Cathy Coltin during Don Detmer’s tenure into something that has taken I think a fair amount of time for us to finish off but from my view I think we’re pretty much done with that report and I think it’s really an opportunity for us to begin to look forward into sort of future goals and work plans.
Now I’m delighted that you and Bob and others took me up on my suggestion that you rather then just immediately charge off into an area that you spend a little time thinking about goals and work plans because I think that’s a very important activity. And certainly the world has changed a lot the last several years and I think if this isn’t the complete answer in the next two days I think it will sort of move you significantly forward in terms of coming to grips with well what is important and what may be what I would almost describe as the sweet spot for the workgroup, recognizing that we can’t do everything but that we want to focus on areas where we really can make a difference.
I’m also very pleased that once again I think we’ve referenced Don Detmer many times now but he was the chair when I came on in 1996, thinking both of our ages here, but I think it’s very useful to seek guidance from previous chairs and I think it’s great that Don was available and could join us for these conversations.
MS. GREENBERG: We’ll invite you back, don’t worry.
DR. COHN: Well, thank you, as I get older and older here.
Now let me just talk about sort of my view of the workgroup and the work going forward in some general outlines, and once again these are sort of partly my views, partly views that I have sort of put together and integrated from conversations I’ve had with all of you, because as you remember one of the first things I did when I became chair was to sort of talk with everyone and try to put things together. But really what I’m interested in doing is trying to make the full committee as well as all the workgroups and subcommittee as effective as possible recognizing that we don’t have unlimited resources. I’m also temped to use the word as efficient but effective is I think really the key piece here.
The other piece I want to do is to make sure the work is as satisfying and successful, because clearly you aren’t in here in this for the money and so I think we better make sure that it really is satisfying to all of those who are involved and I speak both for the full committee, for the committee members as well as the staff in terms of moving forward on all of that. To my view really the first is really this well articulated clear goals and work plans that everybody agrees to and everybody feels is important and not redundant with everything else that’s going on and things that we all have a conviction makes a difference and so that’s really what I think you’re working on these next couple of days.
I think the other piece, and I’m glad that you’ve begun to get some people in to talk such as Steve and Trent, and Trent, you’re staff I guess, right? As opposed to testifying today?
DR. HAYWOOD: What did you say about it?
DR. COHN: Oh, I’m sorry, I was just referring that you’re participating but not actually testifying today --
PARTICIPANT: Yes, he is.
DR. HAYWOOD: Oh, I am the guy for today --
DR. COHN: Oh, you are the guy for today, okay. But anyway, I think it’s important that we talk with customers and if at all possible we have well defined customers for the work, people that are committed enough to the outcomes that they actually want to participate in the work, and that’s been out of my experience something that’s been very useful as we’ve gone forward. It also helps define and guide us as we have to make decisions on the scope and date of the deliverables and all of that. And obviously in all of this and I’ve commented, another key is dedicated and skillful staff and I’m dedicated and I think we all are to making sure that they are supported and helped and we want to make sure that they’re as successful in their work and that we’re doing things that help move them along in their career and their activity.
Now I also want to comment that we need to be intelligent about sizing the work. Now I’m not, I think I’ve commented to some of you before that let’s do letters and not reports and I am going to pull back a little bit from that and not put that as a law but just we all need to talk about it and it’s really a question of sizing the work. There are sometimes big projects that we will decide upon but I can think of nothing worse then taking so long working on an area that by the time you deliver recommendations nobody cares anymore. I don’t think we’ve been guilty of that but at times we sort of get on the edge if we’re not careful and so I would say that even if we take on a large project or a large area, and I’d have you be thinking about how we can size the work so that we can give timely recommendations and input going forward as opposed to waiting until the end of the process.
Another piece is of course is that we have actionable recommendations. Now I’m reminded with the NHII report, which actually did take a number of years in preparation, but one of the differences that and reports like that have made to the Secretary and others is is that rather then high liven sort of abstract objectives, recommendations, the go forth and prosper style of recommendations that people need to figure out what they mean, we typically have tried to get concrete enough that we sort of try to say geez, within the next two years you should be doing this or for the next three years you should be doing that, year five to seven you should be doing B, I mean the sort of things that are described as actionable recommendations.
I think we’ve all observed that the pace of life has gotten much faster the last ten years, I was reading recently that going from I think a normal workforce to the hardest working workforce in the world at this point and so everybody is working 40 to 50 hours a week, people are looking at documents, they’re reading the executive summary and not reading the page 13 of 20 pages of documentation generally. I think they also want to be told what they need to do rather then having to figure it out after they read a report and so I think that what we need to do is to make sure whatever we come out with, even if they may be high level concepts, we need to drill down to the first step, the second step, and third step that moves you into that direction in a concrete enough fashion that people know what to do with it.
Now finally and I think going back to the term of successful is that in all of this we need to make sure that we are leveraging the skill and knowledge of the members of the workgroup as well as the staff. Obviously that’s part of the satisfying piece.
Now in terms of the actual work obviously I’m hoping out of the next two days, I will also comment that I’m an emergency physician and so my tendency is to be a little more aggressive and a little more impatient then some other medical specialties. And I do have to rein that in a little, I know John probably shares my view on this one, I know I got into clinical information systems back more years ago then I like to count thinking that we’d have this solved in about a year and I’d go back to full time emergency practice after that and here we are working on it. But having said that what I’d like is substantial progress being made over the next two days, I would like to see some work plan and some goals, if not complete at least views of it out of the two day work. And I would say both for the, I know that you’re working on the workgroup activities but I also would like you to take a little time to reflect on what the goals of the full committee ought to be in terms of quality. And I’m hoping that they’re going to be aligned but you may decide that there are issues that really are full committee issues that are quality and I think it’s critical that we keep quality as a full committee agenda item as well as just for the workgroup so I want you to reflect on that a little bit.
Another piece is is that when I was looking on the web I noticed the last time we had looked at the charge of this workgroup was when we established it back in 1999 so you might want to at some point look at your charge and make sure it still expresses what you’re all about, I would comment that’s also the last time I saw a work plan which was in 1999.
Obviously the next steps in all of this is that we have an executive subcommittee meeting next week, not that we will finalize goals but I think it’s going to be, we’ll be talking about subcommittee, workgroup and full committee goals, the rest of this year and on to next. I’m hoping that we’ll be able to do some integrating and look at synergies across workgroups and subcommittees, so obviously the more work you can do at this meeting that will help feed into that conversation will be very useful.
Obviously as a comment I will not be here tomorrow but I’m very much looking forward next week to hearing Bob’s report on I think the good work that you’re all going to do.
MR. HUNGATE: Good, thank you, we will have some things to report. I think the comment that you made about the work plan itself and customers and staff involvement are the critical pieces that we’ve deliberately made sure that we have good content from AHRQ and from CMS in this retreat and I think that we will come out with a ratified work plan which fits with those customer needs, which is appropriate. I’d like to take a minute before Brent starts by maybe going around the room quickly again so that Brent knows who is here as his audience. Brent and I met in Chicago actually first when and I were part of the Quality Measurement and Management Project at the Joint Commission, it wasn’t the Joint Commission, it was funded by HRET, the Health Research and Education Trust, and the intent of that project was to get all the purchasers who were making demands, all different, to agree to a comment set of demands. The objective was laudable, it made little progress and died after about two, maybe three years of effort, but I remember Brent’s stellar contributions to that discussion as we went along. And so I welcome you here and again, I’m Bob Hungate and let’s go around again and let you know who’s here, Brent.
MS. POKER: Anna Poker from AHRQ.
MS. MCCALL: Good morning, Brent, this is Carol McCall with Humana.
DR. SCANLON: Bill Scanlon from Health Policy R&D.
MS. GREENBERG: Marjorie Greenberg, NCHS.
DR. COHN: Simon Cohn.
DR. JANES: Gail Janes from CDC.
MR. MOROSKI(?): Ulock(?) Moroski from NCQA.
DR. JENCKS: Steve Jencks from CMS.
DR. HAYWOOD: Trent Haywood, CMS.
DR. DETMER: Hi Brent, Don Detmer from AMIA.
DR. JAMES: Hey, it’s good to hear you again, Don, it’s been awhile.
DR. EDINGER: Stan Edinger from AHRQ.
DR. HALAMKA: John Halamka from Harvard Medical School and CareGroup.
DR. CARR: Justine Carr, Beth Israel Deaconess Medical Center.
MS. KANAAN: Susan Kanaan, writer for the committee and subcommittee.
MS. JONES: Katherine Jones, NCHS, CDC.
MS. JACKSON: Debbie Jackson, National Center for Health Statistics, CDC.
MS. ALBAVA(?): Christine Albava, fellow, AHRQ.
MR. RODEY: Dan Rodey from AHIMA.
MR. HUNGATE: Okay, that’s all of us here in the room and there are unidentified people listening in live on the internet. So Brent, welcome, we look forward to hearing from you.
Agenda Item: Quality Initiative and Perspectives from Intermountain Health Care - Dr. James
DR. JAMES: Thanks, Bob, I’m going to need a little bit of your help because I really did want this to be a free and open discussion and my trouble is I can’t see people who have an issue to raise. Will you help by, I’ll try to pause from time to time but by interrupting me when necessary?
MR. HUNGATE: Yes, I’ll watch for the need from here.
DR. JAMES: And I find that I can hear you quite readily, I think you’re close to the speakerphone and some of the other folks kind of fade in and out a little bit.
A little bit of background just to kick this off, background number one I regard this as really a summary report of the work of the Strategic Framework Board of the National Quality Forum, but we talked about similar issues in trying to establish the theoretic framework for quality reporting for the country. I know that you’re also using the report of the IOM’s Patient Safety Data Standards Committee and I’m drawing very, very heavily from chapter eight of that report and in some sense summarizing things that we’ve documented fairly well there.
I think it was Simon who said earlier, I couldn’t tell quite exactly, you were talking about electronic medical records and the mere fact we have an EMR doesn’t really solve it at all. Underneath it is a key design issue, how do you figure out what you really need, and so that’s what I really wanted to address in these comments.
I’ve asked Anna to pull up on the screen, I presume that’s working Anna, copies of the slides so that everyone can see them. Did you get those, Anna, that I emailed?
MS. POKER: No, Brent, we never got them, I never received any slides from you, I’m sorry.
DR. JAMES: Do you have access to your email?
MS. POKER: Probably not, is it something you could talk through and I could send out to the members afterwards? I never got anything, I thought you decided not to.
DR. JAMES: I sent them, it would have been last night and I sent them to Bob and to you so if you can hit your email I think you’ll find them there.
DR. DETMER: Brent, if you want to send them to me at Detmer@AMIA.org I think I can get it through my PDA.
DR. JAMES: It’s going to be as a couple of PDF files, Don.
DR. DETMER: Okay.
DR. JAMES: Give me two seconds so hold on a second here --
MR. HUNGATE: The marvels of technology.
DR. JAMES: So what it is Detmer --
DR. DETMER: @AMIA.org.
DR. JAMES: Detmer@AMIA.org, I’m going to send two copies, Don, one’s going to be black and white handouts suitable for reproduction and the other is just a PDA file that could be stuck on a PC and displayed.
MR. HUNGATE: Brent, just a second, we may want another backup addressee --
MS. GREENBERG: This might be a little secure disk but if you send them to me at MSG1, like monosodium glutamate, MSG1@CDC.gov, I’ll pick them up on my Blackberry and I can forward them to like Jim Scanlon’s office and they can print them out.
MR. HUNGATE: That will give us maybe a two path possibility.
DR. TRENT: Right, well here they come guys, so bam, so look for them fast, Don, they should be on their way. This makes it a little bit trickier, so imagine you’re looking at a slide, it’s actually a figure that we had in chapter eight of the Patient Safety Data Standard Report. A key concept is, well, to quote W. Edwards Demming is that aim defines the system, that the intent of any system depends upon what you hope to produce through its operation and that’s particularly true with data systems.
We laid out a continuum, imagine at the far left hand end, we talked about comparative clinical data for accountability and at the right hand end we talked about data for learning or for quality improvement. Along the continuum there are a series of possible activities, so starting at the left accountability, the first we called judgment, that might be legal sanctions, malpractice tort actions, professional or social shame, licensing, credentialing, privileging the standards. That’s one purpose for which you use comparative outcomes data.
Moving to the right, the next area we called selection, we described this in some detail in an article that Don Berwick led from the Strategic Framework Board, published in Medical Care 2003. It was the January issue in a supplement from the Strategic Framework Board, it’s the idea that we could produce information that major purchasers or even patients could use to choose a physician or choose a hospital.
Moving one step further to the right we could also use comparative data for payment, for example pay for quality systems. I’m going to come back to that in just a moment because I think it turns out to be a key point. One step further to the right, number four, prioritization. This is the idea of incenting leaders in health care systems to focus on quality outcomes let’s say as opposed to just financial performance. In some sense it might be public shame, it might be the desire to be an excellent performer in a marketplace.
Interestingly that idea of prioritization is probably a more functional term then is the concept of motivation, the motivating people to focus on quality. It turns out most organizations do have a focus on quality, it’s just that it gets shoved down the list compared to other priorities and so it’s the idea that comparative outcomes data could move these important things that we all agree are important up higher on the list.
Finally at the far right under data for learning we have process operations in management and improvement so generating and testing hypotheses for improvement in a clinical care delivery setting. Now interestingly depending upon where you are on that continuum the aim changes, so to over simplify a little bit if you’re at the far left under accountability the real aim of your data system is accurate ranking, can you identify those who are stellar performers versus those who are poor performers, for example, who are delivering good care versus bad care. When you move to the far right under data for learning the aim of your data system changes.
Turns out a primary aim, between the ends of the continuum, at the far left under accountability the focus is inherently on a person. Now that person may be an individual physician let’s say, a single human being. When I use the word person here although I also mean an institution, a corporate person perhaps, a hospital, how well a hospital is doing or even how well a state is doing.
As you move to the right of the continuum it shifts to data for learning where the learning is on process so this is going to be a key distinction as well, is the focus of my data system ranking individuals or is the focus shifted off of individuals onto process. In a very real sense all the clinical research inherently focuses on which treatment works as opposed to which physician is best if you see what I mean, and this is going to turn out to be a fairly key distinction. Well, with that, number one, aim defines the system, that continuum in background.
Imagine you skip to the second slide labeled outcomes assessment. In assessing outcomes you usually move through three consecutive steps. The top of the list you need to understand differences in patients. By differences in patients what I mean is individual anatomy, physiology, biochemistry, genetics, that change an individuals susceptibility to response to disease and treatment and testing as far as that goes. Of course we like to extend that concept, not just disease treatment but the health promotion/disease prevention. It relates directly to burden of disease, I guess I’m not sure where these next two ought to fall but it also extend to the idea of a patient’s preferences and believes, their ability to participate in their own treatment, for example educational levels, interest and engagement in their own health care, and access to resources. The reason I’m not sure about those last few is because I’m not sure if they ought to fit under differences of patients or if we now are kind of blending into the second category, differences in treatment.
Differences in treatment might concern things such as availability of resources, active programs for health promotion/disease prevention, problem and opportunity identification, in other words complete and accurate diagnosis, selection of all appropriate interventions so the idea of referral and treatment indications, doing everything that works but only what works. The actual execution of tests and treatments is what we usually think of in terms of hospital or physician performance and a key part of that we know is patient relationships, the idea of service quality, attentiveness, information transfer, shared decision making, basic dignity and respect.
While in theory we have those differences of patients and then move to differences in treatment and produce a final category of differences in results or differences in outcomes. Generally we divide outcomes into three major classes, medical outcomes, that would include things like appropriateness, did we do the right thing. Complications, I guess in the jargon of QI are called process failures or defects. Therapeutic goals, did we achieve what we set out to achieve from a professional viewpoint and then the patient’s perceptions of their results, that would be some sort of a functional status measure.
Well, if medical outcomes are the first category side by side and perhaps of equal if not exterior importance are service outcomes the clinician/patient relationship being the most critical. Access and convenience of services. And then the final third category of outcomes are cost outcomes.
Well, the basic theory is that if we could somehow account for differences in patients through stratification let’s say in a randomized trial, or randomization itself, or in an observational study designed perhaps through some sort of risk adjustment strategy. If we can account for differences in patients we can measure differences in results then we can attribute differences in performance or treatment on the basis of that.
On my next slide what it says is you’ll see I hope --
DR. COHN: Brent, I apologize for breaking in but it looks like we finally got your overhead so can you give us just a minute to get them up and then tell us which slide you’re on.
DR. JAMES: So I’ll be on slide number three when you get up, it’d be good I think, Anna, if you just pop up that cover PDF file on your projector --
MR. HUNGATE: Your assistant at this point is a combination, Brent, of Don Detmer and the technical folks from here. Don’s got his computer over being hooked up to the slide machine --
DR. JAMES: Cool, I love it.
MS. GREENBERG: Maybe this shows some hope for the electronic --
DR. JAMES: I want you to know guys just to put this all in perspective that I was a few years ago asked to give a talk on health IT at the Institute of Medicine annual meeting and got up there to give my talk and what do you suppose happened, my Notebook failed. And I think there were a lot of physicians in that room who could really relate to that having tried to deal with the MRs as they move through their --
MR. HUNGATE: I see before me a slide labeled outcomes assessment, differences in patients, differences in treatments, differences in results, does that sound familiar?
DR. JAMES: That sounds familiar, is it in beautiful blue with black?
MR. HUNGATE: Yeah, classy slide --
DR. JAMES: Under view you say full screen, you can get rid of all, they’ll be acrobat site junk --
MR. HUNGATE: He’s working away on that.
DR. JAMES: Probably already done that.
MS. GREENBERG: He’s not the CEO for AMIA for nothing.
DR. JAMES: In some sense I ought to apologize too for reviewing things that I suspect that pretty much everybody in this room knows intimately but sometimes good just to get them written down.
MR. HUNGATE: Okay, we’re also out making handouts which will match this so you’re on outcomes assessment --
DR. JAMES: Tell Anna to do one thing, when she asks Adobe Acrobat to print it, click the little box that says print as black and it will look a lot better.
Well, anyway, so outcomes assessment, forward one slide, concept number two, outcomes assessment gone bad. The thing that most people overlook is that all three of those categories are intimately effected by differences of measurement, completeness, accuracy and timeliness of the data involved.
Moving ahead to slide number four, what I’ve tried to do is put together a measurement chain of the elements necessary to get accurate measurement in the clinical system. In some sense this is reflecting I guess an ideal that we know is never quite perfect, we never quite achieve it, but I think it’s good to lay out the elements.
Top left, basic science, the question is is do we know the data elements to predict treatment response and outcome. As an illustration David Eddy has published a little example of infant mortality. What he did was talk all of the known factors that predict infant mortality including treatment, did you get prenatal care for example, and build a model, he discovered that all of the known science predicts about 25 percent of the variability in infant mortality. Everything we know accounts for a quarter of variability in infant mortality, the other 75 percent are unknown to current science. Of course that means if I’m comparing infant mortality across countries or across states I’m making the inherent assumption that that other 75 percent is equal which sometimes may be suspect. Well, that’s the idea of the first step, science.
The second step, measure selection, assuming that we know the science, we go through due diligence to track down the things that are really important, do they make it into my measurement set, did I include all known major factors in the measurement set. Interestingly in doing that sequencing could be very, very important. Just finished reviewing a little article for a major journal that looked at UB-92 data, whether those diagnostic codes are clearly identified as present on admission, usually you note it as within 24 hours, well, how is it recognized within 24 hours of admission versus the complication of treatment makes a profound difference in measured outcomes as you might expect.
There are also issues around measure selection that tie into aggregation or specialization in case selection. So for example we here at IHC under an AHRQ contract assess the current patient safety indicators and the hospital performance indicators that have been developed by Mark McClellan when he was still at Stanford. One of the things that we gave to them was how high did you have to aggregate the data before you started to get reliable results and for I’d guess more then half the measures we had to move at least to a metropolitan service area level or perhaps to a state level before we started to get reliable results.
Well, after measure selection the next question is is were those factors assessed as the patient when through the care delivery process so are the data available to begin with. After the question of were they assessed the next question is were they actually recorded in the chart, and then of course a major step, if they were in the record were they abstracted in an accurate way. And all of those are necessary before I can do analysis and reporting and as many of you know there are many, many failures along that potential chain, the science to measure selection to patient assessment to documentation and finally abstraction for purposes of analysis.
The usual metric we use for quality of data is three fold, A, they need to be complete, sort of getting all the important fields and all of them that are required are filled in, they’re accurate, they’re filled in with well, correct data that reflect the patient’s real condition or the real result, and are timely, they happen in a timely way.
Well, to summarize what we I think fairly well documented in the Patient Safety Data Standards Report most generic accountability systems cannot rank accurately. This is the next slide, the classic example was Henry Crakaron(?) and Cliff Bailey’s(?) work with the HCVA mortality reports. By their own assessment they had positive predictive values so, well, let me back up, eight mortality sensitive conditions, UB-92 data submitted to CMS, HCVA at the time, they had a 9th measure that rolled all eight together into a single mortality measure. Absolute state of the art risk adjustment frankly, as good a job as I’ve ever seen, so very carefully done, Cliff and Henry were arguing that they had a positive predictive value of about .6, independent assessments of the same system placed their positive predictive value between .25 and .4, so just to make that a little bit more explicit, they were classifying hospitals based on their risk adjusted mortality rates.
If you were in the top five percent of the country in terms of risk adjusted mortality you were labeled high mortality outlier and reported in the newspaper. So for eight years, 1984 through 1992, each year these statistics were released and at least in the Western United States hit the major newspapers, it was usually a big article. And their whole stated purpose was to help patients choose to vote with their feet, they thought it would have a positive effect on quality in the health system because well hospital administrators, physicians and other professionals would strive to be good relative to this measure.
Of course the reason that it failed, the reason that it was eventually withdrawn, is because of that little positive predictive value. So if hospital if they ranked as being a high mortality outlier independent more precise clinical measures found that they were correctly classified 24 to 40 percent of the time, 60 to 75 percent of the hospitals that were ranked as high mortality outliers were not high mortality outliers using more precise measures. Now on that slide I’ve said cannot, it turns out that Henry and Cliff did about the best job you could possibly do with UB-92 data, at least I don’t know any way to make it better, it had to do with failings in the underlying data system itself. If we were to go back one slide, failings of the science, failures in the measure selections, whether the patients were properly assessed, documented, and whether it was abstracted, that’s a common pattern that you routinely see in these systems.
Now question number three to move ahead one more slide, this would be page number seven, Don, I just wanted to ask the question what does outlier mean. These are real data from one of our hospitals, we routinely track now, we get clinical outcomes I guess for about 70 or 75 percent of our care in an inpatient or outpatient setting. And the way that we routinely report it is statistical process control charts, always the center line in that SPC chart is your risk adjusted peers.
So this is a hospital down in Southern Utah, St. George, Utah, Dixie Regional Medical Center, shows their overall C-section rate, their expected C-section rate was just under 15 percent within our system risk adjusted compared to other hospitals that did not have newborn ICU coverage so we tried to compare to true peers. What the chart shows is that Dixie Regional Medical Center was consistently over time significantly higher then expected, they were running a rate of about 19.2 percent.
Well, that clear outlier, what does it mean to say that they’re an outlier? Of course our first knew jerk reaction is that their C-section rate is too high, that it reflects some sort of medical performance. In fact technically what it means to say that you’re an outlier, technically from at least a quality improvement viewpoint, next slide, page eight, it means that if you analyze carefully you can find the true root cause of being an outlier.
Now we will run literally oh I don’t know, 30,000 or 40,000 of these reports per month, I mean we aggregate the data at physician level, practice group level, bring it up to clinic or hospital level, regional level, whole system level, depending on what makes sense for each particular clinical result. And we found hundreds if not thousands of outliers and on many of them we’ve gone to the extra effort to track them down. Now realizing that we’re talking about data systems that are not UB-92 administrative data, these data systems are usually explicitly designed for clinical outcomes tracking in ways that I plan to show you in just a minute. Interestingly in doing that across the last eight years more then half of all the outliers that we’ve analyzed have turned out at root cause level to not be problems in care delivery at all, the problem is in the data system.
If we go back one slide again to that one that shows Dixie Regional and their C-section rate, well, when we drilled in on that it turns out that Dixie Regional supplies services to a fairly unique community in the state. It’s down in the far southern desert of Utah just north of the Grand Canyon, we have a series of groups who have moved to that area of the United States mostly to get away from the forces of government, some of the most prominent are fundamentalist religious groups, polygamous groups, make the news from time to time. They’re served by a series of largely untrained lay midwives, so most of the deliveries that happen in that community happen in private homes. But a pretty good network has developed over the years, the lay midwives when they get in trouble with the delivery will tend to phone the hospital and stick someone in a private automobile and send them into Dixie Regional. It means that while the care is in route they’re preparing the OR for an emergency C-section.
Well, a funny thing happens, if you adjust that population for those abnormally high rate of emergency C-sections coming through the ED Dixie Regional stops being an outlier and their rate moves to the regular rate. So what we had to do was go back and change the data system to account for this factor that nobody had anticipated until it happened in reality.
To move ahead again to page eight the conclusion that at least we drew from this is that even a well designed clinical outcome system requires improvement feedback, especially when it’s new, that you start to track down outlier cases to find the root causes because at least in our experience more then half the time, even with a data system designed for clinical outcomes, you’re going to discover a series of defects in your design. And this gives you an opportunity to change it. I think that this reflects just core honesty in using these data systems, I mean when I measure outcomes I’m really asking physicians and nurses to change their practice, change their behavior, can I ask that of them if I’m not willing to meet the same standard relative to my data systems. Well, outliers.
Now while I’ve criticized outcomes tracking systems for purposes of ranking it turns out that that’s not a generic truth, that there are some outcomes tracking systems that rank quite accurately. A classic example would be for example Ed Hannon, risk adjusted morality outcomes after open heart surgery at New York State. Or a similar system that Jerry O’Connor and his colleagues put together in the Northern New England Group. We have a series of good examples of outcomes tracking systems that do rank accurately and work quite well.
Well, slide number nine, if you look at those systems they tend to show three features. First, they tend to focus on a single condition. Now I’m going to turn back and call that a clinical process in a minute. Second, they tend to use carefully designed data systems around that clinical condition reflecting that measurement chain that I showed earlier. In other words they don’t just rely on existing administrative data for purposes of convenience, their designed to a particular purpose. And number three, they use intermediate as well as final outcomes.
I thought this would be a good opportunity to talk about intermediate versus final outcomes, I think it’s an important concept that sometimes can be a little confusing. So the key background concept is something called an outcomes chain. On slide ten I’ve tried to illustrate a simple outcomes chain for diabetes mellitus.
So this is closely simplified as we’ll see in a minute, so I’m tracking diabetes, the real long term final outcomes metrics are well first death, underneath that though are four major complications, retinopathy leading to blindness, it’s the major source of blindness in the United States. The big one, well, one of the big ones, nephropathy, kidney damage leading to dialysis and transplant. Peripheral neuropathy, damage to the peripheral nerves in the hands and the feel, this is the major source of pressure injuries and amputation in the United States. And I guess the really big one is of course macro vascular cardiac disease, coronary artery disease, heart disease, death.
Well, it turns out to track those four big long term outcomes and death that result from diabetes, your timeframe is roughly 40 years, 30 to 40 years, to get enough advance to really be able to make statements about how the treatments are working. 30 to 40 year timeframe puts this well beyond the scope of most practices --
-- [Laughter.] --
-- we have a couple of solid, actually three now solid randomized control trials, DCCT, UKCDS being the first two. What those trials did was link the long term outcomes of diabetes mellitus to a well recognized, well, hypothesis about cause I guess, to blood sugar levels. From those trials we know that hemoglobin A1Cs predict retinopathy, nephropathy, peripheral neuropathy, macro vascular cardiac disease rates and death. Now that means that I can use blood sugar levels as an intermediate outcome to track what’s happening to my patients.
As usual when I used intermediates in this case it first of all massively expands my sample size, so instead of getting one event per patient I get about a measurement per month or every two month per patient, so I might get tens if not hundreds of measurements per patient. And it changes my time scope from 40 years down to a few months. Well, that set in an outcomes chain, it turns out that outcomes chained together as a general principle from those long term final end outcomes to a series of intermediate outcomes, if you drive down the chain you’ll eventually hit a level at which clinicians, physicians, and nurses make decisions, it’s called decision level. Well, for diabetes mellitus for example the things that we actually manipulate to manage blood sugar levels are some sort of hypoglycemic medication, diet and exercise, at the level of a patient or a physician taking action.
So next slide, slide 11, outcomes chains, these are the things that David Eddy called causal chains in his writings about guidelines. They tracked the hierarchical elements of a process/outcome structure from a final outcome to a series of intermediate outcomes down to the level of actual decisions or behavior. Just in passing if you’re going to drive change into a system decision level it’s the only plan at which change can happen. In many instances we have measurement systems that can’t support decision level, they’ve not designed to support decision level, if that’s the case then the only way that those systems can be used is for judgment, for ranking in some form.
Now just to make the case, it turns out that the diabetes example I just showed you is a gross over simplification. Slide number 12 is a real outcomes chain put together by Dr. Larry Staker. We were in an applied research center for NCQA in putting together the original HEDIS measures for diabetes. This little outcomes chain is what the team actually used to pick the current HEDIS diabetes measures. We tried to lay out the whole thing, get it on one sheet of paper, this is not quite perfect by the way but it was functional enough and that led both to the HEDIS measures and later to the dequip(?) measures for diabetes which I think are one of our better proven outcomes tracking systems that we use in the country. So very, very useful tool.
Well, the point of all of this, it’s appropriate to use intermediate outcomes. Well, take a look at slide number 13, the key to an outcomes chain is the reliability of the links so in some instances they have very, very strong links. For diabetes I have UKPDS and DCCT linking hemoglobin A1Cs, blood sugar levels to long term outcome and so I can quite reliably use that intermediate outcome. Sometimes the links aren’t strong at all and then of course I get into some real problems, attributed linkages, inappropriate attribution I guess. But usually if I can find strong links they allow appropriate substitution of intermediate for end outcome which can massively increase data rates while shortening time length.
I should say in passing at the pure technical level when you start to understand the processes and outcomes are inherently hierarchical that you cannot distinguish between intermediate and final outcomes, that it turns out to be an artificial distinction. It all depends on your level of abstraction, so for example the big end outcome we usually track in health care is death, mortality. Well, wait a minute, is that an end outcome? If I were to bump my level of abstraction one level let’s say to a philosophical or a metaphysical or a religious level, well, wait a minute guys, death is just one step in a process.
So what is it, is mortality a process step or is it an outcome? See the idea? But very often groups get hung up on whether we’re tracking final outcomes or intermediates and the point is is that it’s not so much whether you’re tracking intermediates or finals it’s the strengths of the links that count, that would suggest that when you’re building a data system that you might want to carefully examine the linkages to make sure that you’re using appropriate measures based upon how well they predict outcomes of importance or interest to the population you’re targeting.
Bob, I’m going to stop right here, any comments so far?
MR. HUNGATE: Okay, let’s see. Comments? My only comment is I applaud the discipline in the presentation, it’s very, very helpful.
DR. DETMER: Yeah, I’m going to want to weight in at the end with something to try to set this into a context, this is Don, that might be useful, unfortunately you won’t be able to see the slide that I hopefully will find to show but at any rate, it’s a terrific --
DR. JAMES: Well, I have one more background thing I think and then I want to see if I can kind of hopefully extend what I’ve said to conclusions and that, well, we’ll see. Next slide --
MR. HUNGATE: Just a minute, we do have another question --
Brent, excuse me, we have a question, I’m sorry.
MS. MCCALL: Brent, this is Carol McCall, first a comment, this is absolutely delightful. Second is a question, which is when you go back to the slide that has the fairly detailed chain, how many of those for say the IOM conditions that are high priority actually exist?
DR. JAMES: Well, where I’m going to conclude just to warn you is that we actually established a formal methodology around these things as part of the strategic framework for our report and then we endorsed it on Patient Safety Data Standards as we worked our way through it. To my knowledge, well, we’re finally starting to see at a national level some movement, the putting together these shared tools. I can tell you that within IHC I’ve got about 40 or 50 of them.
MS. MCCALL: Well, okay, maybe asked another way --
DR. JAMES: That’s the short answer, and it’s becoming now more of a common tool that we all use. Now there’s another version of it that’s not an outcomes chain, there’s a version of it that’s on a process level instead of an outcomes level and it works about the same way, they’re called conceptual flow diagrams. And I actually like them a little bit better and I can show you those too but let’s just say it’s not just one tool in the toolbox.
MS. MCCALL: Yeah, I guess the question maybe stated another way is what percent of the population, what percent of the disease burden, what percent of some large number of people and/or dollars do we have outcomes chains that have been established and accepted, at least are on their way, so that if we wanted to try to leverage the concept and the approach within administrative and clinical data metric systems and everything that you’ve laid out, how far are we along the path of having things that we could begin to use?
DR. JAMES: We’re still relatively young. The best national example we have is NCQA frankly, most of the NCQA measures have this kind of support back behind them if you go hunting. I wish I could say the same for Joint Commission but at the moment I can’t. I have an interesting ongoing debate with Ken Kiser around this thing because I was hoping that NQF would follow it as a rigorous methodology and Ken’s been using, and it’s understandable, he’s been using mostly a political process, given the structure of his organization I think he has to.
In the last few months we’ve seen what I consider to be some solid progress on getting people to move toward the discipline, right, but one of the things that I’m going to suggest to the workgroup, I think, I want to submit this for your consideration. Here’s the crazy thing, here I am at IHC and currently about 75 percent of my care that we deliver in IHC, Medicare, commercial, inpatient/outpatient, we’ve got the data systems and this is the way we got them and they work. I can show them working to substantially improve care, why in the world do I have to build this on my own? I mean come on, diabetes is about the same in Massachusetts as it is in California as it is in Utah as it in Australia or it is in Sweden as far as that goes.
Single biggest thing we do is pregnancy, labor and delivery and we’ve shown massive improvements by using these techniques, well, wait a minute, again I think that delivering a baby is pretty much the same regardless of where you are in the country and why is this not a shared intellectual tool that we’re developing together as a country. I think that building that kind of infrastructure could have a profound impact on our rapidity of advance in terms of underlying national health information infrastructure, in terms of accountability, well every aspect of the system is what I’m going to come back and argue to you.
So is that fair, Carol?
MS. MCCALL: Yes, it is, I think ultimately if to become a vehicle that we can exploit to do this, because it’s a wonderful framework, they either need to exist, those types of chains, or they need to be within our reach if not within our grasp.
DR. JAMES: Yeah, I could imagine a circumstance where a group like AHRQ let’s say, or perhaps it’d be medical professional societies, maintained our current state of the art thinking around a particular clinical process and shared this kind of a tool and it becomes a key design tool for the informaticist building the information infrastructure. It’s also pretty good for teaching, training residents and the like just in passing. But to date at a national level I guess we’re five or ten percent so we have enough experience with them to kind of know how they work and to kind of make the case I think, but on the other hand we haven’t moved ahead on them vigorously as a country. I think there’s a real opportunity there but that’s more for the whole committee I guess, I’ll just throw that out on the table for you to think about.
MR. HUNGATE: We’ll add it to the list, Brent.
DR. JAMES: Okay. Anyway, the idea of outcomes chain --
MR. HUNGATE: Excuse me, we have one more question, Trent Haywood had a question.
DR. HAYWOOD: Trent Haywood from CMS, I had just two small questions and maybe one bigger question that you’ll come back to related to your spectrum because we didn’t get to cover that because we didn’t have the presentation at that time. The two smaller questions, you had mentioned outcomes needs to get down to the decision level, I just want a clarification around that in a sense of if you imagine, I’m thinking like hospital readmission rates or things that you may actually measure at a higher level, at an aggregate level, whether that requires for purposes of reporting or anything of that nature you would, how that same issue around that, it needs to get down to decision level before you can find the overall utility of that measure.
DR. JAMES: Hopefully without overstating my case I’m going to be a little idealistic on this, the truth is as push comes to shove I’m going to be much more pragmatic but for purposes of the debate I feel like I kind of need to hold down one end of the debate and so I’m going to try to make the argument that we should always build data systems that support decision level and I think though that a good counter argument might be important in this whole thing. Now what it’s all going to come back to, if you go clear back to slide number, page number two, slide number one that I showed you, it’s going to relate back to that continuum of uses.
And one of the questions I’m going to ask you is is that when we design a data system is it possible to design a data system such that it can support the whole continuum, or are we designing for just one part of the continuum. It’s going to turn out that if I design correctly for the right hand end for data for learning I can generate everything up the scale. But if I design for the left hand end I can’t. And that’s going to have direct important implications on things like data burden, unfunded mandates.
Well, a beautiful article in the New England Journal of Medicine a few years ago, a physician decrying the fact that mandates from above, from government agencies, for data, significantly damaged his ability to deliver quality care and it actually makes a lot of sense when you lay it out this way, you can really understand what he was saying. And I think a lot of people experience that as they try to deal with reality of the front line.
But that’s where I’d hope to conclude a little bit, so that was your small question, what’s the big question?
DR. HAYWOOD: Well, that was the issue so I’ll wait until you talk about that little portion later about this --
DR. JAMES: I’m going to come back to it but I’m going to try to build into it a little bit --
MR. HUNGATE: One more intermediate question --
DR. JAMES: -- trying to lay out the argument so that it makes sense relative to that particular statement.
MR. HUNGATE: Before you go on Marjorie Greenberg also has a question.
DR. JAMES: Okay.
MS. GREENBERG: Also very, very interesting presentation, thank you. To go back to the slide, let me see, we now have them so I’m really impressed by how we actually have copies of these slides now but thank you to everyone who made that possible. It’s the slide after the C-section graph and your conclusion was there that more then half of all root causes turned out to be data system failures, not care delivery problems, and I’m immediately thinking well, the data were wrong, like these people didn’t really have C-sections or your usual data entry problems, what have you. And then as you described what the problem was I thought it was very interesting because if you just saw that, if you saw that statement I think the majority of people would conclude the data were wrong and of course that’s, maybe not, I mean I think your average person would and having been in this field for about 35 years that was my assumption.
But what’s wrong is some of the assumptions obviously which is again you’re right, they weren’t care delivery problems, but what this says something about, and I’d just like your reaction to sort of the way I’m framing this, it says something about the population in that area obviously, that’s what you’re talking about, so we’re actually not just talking about care delivery here, we’re talking about kind of the ecosystem of that community.
And then I think it raises, it raises questions about okay, what kind of outcomes are they getting in these home deliveries, what’s the impact of so many home deliveries where patients are having to be rushed by ambulance and how does that impact on outcomes.
And so I think it, obviously you can’t just put them out as an outlier and then criticize them, that’s absolutely true because that turns out not to be the case but it’s sort of the fundamental data 101 it seems is that only when you use information do you start learning things. And so it would seem to me that from a population health point of view this was a very interesting finding and might lead to other explorations that would relate to outcomes. Okay, I’ve gotten people here with raising their hands but that’s just sort of my reaction.
DR. JAMES: I think there’s something that everybody in the room will understand given our backgrounds, we know that no data system is every perfect, we don’t expect them to be after a while, it’s just degrees of good isn’t it, some are relatively poor and some are relatively good and it’s always around a specific purpose. My favorite quote from it, this is one of these that maybe is worth hanging on your, on my wall at least. Andrew Lang(?), famous British statistician, said, he uses statistics like a drunken man uses lampposts, for support rather then illumination. In some sense when you’re in a judgment mode you’re using them mostly for support and we as researchers, the people who’ve lived through the years with data systems, understand that their real utility is illumination, for understanding, to gain insight.
So what you’re saying is that yeah, I shouldn’t be saying bad care obstetricians are regional, what I should be saying is boy, I need to somehow reach out to those communities that are in retreat from the government and somehow get them to understand the risks that they’re taking because of their failure to appropriately use our health care system or perhaps I could even find ways to reach them with the health care system in ways that they can find positive, you see what I mean? So there I have illumination. But boy, as soon as you say illumination you’re on the learning side of the scale, but then we have data systems for which there’s a pretty good argument that nationally we also need a little accountability here too, see what I mean? So how do I handle data systems that can span the continuum? How do I create the infrastructure so that we can appropriately meet all of our needs? It’s not just, I mean sometimes you really do need to motive a hospital administrator to change their priorities, sometimes it may require legal action.
Well, yeah, so Bob can I move ahead or do --
MR. HUNGATE: We still have a couple more comments relating to this right now, Carol and Justine Carr.
MS. MCCALL: One real quick, and as long as we’re on the Deming theme and this particular slide that’s still up says outlier means the following. And I would only kind of add is that here we’re talking about outliers and what Deming is very careful to say is that simply because something is an outlier and when they are you can tend to get to root causes, I would agree. When they’re not outliers it doesn’t mean that nothing is wrong, it doesn’t mean that things cannot be changed. What he says is that you have to change the system, you will not find the needle in the haystack that is the single cause of a common cause issue. But just to be clear, there’s nothing that says simply because something is common cause that you have to like it.
DR. JAMES: Or that you have to live with it either, it’s just that you use a different management method.
MS. MCCALL: Exactly.
DR. CARR: Brent, again I echo everyone’s comments, this is incredibly illuminating, this is Justine Carr. But I want, going back to what Marjorie said, I think that the example of the C-section is illustrative of a lot of what is going on in other venues today, I think for one thing we’re all in love with a spreadsheet with numbers and percentages, or administrative data if it was captured it should have been right, and I think what we’re finding with more recent attempts to use administrative data through some of the AHRQ measures, we’re finding, our board of trustees is spending a lot of time asking our health care quality department why we have, why are we number three not number two, why are we number four, not number five. When we drill down and see are we causing atrial fibrillation in the whole population we find out no, they came in with atrial fibrillation and that it begins with understanding the universe of what’s out there and defining the population that you’re looking at so I think your example was a great example, that it’s looking at people at the hospital and also not at the hospital and I think that as I said we’re trying desperately to take things that exist but without understanding the context and the denominator we’re going down with wrong assumptions.
DR. JAMES: We tried twice to start clinical management at IHC and it failed twice. Each time it took out a senior vice president for medical affairs. The guy who recruited them here, Steve Lewis, gone over this one. Ken Richards, my next boss, gone over this one, yeah I survived, so you remember that old TV commercial where they have the middle linebacker who goes into a work environment and you see this guy walking down the hall in an office and suddenly this big huge guy just takes him out from the side, that’s what it was like. And it cost us, I don’t know, five to ten million dollars per failure.
What had happened is that we’d found physicians who were willing to try to step up and manage, but when Dave Burton and I stepped in for the third try, that’s one good thing about IHC is we don’t quit if it’s important, we did a post mortem on the first two failures and what we concluded was that we’d found people who were willing to try to manage and then we unthoughtfully assumed that the data systems we had in place were adequate for clinical management. And what we discovered that we had for data, this massive investment in data frankly, we had financial measures that were organized for facilities management and they were the wrong data for clinical management.
Well, that’s not accurate, we had about half of what we needed, it was organized wrong. But even after we got it organized along the right lines, kind of what we’re describing here, you had these gaping holes in the data that you needed for clinical management. And so our first step when we took this on was to say okay, how do we get the right data, and it was just design issue that we’ve talked about earlier, and so we started to pull the right data and guess what it worked, it’s design, aim defines the system, what data do I need to manage clinically as opposed to well, we’re back, that’s the second old joke, policeman late one night, 2:00 a.m., sees this fellow staggering around under a lamppost, goes over to help him, he’s obviously inebriated, and he says well what’s the problem sir, and he said well I lost my car keys. Well, that’s probably a good thing, he says I’ll help you find them though so you can get you home and get you into your home.
And they look and they look, can’t find it, and finally the policeman says well where did you drop them, the guy points out into the dark aways and said over there. The policeman in some frustration says well why you looking over here and he said well this is where the light is. Well, that’s what we discovered, we were hunting where the light was but the keys were out there in the dark and we needed to bring a little light over to where we wanted to look. And yes, the light is the data system and that was key for us.
I think that the assumption that we can rely upon consistently existing data systems is simply false and we need to face up to that somewhere along the line and start to design to a different purpose. In other words yeah, you can use it for some fairly limited purposes but it just doesn’t extend far enough, these data systems were generated for well, for financial payment by and large, and while they have some utility they’re going to leave gaping holes. Yeah, we found that we had about 50 percent of what we needed and then we started just to fill in the holes.
Other comments, I guess Bob?
MR. HUNGATE: Go ahead, Brent, finish up.
DR. JAMES: Well, let me finish up and let you have your meeting back here so I’ll try to go fairly quickly.
Cycle of fear, this is covered fairly well in chapter eight of the Patient Safety Data Standards Report, we know how human beings react when they’re judged, guy who wrote about it originally was William Scherkenbach, it’s got a fairly good literature behind it and again the citations are in the article. Scherkenback identifies three reactions, number one he says you rank somebody if you do poorly, first reaction is kill the messenger, also known as denial or shift the blame. Second reaction and the one I really want to focus on is filter the data, also known as game the system or the fact that looking good is often far easier then being good, very predictably human response, not just in health care but pretty much across the board.
Slide number 16, gaming the system. So for example currently on some CMS measures, tip of the hat to Steve, we still think they’re good measures for following, Steve. Currently to be in the top deciles of the counts on most CMS measures you have to have perfect performance. I’ve been looking up people with perfect performance who come through my quality training classes so I have a number of systems represented there, a number of them are posting perfect performance on things like medication reviews for patients with congestive heart failures, let’s say one of the six measures required for congestive heart failure.
What we discover is is a high rate of what you might call denominator deflation where let’s just say very narrowly defining the cases of interest, or numerator inflation, which take the form of disconnected intermediate outcomes usually. I can go into detail on this later if you like, what it means is is that you get measures that don’t reflect the underlying medicine that you’re trying to track. The reason that people are doing this is because they want to look good on the measures and frankly it’s a lot easier to kind of move to the edge of the data system then it is to actually change the performance in significant ways.
Well, what that points to is an absolutely essential need for an independent external audit. Now the best example we currently have in the country of such an independent external audit is the HEDIS measures, so I think we have a good example of how it might work but this problem turns out, I would think it’s probably endemic, just from what I’ve seen in the people that I teach in my courses, so these are people who are there studying quality for purposes of really improving care and boy we have some debates as oh you’ll get four or five different approaches to interpreting those HEDIS measures for example and it’s very, very clear that some are being rigorous from a clinical standpoint in terms of patient outcome and others are striving instead to rank well, which is an obvious response to a judgment system. The only way I know at least to get around that is some sort of an external audit system and we have good examples of those too.
Slide number 16 as we’ve pulled this thing together, redux really if aim defines the system, so when I’m at the left hand end of the extreme, comparative data for accountability for judgment, my aim is end care or accurate ranking, it turns out that in most circumstances current administrative data systems cannot rank accurately. So if Don Berwick were here we’d be having a little argument, we’ve had it many times, I would be citing the literature that demonstrate that patients at a patient level, it appears that they do not pay attention to medical outcome statistics when choosing a physician, a hospital, or a health plan, statistics are not an effective tool for helping people make those decisions, even very well educated consumers like physicians don’t appear to use them as much as they use the relationship with a personal advisor, your personal physician let’s say. What Don would say correctly is that we’re releasing these data publicly, these ranking let’s say, these judgment measures, yeah, we’re releasing them publicly to the patient but our real target audience is the profession.
The danger is is when a measurement system really targeting the profession loses credibility at a professional level. So Henry and Cliff published some frankly truly excellent risk adjusted outcomes measures from the HCVA mortality data but then we have a series of enterprising young academics who evaluated them against those standards on a case by case basis in the literature and find a positive predictive value of 25 to 40 percent and destroyed the utility of those measures within a professional community, causing HCVA to eventually retract the measures, stop doing it. You see the idea? So I think this is a real problem we need to think about, it’s not just credibility at a public level, it’s credibility at a professional level and realize when I say 25 to 40 percent positive predictive value those models are still quite significant from a purely statistical standpoint, it’s just that it’s so hard to interpret, it means that if I go to a high ranked hospital I have a 25 percent better chance of actually going to a good hospital, or a 40 percent better chance. So it does change my chances, it’s just a very hard message to deliver.
Now the trouble with judgment systems is they often rely on existing administrative claims data, it’s a matter of expediency. The trouble is is that good ranking demands very accurate data and those data systems just can’t perform to that level and are not likely to be able to perform to that level in the future either.
Well, the big issues, the ones that I really wanted to come down to, this we began to cover in the medical care article that Don led out on, one of the questions you asked is can I develop a data system that spans the whole continuum, it turns out that if you design well at the process level data for learning you can design in such a way that you can roll the data up and get, well the process management data I need, I can use it for prioritization, I can use it for pay for quality, I could actually use it then for a national accountability system. Data for learning contains accountability data when it’s properly designed, the trouble is you can’t go the other way.
If you build a measure, some sort of a high order measure, it doesn’t necessarily mean I can get down to process level. And so one of my arguments is that as we think nationally at a policy level about how we build information infrastructure we should be designing our data systems in such a way that we can use them across the entire continuum, that they can serve all needs. One of the major reasons related to clinical data burden, when I’m doing data for learning they tend to be the sorts of information that are essential for patient care delivery, so I think if integrating them into an electronic medical record, integrate them into the clinical workflow, there’s a source of information that people are generating and using routinely, well, make it convenient and easy not just for the patient care but then to roll them up across the continuum, a principle of good EMR design. Because of the integrated nature the data burden is transparent, I mean it really reduces it. When I start though to impose measures from the top, well, it’s usually seen as an unfunded mandate.
So for example, Steve, we’ve been having one heck of a debate inside IHC about whether we should pull resources out of proven performance that have been significantly improving clinical outcomes in our clinical programs so that we can look good on CMS measures, which I pretty much promise will not change clinical performance. And if I were in my most cynical mode I would argue that the CMS measures are doing tremendous damage to quality of care in the United States. Now the fact of the matter is that we won’t let that happen but realize that behind it is a policy question around data design, how do we design the infrastructure in such a way that it can function at all levels as opposed to becoming an unfunded mandate that then sucks resources out of actual quality improvement. That’s the argument, that’s the thought behind it.
Finally, one last idea, slide number 17, this is my simple one slide, massively too complex summary of the methodology that we endorsed at the Strategy Framework Board, this gives a little more detail actually, brought that home and started to use it here and it works pretty well. We designed it off the way that we design good randomized control trials, actually used Stu Polcock’s(?) book on randomized control trial design. It’s just to start with the end in mind and then work your way backwards in designing a data system for clinical outcomes.
We fundamentally endorsed the same method but it came up earlier, I think it was Marjorie who was saying something about it, could we as a country, what kind of leadership would we need to come up with a set of design tools that we could share so that as we move into a national health information infrastructure, that those systems are designed to produce the information we need for front line process management improvement, real clinical quality in other words, but at the same time can roll up and give us the information we need for accountability across the system.
Now I think in a little while the performance measures subcommittee of the current IOM effort is going to report, if you go back one slide to aim defines the system, it turns out that while judgment is difficult pay for quality sits in the middle and arguably you could do pay for quality systems that really made sense. To illustrate, the current CMS DRG system has an adjusted R squared of about .3 to .35, variability and cost per case, but it’s still sufficient to build a fairly decent payment system. I think that you could use the current systems for things like payment prioritization fairly accurately. I think that it gets really tough when you move into selection of judgment, with those kinds of systems, so I think the pay for performance effort I think really has some legs under it potentially even with what we have today. But even in those areas if we were to substantially improve our underlying infrastructure we could really make a better world for health care.
So with that let me shut up and sit down and listen a little bit and respond to comments, Bob, if I may.
MR. HUNGATE: Yes, we’d like to do that, I see Steve Jencks would like an opportunity here.
DR. JAMES: Well, sorry, Steve, and I was being pejorative, I was being to the extreme, so I just really, I do believe in what you’re doing it’s just that that’s one of the risks.
DR. JENCKS: I think that there’s probably very little different in our assessment of the situation and the risks here, and I think the really important question at a level considerably above this subcommittee in some ways but what one really has to keep in mind is how do we implement national expectations in a way which does not go down these side streets where people are sub-optimizing for particular measures and doing nothing for their systems, not really examining root causes but hiring a nurse to go around and make sure everybody’s gotten a beta blocker. It strikes me that unless we see that --
There’s a little diversion area activity going on here.
So I would argue that one of the things that we have to keep in mind, not necessarily as something that this workgroup or NCVHS can deal with directly but that it can contribute to dealing with or can undermine a little bit is how we nurture the construction of the kind of information systems on one side and the kind of thinking on the other side that you’re quite correctly arguing can be undermined by certain kinds of responses to published data.
DR. JAMES: I’ve got to agree with you, Steve, I think that is kind of the task, but in all, I think I’m right about this one, I think that one of the things that we could do is establish a good discipline for how we design those data systems, we won’t get it immediately but we could set a course for the country that we can begin to move toward. And I think that could have a very positive long term effect, make all of our lives better and substantially improve health care for those who seek our help. I think it’s a design issue in other words, I think the core issue is the design issue.
DR. JENCKS: Let me just suggest something I think we’ll probably come back to a number of times over the next two days, which is the need to have a vision of where we want to be in ten years because if we simply for example select measures on an opportunistic basis of what we can get data for and what we have good clinical consensus on, that’s not going to take us where we want to be in ten years.
DR. JAMES: I think for a group like this Quality Workgroup for the NCVHS I think what you’re saying is exactly true, I think that your mission is substantially more demanding, broader, has a lot more potential, then what some other groups may have undertaken.
MR. HUNGATE: Trent?
DR. HAYWOOD: Just to follow-up on kind of where you were at the conclusion, I’m curious to hear, because this is obviously our activity is not occurring in the vacuum and part of the push on quality is also a tension around timing, like what can you do that actually allows for a quality to have a foothold or a stake in the ground as it relates to our overall health care policy and payment structure that actually allows us to move forward towards that ten year goal that Steve may be suggesting. So I think part of that is a timing so I’m curious to hear under your thinking what you are suggesting as it relates to kind of that continuum or that spectrum about where you anticipate actually moving forward quickly, recognizing that this is not occurring in a vacuum, in other words that there’s other activities that are more focusing on kind of cost containment or cost shifting in comparison to quality.
DR. JAMES: Well, one of the things that you have to remember, that one of the things that’s informing my whole approach to this thing, how to say it, last year we took about $30 million dollars in cost of operations out of IHC by improving clinical outcome. It’s making us one mean competitor in this marketplace for which we’re suffering a bit by the way, so for me one of the primary ways that I control costs is by improving clinical outcomes for certain areas and I believe there’s huge opportunity there.
So if you were to force me as I heard you saying at the first to put a more specific point on it, now I’m going way out on a limb, I’m not at all sure that this is right, I could imagine the following, I could imagine that for example that we got funding for AHRQ would be one potential body to do it, where we started a collaborative effort to identify high priority clinical processes, we could start for example with the IOM priority committee output, and we started to put together and post the outcome chains, the conceptual flow diagrams and the resulting data systems just as an intellectual property that we share as a country, that we engaged this very specialty societies in using those tools to define them. And then that that became the foundation for how we think about the activities of an NCQA or a JCAHO or a CMS as far as that goes. It becomes a major contributor, a design element for the whole national health information infrastructure effort.
So it is infrastructure, it doesn’t maybe have direct impact. Now the truth of the matter is is that groups like NCQA are already moving in that direction but I think that we need some sort of a national leadership role and yeah, if I don’t know, AHRQ may not be the best group but I can imagine for me they’re probably the best group, that somebody would sit down with a little bit of money and start to put this together so that people could design to a common framework.
MR. HUNGATE: Okay, Carol?
MS. MCCALL: A couple of comments and then a question. As I’ve listened to all that is here I find myself asking a couple of things. One is what is our biggest point of leverage? And then the next is what’s the biggest way we can mess this up? The first one, my first thought is, I’m using you or your last slide, building an outcome system, my first thought was that one of the biggest points of leverage is this conceptual model and in the way that it really has to start by how we conceptualize the problem, how we literally begin to think about it. And I like the idea of having specific goals, maybe related to IOM or something else. But that seems like one and another one, and this is more kind of an open question, is are there some links, is there a point of leverage to some what Bob likes to call force for change. There are forces in the environment and they’re starting to move their vectors out there and can we do something that maps to the long term conceptual models but can be used with an existing framework but doesn’t do too much damage because you don’t want to do something where you start on the left side, not the right side, to your point, Brent. So are there, the question specific to you is are there links to some force for change that are underway, that we could leverage, number one, and number two, what is the biggest way that we could mess this up.
DR. JAMES: Well, so this is fairly speculative, I wonder, I know that Secretary Levitt is working fairly hard on the National Health Information Infrastructure, what we’re talking here is content of how you structure it internally, I wonder if there might be some linkage potentially there. It’s pretty clear that you can use these concepts and link it to patient safety and one idea that’s on the table is linking particularly adverse drug events and patient safety to the current FDA problems with after market repeal of drugs like Vioxx and Celebrex, things like that, so that might be a potential linkage I guess.
I think that the major effort that’s underway at CMS, JCH and NCQA to get more accountability so the whole IOM pay for performance measure, that could be a potential linkage too. So for example we’ve been in conversations with the Dartmouth crowd with Mayo, we think that you can make a really good pay for performance system based off these concepts that would be substantially tighter so there might be some leverages or linkages there.
I’ll have to honestly say I’ll have to think more, what are your thoughts about how we could screw it up? How could we make it go bad? Because I’ll have to think about that one for a bit of where the real pitfalls are.
MR. HUNGATE: We haven’t done that, we’ll have to work on that a little bit. Simon?
DR. COHN: Brent, this is Simon Cohn, I’m listening to what you’re saying and clearly I’m well aware that you’ve made a career in this area so I think my earlier comments about being an emergency physician and wanting to see some early outcomes, you may be, your comments may be obviously the other side of that. But one of the observations I guess I would make listening to you, and I’m not sure I’m right about this one which is why I’m asking, is the type of system that you’re describing is one that really need to be very flexible, only in the sense that my view is that the sort of outcomes work that you’re talking about or the design work, if not mistaken probably changes every year or two and you discover well, geez, we didn’t do it right that time, we need to revise and get new elements captured specifically or review new processes, etc., etc. I may be wrong about this one but isn’t really sort of the fundamental design principle you’re using is some sort of flexibility?
DR. JAMES: What you say Simon is largely true, interestingly though even though yeah, our good protocols change almost monthly and that effects data systems, no doubt about it and so flexibility is a key design principle. Interestingly there’s a solid core, so the way I think about it is that here I’ve got a cutting edge where the change is happening and it’s out on that cutting edge where the change is happening that I have to be pretty flexible, but back behind it I have the things that we kind of know and they tend to be fairly stable. Now over a couple of decades the cutting edge is probably going to move from area to area to area so I guess what I’d say in response is I believe that that’s a general reality that we’ll always face, that we’ll always be in a state of update and it’s a matter more of the pace of change rather then whether there will be change. So yeah, I think there’s a core.
If I think of it as a policy issue what I’d say is that the health care overseers of the country come together and they say okay guys here’s the core and we know you’re going to be tracking other things beyond that. And then we’re going to use what you learn by tracking those other things to be constantly updating the core at some reasonable rate. See the idea? Because isn’t that how any of those systems basically have to work? I’m pretty sure that that’s what Steve does for example at CMS is he thinks about these issues.
MR. HUNGATE: Go ahead, Don.
DR. DETMER: Brent, obviously what, this was a terrific talk and in the interest of trying to do some stage setting I’d like to just pull back up a little bit to 30,000 feet and make a couple of comments and see if you’re in sync with them, I think you will be. First of all I have a slide up here that I know you’ve seen, it informed the IOM committee as we did the Chasm Report, and then on the bottom there’s a degree of certainly about how to treat something and degree of agreement on how to treat it and down in the corner there’s a simple area where you do think yeah, okay, I’m dealing with diabetes, I got markers like glucose levels and such, but in that complexity zone where you’ve got a patient with 17 different diseases all interacting, good grief, you can’t figure out for sure what sort of to do.
So that’s the picture and then I think where we were coming from in that IOM report was saying gee, our system really should be safe, effective, efficient, patient centered, timely, equitable, those sets of rules to guide us when our thinking really is in that complex zone where we really don’t have data, we really don’t know for sure, but we can at least ask ourselves well in this situation what’s likely to be the safest, what’s most likely to be efficient, effective, and if you have enough flexibility in your system so that you allow people to think on the fly to self organize in an adaptive kind of way over time we may shift things down to where in fact that they are in fact more mutable and they do become not simple in a simpleton kind of way but at least more manageable.
And then I think this issue we were talking about is how do we get the infrastructures that will allow us then to kind of deal with the realities of that so that we really do have learning that occurs, and it strikes me that the IOM, excuse me, the NCVHS working group, NHII model, comes down to the NHII architecture and I’ll be talking a little bit about that later with John Halamka later this afternoon. But it seems like that vision of three interlocking records that need to be able to inform themselves literally in terms of, it’s that diagram that has the computer based personal record, the computer based patient record and the computer based population record, if the architecture of our IT systems doesn’t allow these three kinds of sets of data and records and the knowledge systems that relate to that to inform one another because they’re architecturally built separately so if that vision doesn’t allow these things to in fact cross inform we’ll never get there.
And then to finish my comments, I think what I’m hearing you say is what we really want are better decisions at a variety of levels and in a way I think the buzz on this is kind of been talking about evidence base, just in time just for me decision support, that really does allow clinician override but then tracks those, the implications of those overrides so they can be analyzed much like you’re doing, much like the New England cardiovascular group is doing, which does rely on an ITEHR environment but ultimately if this model builds out we’re going to have increasingly this electronic personal record data that clicks in mortar kind of record like Halamka said, done so beautifully I think up there, that can help this all move forward. And I think to quote another one of your comments that you didn’t make today but I think you said we need to make the ruts for the wagons to follow, on the other hand we don’t want to deny the science in this thing because we really want to build it increasingly over more and more science.
Last comment, I think you also made the point that trying to really put quality and informatics together in this is a contact sport, somebody made that comment, and I thought I heard you say administrative claims data burden, or did you say administrative blames data --
DR. JAMES: I didn’t actually say that but I wish I had have. The other piece that we need to remember in this, Don, though, is the two other pieces that were in our report just to build on what you were saying, it’s the gap in current performance so even on things where we should be fairly certain, the health care system, we’ve screwed up fairly routinely.
And then the second idea behind that is that you can prioritize it, and so I don’t know, it turned out for us to be a 90/10 rule, if I pick ten percent of our clinical processes, it was 62 out of more then 600, it was 92 percent of all of our care delivery and impacts some patients. And so I can prioritize within that if you see that idea, and so I’m going to do the big things first. But yeah.
MR. HUNGATE: Okay, Steve, you had a question?
DR. JENCKS: I really want to underline, Brent, what you said about flexibility and use the words dynamics even more. One of the things that we made a serious mistake on building our performance measurement systems was just totally misperceiving the ratio of effort to build the measures versus effort to maintain measures that are good and appropriate. So it would really, if you want, your question was how can we really screw this thing up, well, one way we can really screw it up is by creating a relatively simple minded system that lacks the flexibility to accommodate to changes in reality.
DR. JAMES: I think you’re right.
MR. HUNGATE: Another question, Carol?
MS. MCCALL: Brent, you had asked a question, unfortunately I had to step out very quickly to take a call but I’m back and apparently the question was what do I think how could we mess this up. I think --
DR. JAMES: Sounds like it’s a goal, doesn’t it.
MS. MCCALL: I do the comment is dead on, that’s one, so I won’t repeat that about how we build the system. I think another one, how we can mess this up, it has to do with kind of your continuum from the left to the right about ranking and accountability versus learning. I think the other way that we can mess this up is by focusing too much on the left right now and thinking that that is the goal because I do believe that we will find that we’ve painted ourselves into a corner and that we’ve literally got no where to go from there. So as we look at some of the forces for change, as we look at systems, as we look at trying to do things in the short term, I think we can mess things up by doing that and so there may be some conceptual models or other points of leverage that we can use to make sure that we don’t start at that end. We may do things at that but we won’t build from that end.
DR. JAMES: You know now that you say that one of my real, here’s one of my real personal worries and I don’t say it too loudly too often because I think that the effort is important to move ahead, I worry, for example, the IOM Performance Measures Committee that will recommend a series of measures, that we will place too much reliance on their accuracy, yeah we’ll be targeting the professionals through the public but what will happen is they’ll lay the measures out there, they’ll start to report, and again some entrepreneurial young academic will start to compare the measures that we’re using, that we’re forced to use really by pragmatics, the current state of the system, that will trash them hard by comparing them to some gold standard measure. And once again we’ll invest four or five years of hard effort in a particular direction where we all agree on the goal and the target but because we just weren’t rigorous enough underneath that we’ll get our legs cut out from under us.
I think that’s happened a couple of times, I think if we look back historically it’s not unprecedented at all and that’s what scares me right there and I want to make sure that we’re tight enough on this stuff, mostly at setting expectations, is telling the profession, it’s admitting up front that oh this is an imperfect measure, right, but here’s how we can still use it in a positive way even though it’s imperfect, so that we’re not leading people to believe that it’s more then it really is. I don’t know, does that make sense? That’s what really scares me about the whole thing.
MS. MCCALL: Yeah, I do, this is kind of, I’m going to tack a little bit to another direction, here’s another question. We need a lot of change, which changes, some will happen naturally, which ones will not?
DR. JAMES: This idea of having a shared model will not happen on its own, the shared discipline, the shared toolbox, theoretically they should be part of the job of the NQF but given the internal political structure I think NQF needs some help to pull it together at that level.
MR. HUNGATE: I like that answer, it gives us a reason to exist, it fits very well, it’s a gap maybe we can strive to. I think we’re going to have to call a halt to what I regard as a marvelous discussion, take a brief break and come back. Do you have any parting thoughts as I call that halt, Brent?
DR. JAMES: Just to say thanks guys for accommodating me and I’ve really enjoyed this, I can’t tell you how disappointed I am not to be back there with you.
MR. HUNGATE: We’re sorry too, we would have enjoyed it even more. But this was good enough.
DR. JAMES: Until next time.
MR. HUNGATE: Thank you. Let’s try for a ten minute break.
[Brief break.]
MR. HUNGATE: Okay, we’re up to about 10:20 in the morning now, which is okay, I’ve learned the limit of hierarchical control and believe in chaos and we’re doing that a little bit on this subject and that’s just fine. So what I want to do is ask Bill and Carol to give a quick, rather then a 25 minute maybe more like a five minute discussion here of the opportunities part of it, customers and perspectives, and then move right into the discussion on current performance measurement activities.
Agenda Item: Overview of Opportunities for Quality Initiatives - Dr. Scanlon and Ms. McCall
DR. SCANLON: Well I think that this is actually going to reflect the type of sort of struggle and deliberations that the workgroup has had and sort of both starting when we met sort of in February as well as in a number of phone conversations to try and set the agenda for today, it’s partly this issue of trying to get your mind around the concept of quality and what contribution can the NCVHS make and that I think is the primary sort of motivation for having a retreat like this without outside input to tell us in some respects your perspective as to where we best fit. Brent’s discussion like what everyone said was fantastic in terms of showing us not only the breadth of what we are dealing with but some very good insights into some of the detail that we need to be very, very concerned about.
I actually when I was on the committee before the hospital mortality statistics was a topic that was discussed at every meeting right after they had been released and I went off the committee before HCVA stopped and I’m sure the discussions went on until HCVA stopped the release of those statistics. That goes back I think to sort of Carol’s point, or Carol’s question about how can we mess up, and I think one of the things that we can mess up in, not the committee but sort of us as a society is to do something prematurely that ends up sort of being flawed and ends up creating a backlash that sets us back for a period of time. So being careful is extremely sort of important.
In that regard I think one of our central part of what our discussions has been something that Steve Jencks recommended to us to think about having a vision, a very broad vision, and then we would start from that point and work toward the more incremental steps that it will take to try to achieve that vision. But at least knowing that we are moving in the right direction and that the incremental steps are not sort of random and that you sort of end up at the end of ten years discovering you’re at the same point that you started. But having said that it’s not sort of a significant task sort of both have a vision and to sort of to have the roadmap that starts you sort of in the right direction.
In our discussions in preparing the agenda for this meeting we came up as a group, Carol was the one that actually started us along this dimension, was the idea of a matrix, thinking about sort of what are both the goals that we would like to accomplish through using sort of information and then sort of who would be the customers sort of for that. And that is the matrix that we passed out and that you all have, we don’t have an overhead of it. It’s meant to be sort of a very preliminary document, not in any sense exhaustive, and it’s not sort of meant to be exhaustive in terms of the dimensions of the division that we need to think about.
There’s sort of a third clear dimension here which is the idea of how do we use information to accomplish these goals, information that’s going to be coming from various sources, being used by various sources in different ways, and I think when we added the word facilitator we saw in part sort of the role of government is one in which to make sort of the information flows happen in ways that will be sort of most productive.
Government I think we need to recognize is poised now, or at least it seems seemingly poised, and more enthusiastic then in the past, about trying to use information to influence health care and health. Maybe it’s a sense not of sort of capacity but of frustration, that there’s been so little progress in terms of trying to make forward movement in terms of either improving quality, controlling cost, that we need to do something so the ideas of sort of paying for performance, quality reporting, websites that sort of provide comparative information, etc., are all being sort of pushed but there is that issue of are we sort of on solid ground as we move forward in all of these dimensions or are we creating sort of risks that we need to be concerned about. And then sort of there’s the second sort of issue which is how can we move these processes in a positive way, more rapidly to achieve the kinds of outcomes that we would like to accomplish.
There’s a variety of dimensions I think that we can go into and decide sort of this issue of what are the mechanisms that it will take to implement sort of a vision and what would be the kinds of steps that we need to have to, or to take to accomplish that division. There’s also going to be an issue of sort of how do persuade people that this is the right thing to do. In other parts of the committee I’ve heard the term business case used a lot with respect sort of to IT, we need to use it in terms, in the context of quality as well, sort of why should these various actors be sort of willing to be participants in our vision or the societal vision to try and improve quality through sort of information and that’s in some instances not going to be a trivial task that we’re going to have to undertake.
So I’m going to stop because as Bob said we would like to get your input much more then sort of our speaking to you but I’m going to turn it over to Carol sort of for her thoughts on this process.
MS. MCCALL: A couple of additional comments, just to kind of set some context. Everybody at least that’s here in the room should have this grid. And the grid was really meant to help guide the discussion and I think that there’s some other things that we can use now based on the conversation this morning as well as the background paper that Bob had sent out. So just to kind of add to the context that I think will be vital.
The first is aim and I think that what Brent shared with us this morning does a very good job around the discussion of aim, it’s not all kind of boiled into kind of the WordPerfect version but I think it was a very delightful discussion around aim.
The second is I’ve heard people talk about a need for a vision and I think as a starting point Bob has some things in his paper and it’s actually on page two and he talks not only about the charge of NCVHS but also the charge of the working group and makes reference to some of David Brailer’s comments. In particular at the bottom he talks about goals three and four, to personalize care and to improve population health. Now on our grid we’ve actually expanded health to include individual health, population health and system health, and those are broad goals, they haven’t been made specific yet so we have an aim and we have some broad goals.
If you then turn to page three of the document, the next page, it talks about the meeting process and about halfway through the page Bob has a great term that he calls forces for change and a lot of the forces for change in the system, some of them are beginning to move, some for better, some for worse, but they are beginning to move. And they are mechanisms, they are means to an end, listed here are things like pay for performance, consumer health care, patient safety and the like.
And underneath there what we said is that for these forces we really want to understand some key things as well as as we understand them pay particular attention to a couple. As we understand the forces, in particular given the workgroup’s charge, what is their reliance on the existence of meaningful information and also what is their ability to generate meaningful information about quality. And if Brent were still on the line he’d probably say both and yet it will emerge out of that evolution.
And then as we understand those things we want to pay attention, again with our charge and role in mind, what changes need to occur to bring things about, what changes won’t happen naturally, and that’s where we can have some impact. And so those become the mechanisms and the things we need to know and again back to this grid or this matrix was just meant to be a guide, to understand some of the higher level goals, some of the different constituencies that are involved, and as we understand the mechanisms and the specific answers to questions help us create enough information from the next day so that we can then put forth what is a meaningful and actionable agenda for this group.
So that’s a little bit more thought into the background about how these documents have come together.
MR. HUNGATE: Okay, Marjorie.
MS. GREENBERG: I might just comment that they really map nicely against the concentric circles that Don showed I think because, well, the individual health, could be both the provider and the personal record but obviously the population health and then the system health which we talked a bit about trying to capture system in that vision and I think, but it is part of it and so I think this is quite nice. Of course I’m always happy to see that even population health is not only here but it’s the biggest section.
MR. HUNGATE: Okay, Justine, I’ll turn the next discussion over to you.
Agenda Item: Panel 1: Discussion on Current Performance Measurement Activities - Dr. Carr, Moderator
DR. CARR: Well, thanks Bob, and I want to also be cognizant of Simon’s mandate that what we want to do is explore the possibilities but with an eye towards achieving a plan that has tangible deliverables in a timely fashion. And I think as we are educating ourselves, reading the vast growing array of information available to us, it’s impressive that so much is going on and so I think two ways to think about what this workgroup might want to do would be one, is understand as this vision is evolving nationally are there gaps that we’re seeing or hearing that we would want to help identify, or is there some dimension that has not yet been addressed that we would want to bring forward. I think those were two of the things we talked about in the early stages.
But I think what I’d like to do now is ask each of our invited guests in turn to make some opening comments and maybe set the stage for the kinds of things that we want to further develop in our discussion in the next hour or two and so I think I would look to Richard Klein to begin.
Agenda Item: Panel 1: Discussion on Current Performance Measurement Activities - Mr. Klein
MR. KLEIN: I’m Richard Klein from CDC’s National Center for Health Statistics. Our particular office at NCHS is involved with supplying to the department down here for its statistical support on its health promotion/disease prevention initiatives. Many of those, particularly Healthy People, we’ve been through several iterations of Healthy People, involve the selection of indicators, we’ve been through massive discussions on developing indicators, criteria for indicators. For Healthy People not all the indicators are directly related to quality, we have a number of different types, but certainly a lot of the quality indicators are there and in fact have been used by others that Ernie will talk about in a minute for other projects.
Also here today, didn’t know they would be here, are the colleagues down at the department level that actually run the Healthy People initiatives so maybe some of the issues you have may relate to them because they actually developed the consensus process by which all this happens which I imagine it’s something that you’re interested in.
DR. CARR: Thank you. Ernie.
Agenda Item: Panel 1: Discussion on Current Performance Measurement Activities - Mr. Moy
MR. MOY: I wanted to frame this first, I think Dr. Clancy is going to talk tomorrow about the many wonderful things that AHRQ does in the area of quality measurement and so I’m going to restrict my comments to what I do on the National Health Care Disparities Report and the National Health Care Quality Report. I think in some ways my job is a lot easier then that of this committee because we have a very, very focal charge, this is we produce these reports in response to a Congressional mandate and very specifically our primary audience then is Congress, national policy makers. And so they’re asking us to provide them with a snapshot of the status of quality of care and disparities in care for themselves, for that very specific audience.
In producing this report I think we’ve encountered a number of different gaps and it’s one of these things well we wish that someone could help fill these gaps, sometimes this is other parts of AHRQ that need to fill these gaps but other times there are other areas perhaps that we don’t fill those gaps necessarily as well and maybe part of the purpose of today’s conversation is going to be to talk about some of those gaps I gather. So I will talk about some of them, again this is from my context in writing these two reports as primarily a user of
data measures and of quality data, from that perspective.
But first of all I don’t want to have to deal with all the developmental measures and stuff that is not well formulated and so one of the very simple areas that I found difficulties is when I’m trying to pick out well established consensus based measures, that in some areas like cardiology there are many, many measures, in other areas we’ve identified there are relatively few measures and there’s really not a whole lot that we can say. And I’ll say that when the priority population conditions came out that was particularly emphasized because some of those areas we think that the science behind some of the measurement issues are very, very poorly developed as it relates to health care, which is again the focus of our report.
So for instance having obesity identified as a priority population and we don’t have a lot of science about what particular kinds of health care would impact on that makes it difficult for us to then incorporate this priority population into our report although we do work on it. It forces us to do things like rely upon well, some groups recommend this particular service for obesity even though we don’t think there’s a lot of science supporting, there’s not enough science for us to be able to document quantitatively exactly what are they the potential benefits and costs for instance for a particular service but we try to include it to address this particular issue but certainly development of measures with more scientifically rigorous backing would help us.
The other issue that’s related is even when we have identified good measures for particular conditions often there are an absence of national data and so for instance in the areas of HIV and mental health substance abuse we encounter a lot of problems, we think that there are some processes of care that have been fairly well validated but we don’t have national data, we have regional data, state data, local data, but not truly national data which we think is a perspective that Congress has an interest in.
Similarly along those lines when we look across the IOM dimensions of quality which we use as our basic model we find that safety and effectiveness are very well fleshed out in terms of measures but the other ones, the tail end of that list are not as well fleshed out so timeliness, patient centered, and efficiency are areas where we think there are major measurement gaps and data gaps and we wish somebody would kind of spur that movement along a little further to develop measures and data to support those particular areas.
I also wanted to bring in the disparities report because I think that disparities is a relevant issue to quality, IOM does identify equity as one of the core dimensions of quality, and so we think it’s important. We think it’s important both for the national policy makers to view these two as linked topical areas, and then we also think it’s probably important for other users of this to view them as linked topical areas as well, if nothing else for instance as we start to increasingly have public reporting and pay for performance measures what is the impact of different providers and facilities providing to different populations, are we creating a level playing field.
So if you’re a facility that’s taking care of populations that we know from national data are much less likely to achieve these high quality outcomes or achieve these processes, can we hold them to those same standard or do we need to adjust in some way for different facilities and providers taking care of different populations. And so I think disparities is relevant to the quality conversation and again I would encourage this group to look and consider the incorporation of disparities information into the quality measurement activities they propose.
If you do so though you open up a whole bag of worms in terms of other issues, how to collect these demographic, what kind of demographic information to collect, how to collect it, it’s not standardized certainly once you get beyond the federal level, etc., so it does open up new cans of worms there. So I think those are the major gaps that we’ve identified in the development of the reports, measurement data and then disparities information.
DR. CARR: Thank you and I know one of our, because we are a workgroup of the populations committee that synchrony of those agendas I think is important to take note of as we’re going forward today.
Dr. Roski.
Agenda Item: Panel 1: Discussion on Current Performance Measurement Activities - Dr. Roski
DR. ROSKI: First of all let me apologize, I will need to step out around noon for a phone call but I’ll be back.
There are a couple of things that sort of keep us busy at NCQA these days, my role at NCQA is as vice president of performance measurement to ensure that we enrich the data set as much as possible by introducing new measures, ensuring that the existing measures are maintained, meaning that they are adapted to the most recent clinical evidence that is available, and also to conduct research with respect to performance measurement, what new areas we could explore and then contract work that we do actually for CMS, for AHRQ, for a number of states with respect to performance measurement.
A couple of comments, one of the issues that, and I will confess that I’m not that familiar with the task of your committee, but one of the issues that sort of struck me is the theme that we discussed earlier today and that is to be clear about the aim and for what purposes is information supposed to be provided, is it information for patient decision making and selection, is it information for providers to make decisions about care in the care delivery process, is it information for payers to judge who or who to not to reimburse or to differentially reimburse or who to include in their networks for example. Those would be questions that I think would be critical to determine where you want to set the focus for your activities and I’m thinking of the continuum that we saw earlier today ranging from accountability to information for learning that you may want to be clear about.
In terms of measure development in some ways, I read somewhere in the documents preparing us for this meeting, is it appropriate to switch maybe a focus from quality to performance and certainly in our experience quality doesn’t seem to excite very many people these days. We’re getting our shins kicked at NCQA that we’re not providing relevant information for decision making. What that basically means is we’re not providing information about the value that’s being rendered in the health care system. Or to say it differently what does it cost to produce quality and what is NCQA or what is the system doing to address the suspected rampant waste in the system that may be partially responsible for the health care cost problems that we have in this country.
So we’ve focused very extensively over the last I would say 12 to 15 months on the problems, getting away from defining quality as a problem of under use, meaning are patients getting all the things they should be getting, to a focus on are patients getting too many things that they really don’t need, or are they getting the wrong things all together that are dangerous to their health. So those could be problems of over use and misuse.
Along with that what we’ve been trying to understand is a system that is trying to develop or trying to produce optimal results, what would that need to look like. And one conclusion that we’ve reached is that it truly needs to be a system as opposed to various care processes rendered by various providers that may or may not talk to one another and may not have all the information that’s germane to rendering appropriate treatment. So performance measures that would focus on the systemness of the system and trying to determine to what extent care systems are present seems from our perspective as very important particularly as we’re now galloping I guess into the age of trying to assess the performance of individual providers. And there are certainly in our minds questions about under what circumstances that actually makes any sense and under what circumstances could it actually set us back from focusing on individuals as opposed to the many individuals that are usually involved under care, typically of complex patients anyway.
Coming back to the issue of how do we identify value in the health care system, clearly the purchaser community, the health plan community, is very much, CMS I know is very much aware of the work for example that’s being conducted by Fisher, Wendberg(?), et al, if you will basically demonstrating that there either is little relationship between the resource intensity we expend on treating people and the results we get from that, meaning that you can get apparently to the same results for a whole lot less money in Minnesota then you can in Florida. Now the reasons for that are not totally clear but it certainly has the employers and payers on the barricades demanding that information about cost and resource intensity is being considered as we make determinations about the overall performance and quality of a system.
Now to that end not only are we getting our shins kicked by purchasers trying to, demanding that we develop performance measures in that area and do so quickly, as we’re trying to move this forward we’re getting also our shins kicked from the provider community in terms of not being quite comfortable with this concept, usually pointed to methodological and scientific challenges to actually come to some definitions here. So frankly what a real contribution might be here is to at least propose a taxonomy of what does value actually consist of in the health care system, is it simply quality divided by its cost or how should we look at that and what should the American health care system, how should the American health care system who probably cannot close its eyes to the cost inflation in the system define the value and how can we then operationalize that in terms of performance measures.
Talking about performance measures, we have a fair number of performance measures that are for example part of the NCQA HEDIS system or the JCAHO system and some measures that CMS has put forward, they’ve been endorsed by NQF. Now certainly a question for us is ultimately we believe that performance measures or developing performance measures is a public good. Now AHRQ has estimated that it may take in excess of $200,000 dollars to develop a truly well specified performance measure that can be implemented. Now if we imagine how many more performance measures we need one quickly comes to the question of well who exactly is paying for that. Frankly most of the cost of the HEDIS measure development is borne by health plans who are going through accreditation and where those fees are essentially subsidizing the expansion of that measurement set.
Now in our estimation that’s probably not a fair proposition to a health plan because ultimately they will pass it on to the payers and ultimately to the consumers so it’s a vexing question to us, who exactly has responsibility for that, in particular as organizations such as NQF are trying to endorse performance measures as national consensus standards who actually has the motivation to deliver performance measures to NQF so they can be “integrated” into the public domain yet expend an enormous amount of resources actually developing them, not being able to reap any benefits from that. So there’s a real financing question that I think we should be dealing with.
Also with respect to measure development, I think what we haven’t done a great deal of is, or what we may need to focus on in the future with all respect to what we heard earlier today, a demand from an overarching measures of quality, in other words as people are trying to make decisions about where to seek care there is a question about can anybody really understand 60 different HEDIS measures about any particular health plan and then based on that make a decision, there needs to be some way to simplify the story if we want people to be actual consumers of performance information. So for example if I apply the mother test and I would ask my mother to make a health plan choice based on the excellent HEDIS performance measures that we’ve been working on for so many years she couldn’t do it because it’s not understandable frankly to the normal mortal person.
And consultants frankly are able to seek engagements with employers to help interpret what all of this may mean and then translate it into ROIs and so forth, so there is a need in my mind for a focus on composite measures, figuring out how can we simplify the story so that we potentially also could get off these performance measures of one diagnosis at a time. And that’s not withstanding what we heard earlier today in terms of having outcome chains and so forth.
In addition one thing that we’re very cognizant of is the notion of what it takes to actually implement performance measures, I don’t know if any of you have had the unfortunate history of having been involved in HEDIS data collection, it’s a tremendously complicated highly, highly specified process. And to some extent I’m a little leery that some people may think that with the endorsement of performance measures by NQF we have the problem of performance measurement solved. I think the much bigger problem is how do you actually go from consensus to actual implementation.
Now luckily virtually all of the performance measures so far identified by NQF as national consensus standards are already or have been for many years implemented by one organization or another, be it JCAHO or now the HEDIS measures are going through this process. So the other questions may not be as relevant but as for example we’re talking about the vast array of ambulatory care for example as rendered in fee for service settings who actually has any idea about how we would implement any kind of system that would give us valid and reliable information that is audited, I think that’s a very important point that we heard today, that we can have any trust in.
So that may be something that this committee can help provide some insight into. And I think with that I’ll end my comments.
DR. CARR: Has Dr. Jencks returned?
MR. HUNGATE: I don’t think so, I think Brent has probably signed, oh, Trent, and Trent is the one who is expected, Trent Haywood is expected, he had a phone call but he expects to be back. The other jobs go on.
DR. CARR: Well, again, just trying to structure where is our area of focus, I mean independent of our matrix I’m hearing are we thinking about the work of today or are we going to focus on the work of tomorrow, are we going to focus at the micro level as we heard from Brent James this morning or are we going to be looking at the macro level in terms of populations, Healthy People, are we focusing on the cost in dollars to make this happen or the dollars saved if we make whatever this is happen, also the cost of resources, the implementation of these measures the number of people whose, number of FTEs added to every payroll to make this happen, are we talking about end points, are we talking about processes, and in terms of measures are we focusing on their predictive value, their availability or their alignment with any utility and then who is the audience. I think we heard a tremendous representation of all of the dimensions of this and back to Simon’s admonition that this committee has to have a discipline of looking at, of defining what we’re going to do, making it achievable and timely.
I’m looking for comment on this, Marjorie, yeah.
MS. GREENBERG: I agree with everything you said and thank the presenters, I just wanted to ask Richard as a colleague and knowing what a huge job it is to track these measures and I think I was finding the email from Julia back a few months ago, she mentioned that NCHS under your leadership monitors over 467 measures of health care status and health care quality. And that really makes me think about what our last speaker mentioned about how when you’re dealing with all these measures do you simplify it or do you tell a story and this I know has been a long time debate in the Healthy People arena and I think as each iteration maybe tries to address this by having some sort of overarching goals, I think more in the last iteration like reducing disparities or whatever, but I just wondered if Richard you could comment on that and what your experience in working with such a large measure set has, I mean I’d be interested also in some of the measurement challenges that you think might be relevant because you’ve seen them all I think, but also just how you, some way to bring this all together to make it meaningful or are people getting drowned in these different measures.
MR. KLEIN: Yeah, it’s been a challenge for us. Healthy People has been around since 1979, introduced under the end of the Carter Administration, it’s been through until then a number of different types of political environments and social environments. The measures tend to grow, this is one thing we noticed about measure sets because what happens to these processes is at the next iteration you want to be more inclusive so you include more people and you include, we ended up with ten measures being proposed for sleep apnea the last time because we had the sleep people there. So the people you include the more measure proposals you’ll get.
Healthy People is also misleading to call it 467 objectives, that’s how we get away with a minimal number. What people get around that is have a number of sub measures all of which are individual performance type measures, A, B, C, D, so we’re up to when it was published in 2000 we had over 800 measures, we’re going through a mid-decade evaluation right now, we call it the Mid Course Review, where we can do some tweaking, we can propose to drop measures which actually has happened, I want to talk about that of how we dealt with measures that didn’t have data. Or some of the ones that didn’t have data before now have come in with sub measures, we’re probably up to 1,000.
We try to deal with at the beginning of Healthy People 2010 which was at the end of last decade and the beginning of this one, we tried to deal directly with this issue of needing a lot of measures because you need to cover a large waterfront, needed to cover all of public health, we needed to have the buy in and inclusion of all the department which means you have to have all the little niches but at the same time can we come up with a summary way of discussing health so that the pass the mother test that he just talked about and pass the public test because also an audience at least of an effort like Healthy People has to be the general public, it can’t just be health professionals.
So we’ve had two different approaches and they’ve had various degrees of success, the first that came out with Healthy People was another committee like this that we got together and say we had the charge from Dr. Satcher(?) who was the assistant secretary for health at that time, to give me a set of measures I can put on two hands that would be, we ended up calling them leading health indicators. So some summary part of this set that would link back to the set that would be a subset of measures but that would be kind of key indicators of health, we got the IOM involved, we went through a fairly major process, I guess it went on a couple of years, did come up with ten but in the tradition of Healthy People a lot of those ten have subparts and we ended up with 22 individual measures.
These we made a pretty big deal about at the release of Healthy People, implied that this is what we were going to track and that is what people, we’ll track the 467 but we could kind of focus on these, it would give the people that want to match up to an indicator process, in our case the audiences tends to be state and local health departments and political entities that would like to then benchmark to the public well now they have a small set, of course a lot of these weren’t available at that level.
We decided to focus not on outcome measures for these, that’s what the result of this discussion was, we were looking for things that could be effected in the short term so they tend to be more in what I think we’d be calling performance type measures, they’re risk factors, they’re services, things like obesity, physical activity, and some delivery of services. So that the idea was you could move the bar so there is no leading health indicator on lung cancer but there’s one on smoking for example, things that could change in the short term.
I don’t know to what extent the process that we ran into would be relevant here but this was released in 2000 to great fanfare and in a few months we had a change of administration and being everybody’s subject to these things and for one reason or another the next group that came in wasn’t enthralled with the leading health indicators, just didn’t like it. And so they’ve been totally deemphasized for five years now, we were set up with our database and everything to really track these and they’ve not been a major focus of the department and every department comes in, they have a right to decide what they’re going to focus on, they said we’re not going to focus on this.
So I don’t know to what extent you have some hierarchy over top that says that you can come up with, it’s a consensus process, this was a major process involving the IOM that just got not political support. So we’re set with this set of leading health indicators, we continue to track them, they’re a selection on the major database that we use to track these available to the public, but they’re generally not used and when we come out with our report on the Mid Course Evaluation of Healthy People and how we’re doing, if there’s any mention of these at all it will be very small. So that’s the first thing, so we did try to come up with a summary set but we probably had limited success.
The other thing that we’re doing that really goes back to Healthy People 2000 which tried to develop last decade and the decade of the ‘90s, and that’s to try to come up with single or a very small set of summary measures of population health that could describe all 500,000 measures that talk about health in the aggregate, these tend to be based on what’s been around for a while, measurements, qualities, dollies(?), hallies(?), everybody familiar with what I’m talking about, various adjusted life years, they’re measures that take mortality and morbidity and come up with a single number. We did some work on that last decade, this decade we decided to make it a major research focus and we had somebody spend literally full time on this, hired a researcher and group of people to look at this for Healthy People 2010 to come up with a set of summary measures.
The issue with this is the more aggregated you get the more difficult it is to interpret and understand the measures, the less consensus, there’s no consensus, this is a world wide, there’s been conferences all over, they have them every year in Europe and everybody’s got their favorite measure and some want disability adjusted life years, there’s a tremendous issue on weighting, you have to weight every health state. What you do is essentially take mortality and then decrement that by deciding whether somebody’s healthy or not given whatever parameters you decide and we’ve picked, we’ve now settled on four measures that we’re going to use for this Mid Course Healthy People evaluation, if I can do them off the top of my head, one is the years of life without a major chronic conditions, one is the years of life in excellent, very good or excellent health. One that we’re still playing with would have to do with the years of life without a major risk factor and I think we have smoking, physical activity and obesity in there. And then there’s one more that will come to me, I wasn’t prepared to talk about that.
The point is that we have these measures, we have some data that we get from health surveys, they’re going to tend to change very slowly --
MS. MCCALL: Question, I find myself becoming just really enthralled in the process that you’re describing and so many things that I could ask but the question on the tip of my tongue right now is you have all these metrics and you’ve buffed them and polished them and they’ve come and they’ve gone, what do you do with them? What are they for?
MR. KLEIN: Well, I could say that from my perspective that’s out of my range but we obviously can’t say that, might let the folks from the department speak to it. I think the way we look at it is that the whole aggregate of measures is basically a resource for others to benchmark to, so what you have is, when Healthy People 2010, when they initiated the process, this was in about 1998, everybody in the department was convinced that we were going to cut down from the 300 measures we had in 2000, nobody could deal with 300. And when we went out to, we had a number of public meetings, and the general consensus was nobody wanted to cut down, even people that thought it was too much, the folks that felt they’d like to pick and choose so I think as a way of like a menu is one of the real values of all these measures.
To make sense of them in an aggregate like we’re trying to do for this Mid Course Review is very difficult and you end up with the kinds of things, and Ernie has had the same issues, where we end up counting. Well, 50 of them went up and 40 of them went down so I guess we’re doing okay, without considering, weighting, are they all considered equal. I think the real value in the process is the political consensus that it engenders, everybody comes together and agrees on this and to the extent that that filters down into some kind of other process, which I believe with no evidence that it does, I think that’s the main value.
One thing we should talk about with Healthy People that’s different I believe then the other indicator processes is the fact that it did set a target and that sets you up for all kinds of issues.
MR. HUNGATE: I went to the kick off meetings for the Healthy People 2010 and there were two members of NCVHS that were on the panel that announced it, John Lumpkin and Dan Friedman, and Dan’s from Massachusetts, he and I have become good friends through NCVHS and I’m still learning about population health. But I’ve always thought of Healthy People and the program as the core piece of the system architecture of health measurement, that it’s a kind of a national profile of what’s the health of the nation. It could be used as objectives, you’ve put in some suggested improvements that might be effected, there’s no assigned manager to do that, nor in this society would there be. But I would really like it if my family practice physician had that information on his panel of patients, how are these people doing, what are the issues that need to be grappled with. So I wonder how what goes on in that program links to what happens in the National Health Care Quality Report, how it links to what NCQA does, what is the synergy between those. I think that it’s a method of communicating the population writ large, which has some history and some trends, and makes people think about the things that they live with.
DR. CARR: I’m going to welcome back Trent and ask him to speak, we’ve just been saying a couple works about what are the contemporaneous issues. Again, as I’m listening to this I’m struck by the vast complexity at the detail level and I’m wondering, this committee, four meetings a year, where we can fit in. I think one of the things I’d like to put out there is as we are looking at the development of the electronic health record, and I’ve been emailing back and forth with Simon and others about the safety report that came out last year from IOM that talks about AHRQ developing a taxonomy of safety that can fit into the electronic health record.
I’m looking for sort of tangible processes that we can move forward so that something will happen. I’m struck by the greatest minds in the country are struggling with the metrics, the reliability, the predictive value, the completeness of them, and I’m awed by the expertise of that group and I’m thinking this group can’t enhance that, this group can emphasize something that needs looking at and so, because we’ll be hearing later today about the electronic health record, the vision, I’m wondering if we want to interface with how will we know about quality when the electronic health record is here, what are the elements.
MS. MCCALL: Just to kind of extend those thoughts a little bit further, but back to Rich what you were talking about, in your story of what you went through I was struck by a couple of things. One is the paradigm that we’re carrying around, the mental model of consensus, that we must reach agreement on metrics and it will be the union of that set, before we can proceed. And so I see that running through a lot of things that we do and yet one of the things that we may want to consider back to the discussion about a shared intellectual property is that it’s not about consensus on everything, it’s consensus as Don had in his slide on maybe a critical few that can begin to move us forward. So I think we’re letting the perfect be the enemy of the good.
The other thought that I had was that a lot of what your purpose of those metrics right now is for learning, as you said it’s a resource, how can we use it as a resource to learn. And yet on specific metrics we do have to create I guess what I would call a playground for people to actually play with new metrics. Ernest you made a comment about obesity, it becomes this kind of Catch-22, I don’t know and therefore I can’t move, I can’t put a CBA on it so how will we learn. So people may want to try different things outside, around the edges of where we do have agreement.
And the infrastructure and the metrics and the intellectual property have to have mechanism, technical and intellectual that allow us to play around the edges so that we can expand and change. But that one of the key things may be to create that center core of things where it is clear, the light is bright and our keys happen to be under it and isn’t that a good thing. So that may be a place where we can focus and try to find those critical few places.
The other thing would then be to link those, the metric is not the goal, it’s trying to change something is the goal. Then take those metrics and weave them back into a matrix of constituencies, of what do they serve, what higher purpose do they serve, and what levers, pay for performance or whatever can they get woven into the fabric of because that’s how they live, that’s how they become alive.
DR. CARR: Thank you, Carol, very good point. Trent.
Agenda Item: Panel 1: Discussion on Current Performance Measurement Activities - Dr. Haywood
DR. HAYWOOD: I apologize, I wasn’t able to benefit from what had been said previously on this activity. What I will do is just make this informal then, pretty much hopefully that we can have some dialogue about where we are currently with our activities at CMS as well as the larger enterprise on kind of quality measurement.
I think it’s safe to say pretty much where we are is probably better then where we were two years ago but it’s not necessarily where we need to be it’s quite clear, keeping in mind that even though where primarily a lot of the time the focus is on the measures I know internally we talk about the measures are not in and of themselves the ends by rather a means to which where we’re trying to actually improve overall quality. And as you may recall this process started under then Secretary Thompson but has continued under Secretary Leavitt, under the notion of really trying to improve quality and using the powers that be, particularly with the Medicare and Medicaid program to try to move forward on what we were calling quality initiatives and continue to call our quality initiatives.
And we started off as you may recall looking at nursing home quality initiatives. One of the reasons why we started there was because we could start there to be honest, meaning that there were some that had concerns about the quality of nursing home activity and so we had encouragement from external stakeholders as well as nursing home practitioners about really trying to address that public perception around the quality of services that are provided in those institutions. And to be quite honest and in comparison to other health care settings decision making at least from the consumers in that particular marketplace is probably more driven by risk adverse decision making then in other settings. In other words that they’re more likely to actually look at that information and utilize it and have concerns about quality in comparison to maybe at an individual physician office or at a hospital setting where primarily the brand in and of itself lends itself to conclude that there is high quality there.
And so that’s where we started, in addition the key point that came to mind that we’ve always talked about is kind of these building blocks as we move forward recognizing that you need measures, you need the practitioners or the providers to be able to have some data collection tool. You need to be able to accept the data so you need some data submission process that allows for the submission of those performance measures.
And then lastly on that block we ultimately decided based upon that whether or not we’re going to use it for purposes of quality improvement, internal quality improvement purposes, for public reporting or accountability, for some other activity linking it to payment, whether it be through our demonstrations or some other form of payment mechanism.
So we’ve always talked about these kind of the building blocks and that was another reason why starting out with nursing home and home health, there was pros to cons and that meaning that you already had a process, whether it be the minimum dataset in nursing homes or oasis in home health that by and large the community in those settings were familiar with the dataset, they knew the strengths and limitations to be honest of the dataset and deriving measures from there allowed us another way to actually jump start the process to get going quickly out the box again recognizing that if you’re talking about ideal, that may not be the ideal way of actually constructing performance measures by limiting the use to those where you actually have data systems already in place.
I think from our perspective it actually has shown to be beneficial, both in terms of the reception and getting visibility around the notion of actually providing information to consumers for nursing home quality improvement, home health quality improvement, as well as when I looked at some of our data from the quality improvement organizations, the QIOs are working with those institutions, that we are actually starting to see some improvement around these performance metrics in those settings.
And then we move to the big areas around hospitals and now, I should have background, ESRD is always the forgotten child there, ESRD was long ahead of the curve to be honest, before we even went down this route had actually been providing that information for quite some time and so we had 16 clinical performance measures under ESRD for purposes of quality improvement as well as three metrics that we use on adequacy, anemia management, and survival, patient survival ratios that we have publicly provided as well. So ESRD had already broken through a lot of this ground before we even moved to nursing home and home health.
But we always recognize that I think hospitals and physician offices were going to be some of the bigger challenges, particularly as it related to the issue around data collection, being able to actually have a data collection vehicle that would actually allow for that process to unfold, for us to be able to have our infrastructure in place to actually be able to receive that information. And then finally linking it to either public reporting or pay for performance or something along those nature.
Earlier on it was beneficial for us to actually have had nursing home and home health first because those that were watching the marketplace new where we were heading so hospital associations actually took the lead and came forth and said that given where we currently are they recognize that we were going to move toward the hospital industry and they wanted to do it actually on a collegial fashion so instead of a mandatory process, which occurred in the other settings, it was all voluntary on the hospital side. And so we worked with them and what we ended up doing as you may recall is look toward the National Quality Forum as the only consensus development process that is currently out there, looked at their 39 measures on the hospital side and said how can we actually get going with implementation.
So we started out with ten of those measures to just get going and what has been benefited over a little over two and a half years around that process is really working on primarily issues around getting at the individual hospital level if you will to help understand around what the processes are, build out our infrastructure, as well as making certain this still continues to be quality improvement and not just a matter of accountability for the purposes of being able to rank individual institutions. And that’s been quite beneficial because we continue to have support from the hospital stakeholders on this activity as well as collectively moving forward and building that infrastructure, whether it be the infrastructure of the National Quality Forum, the infrastructure around what we call the hospital quality alliance whereby we spent a lot of our times discussing how we’re phasing things in, how we’re going to operationalize and particularly how are you going to message it so that the communication doesn’t actually lead to the detriment of the overall activity. And so there’s a substantial portion of our activity that’s as you can always imagine just around communication and messaging and making certain all the stakeholders are working collectively.
And then lastly actually what happened on the hospital side is that with the Medicare Modernization Act obviously it linked that public reporting that we already had out there to a financial incentive, so to their market basket update. And while all that was occurring we were always behind the scenes working closely with some of the physician organizations, the AMA Physician Consortium as an example, NCQA and others, to really start working on the physician side of the house, starting to get at the ambulatory side, looking at ways in which we can actually work with physician groups recognizing that they’ve had even more of a fragmented system in comparison to the hospital setting and the other setting.
And so we’ve developed several different activities around there, one of them was the Doctor’s Office Quality Project where we looked at some of the measures that were already, that we thought were already out there, some of NCQA’s measures, some of the HEDIS measures, as well as what the AMA was doing on the physician consortium just around the notion of really can we actually at kind of the mom and pop doctor’s offices out there in three different states, New York, Iowa and California, really start to get at how tough it is or is not to be able to collect information, making certain that they even understood what we meant by trying to collect the information, getting a sense of the lessons learned about trying to be able to do that such as how many physicians even know who their patients are, meaning that they have patient registries, that they have a sense of how many diabetics they’re treating or how many heart failure patients they’re treating, whether or not they’re on certain prescribed medications or not. All that has continued to occur on the Doctor’s Office Quality Project.
What we also did under that project is not only look at clinical, quality of clinical effectiveness measures, we started to stick our toes in the water in the other two grounds in a sense of structural or systems measures so there’s office system surveys that we’re looking at as well to get a sense of what is the system of care in which that quality, those services are being provided so there’s an office system survey that’s part of that Doctor’s Office Quality Project.
And thirdly it was going to be a precursor to ambulatory CAPS(?) but as with all things it takes longer then we anticipate for implementation purposes so by and large now they’re actually using what we think will ultimately become ambulatory CAPS and so that’s the third prong there. So I know we’ve spent some time talking about clinical measures but there’s the patient’s perspective of care and some structural measures also that we’re looking towards.
And so that’s the Doctor’s Office Quality Project, more recently we’ve also had a lot of activity underway with health information technology and provide some of that structural system support. One was actually Doctor’s Office Quality Information Technology, or DOQIT, or DOQ IT, whereby again working with our quality improvement organizations we really were trying to look at ways in which we could actually look at quality but in the ways of that structure or that system of support and how we can actually incentivize doctors to actually adopt health information technology not solely for the purpose of saying that they’ve adopted it but really for decision support and being able to start to tease out as we move along this continuum of being able to make decisions based upon data, that they’re also able to not only have data available for their purpose of decision support but maybe upstream as well for purchasers or others that may want that information.
And then with the passage of the Medicare Modernization Act there’s even more incentives, there’s some demonstration authority whereby we’ll actually be working with physician offices in some of those same states again trying to encourage them about adoption of health information technology and standards and performance metrics but providing financial incentives in terms of this demonstration authority, so that’s section 649 demonstration.
There’s also a physician group practice demonstration that you’re probably aware of where this was mandated in the Benefits Improvement Protection Act or BIP of 2000 whereby it allowed for us to actually look at quality but in terms of overall efficiencies if you will or resource utilization. So we worked at the large group level so it was a lot of physician sites around the country whereby we’re going to be looking at that demonstration over the course of three years to get a sense of not only cost savings in the traditional sense whereby physician leaders had come to CMS correctly saying that one of the concerns that they have is that there’s no financial incentives on a lot of these quality activity and there may actually be perverse incentives, meaning that the current financial structure may actually be an impediment to improving their quality.
So what the physician group practice demonstration was designed to do is to try to decrease the barrier and actually be able to show that improvement in quality could actually lead to more efficiencies and so that we can actually allow for the quality improvement to somewhat pay for itself. So that activity is just getting launched and getting underway.
There was again a substantial amount of activity where we had to have a lot of dialogue with those physician group practices around what the correct metrics should be, how we’re going to actually do the data collection, at what particular level is the metric going to occur as well, so that’s another example of where we continue to have activity with the physician leadership around what the right accountability or the metrics should be and how it’s going to ultimately link up to payment. Within that physician group practice demonstration there’s weighting of particular metrics, that was actually negotiated with the physician leaders in those to make certain that it would represent the clinical quality that they wanted versus represent what the purchaser may value more so there was some negotiation there.
Moving to where, I’m looking at this work plan because I saw some of the questions here so I was trying to allow for some open dialogue that relate to some of the questions that you had in the status of the work plan. The first question was around the notion of is the focus to be health quality or health care quality, and I’m taking that to mean the services of health care as provided on the second prong. I think that’s going to be open for conversation with this committee but I assume if you’re making recommendations to the Secretary it may be one of those questions where you get the answer being both, because obviously with the Secretary there’s certain agencies under that authority are actually purchasing services or providing services and so to the extent that they’re doing either of those you want to look at the quality of those services while at the same time recognizing the role that the department plays in the overall health care population, the quality of the health care population. So I think it’s probably the answer is going to be both but you may decide which of those you want to consider or prioritize.
Along the lines of this second bullet point from the customer standpoint around performance measures instead of quality, I can tell you it’s an interesting question and we’ve done both, we’ve done it interchangeable at times to be honest. An example of that I give is with the Ambulatory Quality Alliance, this is an alliance where similar to what we did on the hospital side where we said we really want to work with physician leaders, with the external stakeholders so it’s a public/private collaboration so that we’re not the only ones out there doing it and we’re not pulling physicians between the public sector and the private sector as to what information is actually being requested. And so luckily the Agency for Healthcare Research and Quality, the family practitioners, the College of Physicians and the American Health Insurance Plans came together as kind of the conveners and the cofounders of that Ambulatory Quality Alliance and allowed for that activity to move forward.
So the current landscape, let me backtrack for a second before I answer this question, so let me just tell you what I think on the physician side the current landscape is as you think about the marketplace, currently it seems to me that the landscape is that you have the National Quality Forum invested to find a common pathway in the terms of consensus development process so the convener, and there’s pluses and minuses to that and people will argue rightly or wrongly about kind of the tradeoffs about having a consensus derived process, particularly the more that you’re rigorous about science and things of that nature you may have a different view then consumers or purchasers. And so that is a vehicle or the form in which all that activity occurs. But it does require and is continuing to be more and more rigorous about requiring scientifically valid and reliable performance measures. And so there’s a bar to entry or there will continue to be a bar to entry into that process.
One of the areas that we think that actually could get people prepared for that activity is what the NCQA is doing and what we believe that they’re going to be doing with the Physician Consortium for Performance Improvement, whereby the Physician Consortium for Performance Improvement has roughly 62 medical societies, maybe more then that now, 62 medical societies where they’re coming together, some of the measures will be more along the lines of QI and may not meet the rigors of a consensus derived process but it allows them to get going across the board so all the specialty societies and so CMS has been working closely with all these physician specialty societies to help them get going on their process, looking at where they may have guidelines, looking where there are gaps in terms of their measurement, and really wanting them to work with the Physician Consortium and NCQA to get to the point where they actually have measures that could eventually get into the National Quality Forum Process.
And then along side that I had mentioned the Ambulatory Quality Alliance which similar to the Hospital Quality Alliance is really the to me the operation arm where you have all this measurement activity but at the end of the day you want to still have some consistency both in terms of implementation and in terms of messaging and communication to the external stakeholders. And that’s where the Ambulatory Quality Alliance has played that role for us and it’s in its maturation process but its continued to play that role for us whereby we can work on the private/public sector so that there’s some level of consistency around what it means to have certain measures be implemented early and what is the messaging for all the different external users of that activity.
Now going back to this second question, with Ambulatory Quality Alliance this is one of those discussions that we had because there were definitely some that had the concerns about whether or not some measurement activity could actually be considered quality measurement activity. And so the compromise that was reached in that was that instead of arguing about that and there’s not just semantics, there’s some legitimate arguments about that, given that our overall goal is really trying to improve the performance of the health care services we chose to go with performance measurement. There was a compromise, it was a for us to move forward and just take that particular issue off the table, to be honest, so that we didn’t have another shot when we didn’t need it, we didn’t need to take on that particular issue of arguing about that, just put it in there as performance measure, it’s more broad, it’s more comprehensive, and it allows us to move forward which is what we did in that setting.
Consider the impact of surveillance assurance or improvement, I think that goes along with somewhat the first bullet point but also goes along I think with what I heard Brent James talking about about that spectrum, it really depends upon the utility and who your end user is and what you’re trying to accomplish. Obviously with CMS we’ve done several different things, an example of that is on the nursing home side and home health too, primarily those are outcome measures, percent of residents that have improvement in certain activities, so they’re outcome measures versus on the physician side and the hospital side primarily right now you have process measures. Some of them are intermediary outcomes but primarily there’s process measures. Again, this is one of those areas where there’s a lot of argument where I think the answer is both, I think you can to the extent possible you want to do both or all of the above or whatever you traditionally pick at the end of a multiple choice question set.
And the reason why I say that is because it is, it continues to be another one of those where people for some reason can’t get going because they continue to argue which one of these you should do, should you do outcomes, should you do process, I think the answer is we say both. Now there’s pros and cons to that process or that choice but one of the reasons particularly why we in comparison to some that would argue only outcomes, that we think process works, it gets down to that, the issue that Brent James raised about outcome measures and being able to get down to decision level, well, process measures traditionally if you pick the right ones allow that easier. In other words it allows it to be more actionable at the individual level that you’re trying to have accountability so a physician potentially, did they provide the service or did they not provide the service. To the extent that we are getting to that level are you using for those purposes, that’s one of the reasons why we’ve looked at process.
The other reason why to be honest you’ve done that is primarily right now your financial structure pays for services and services tend to be process oriented and not outcome oriented. So it’s another reason why you would continue to have process in there even though your goals may be outcomes as well because that’s the current financial structure, that’s how we currently pay people so to the extent that we’re paying people along those performance measures we’ll probably always have process measures.
DR. CARR: I’m going to just step in for a minute because I guess what I’m listening for, I heard it clearly this morning from Brent James about what he sees the committee could do in terms of making a national toolkit, I guess the question we want answered the most, and these are sort of ways of thinking about it, but what is a gap or what is an uncharted territory that this committee could assist with in terms of holding hearings, synthesizing data, and making a recommendation? Am I saying that correctly? And I’m actually putting it to all of you because I think that’s, you’ve educated us a tremendous amount in terms of what we’ve heard this morning but what we will, what will help us the most is to be able to articulate what is that, what is that area that you need help on, expanding an agenda or addressing a gap or getting more information?
DR. HAYWOOD: To me it’s a tough question to answer without, it’s kind of like, it’s hard for me to answer what a committee should do without knowing what the overall kind of what the goal for the committee was --
DR. CARR: And that’s okay, certainly more what do you need and we’ll leave it to these fine gentlemen to figure out which of all the things we can do.
DR. HAYWOOD: So as far as where there are gaps there’s a lot of people that are in the marketplace obviously right now, there are still gaps around infrastructure as Brent James mentioned so there are definite gaps around infrastructure and what the process may be to actually allow for that common data infrastructure if you will, or that platform. I think there continues to be some dis-concordance around what we may be doing, meaning purchasers, so whether it be public purchase or whether private purchase versus what say state oversight activities or entities may be able to do, and that’s one of the things that we consistently heard when we went out and did listening sessions as it related to hospitals, your purchaser is asking for a certain amount of information which makes sense from a purchaser standpoint but from a regulatory state oversight body they may be asking for different information.
And so I recognize that there seems from the end user or the person being held accountable that there’s a disconnect there and so I definitely think that there is a need for someone to be looking at state/federal partnerships like what is the, where is the correct balance around the notion, around all this quality activity and performance metric, what is the correct balance for states and federal entities to kind of partner along that.
I know CDC and AHRQ and some of those have spent a substantial amount of time look at some of that activity, some of the AHRQ HCUP data and things of that nature, so I think there definitely needs to be that role and I would offer that up as one suggestion as an area that would be beneficial that we’re interested but to be quite honest it won’t be the first thing on my radar screen.
DR. CARR: I think this, it’s helpful if we can sort of, I look to Richard Klein now, same thing, what’s out there that needs addressing that’s not being addressed that is the kind of thing that might be addressed by this committee.
MR. KLEIN: Well the second part I’m not sure, what came to my mind is a difficult area that when you have a set of indicators one thing that people have not addressed a lot because it’s very difficult is the relationship between the indicators, relationship between the metrics, between your process measures and your outcome measures, particularly when you’re in the business of setting targets, like we are for Healthy People, in many cases the indicators or the metrics that you used can be opposed to each other, where we increase survival we’re going to get more disability, we reduce heart disease probably going to get more cancer. And there’s a tradeoff out there that we just close our eyes to and don’t address.
Heard a presentation yesterday by the National Cancer Institute dealing with an area that probably a lot more is known about, they focused on colorectal cancer and took the Healthy People objectives in this area, which we have some screening measures, associated with that we have risk factors and we have an outcome of colorectal mortality. And got involved in a fairly extensive modeling project, I think they spent about a half a million dollars with Harvard on this one idea, and showed what kind of change would you have to have in screening, in risk factors, to get X amount of change in mortality which is really what we’re talking about. When you’re talking about the outcome measures it’s what is it we have to do to get there.
Another modeling project that went on similar to that was on tobacco done by, I don’t know who did it, CDC I think, that showed interestingly enough if you wanted the biggest bang for your buck on health you would put your resources into smoking cessation for adults rather then prevention initiation for youth, which blew a lot of people’s minds but if you think about it it really does make sense, you can deal with the kids later because they won’t listen to you anyway, just get them before they have the health problems.
So I think just to summarize that, and I don’t know if you can do this, I mean we’re really just in the infancy of it, but we know we have to think about the relationships, just to put a set of indicators out there without thinking about how they interrelate and maybe even oppose each other I think would be a great contribution if at least you showed that we had had some consideration in that direction.
MR. MOY: I think you hit upon a very important issue, this group, and that is identifying simply that different audiences have very specific different needs for these quality measures and I think that one, just stating that probably is important. But you may be gathering information about what the exact needs and projected uses of these measures are, we’re not in the business of just collecting them to say we collect them, we want actually to do something with it. But what we want them to do with it is different depending on the audience.
I know as an example for instance we focus again on our Congressional audience and we’ve really tried to beat our brains about exactly what they want from the information that we can provide to them. And I think in our minds we summarize it as simply they want to know if things are getting or getting worse, they want to know how fast they’re getting better or they’re getting worse, what areas we’re doing good in, what are the areas we’re doing bad in for priority setting purposes. And our needs are very different from someone’s purposes, they’re very different from the public reporting purposes that CMS engages in quality measurement, they’re very different from the NCQA purposes for providing information for providers. And I think working to identify the very specific uses and data needs for these different purposes I think would be something that’s helpful and I think it’s something you can do by talking to different audiences you talk to.
MS. MCCALL: Ernest, one clarifying question because I think it’s a fascinating point. What would you say for the work that you do is the most meaningful difference, you say look, I need this and I know that so and so over there needs it differently, what’s an example that we can take away with us?
MR. MOY: In a very generic sense it’s we want to tell Congress we think what is the status of quality and so we’re not really, we don’t really care about individual measures, we’re using them as markers for our system. And so we’re not committed to necessarily a measure in this particular topic, I mean they’re markers for us. We do want to be able to say whether heart disease needs to be a high priority, or diabetes, we do need that level, but we don’t need individual specific measures, we don’t need the comprehensive measure set that Richard tracks because we don’t want to fix this or that, we’re looking for markers.
DR. MCCALL: Very good, thank you.
DR. CARR: I think too following up on some of Brent James’ comments, that if there are specific outcome measures, even thinking of the CMF as measures, if what we’re finding is in pneumonia 99 to 100 percent of people are getting their O2 saturation checked. And yet we have payers who will, or organizations who will, you’ll get three stars if you’re 100 percent and you only get two starts if you’re 99 percent. And I think articulating, I mean I’m exaggerating a little bit but it’s not far off, and so if the audience is consumers who are trying to decide where they want to go with their pneumonia they’re getting very misleading information. If you are the organization, if internally you’re looking at your emergency room practice and you’re saying who was that one person we didn’t see that’s a very valid piece of information. So I really do underscore what you say and I think a lot of times the data elements are great and then the roll-up of whether is it done or not done, is the answer simply yes/no or percentile, quartile or good/bad, that I agree would help make the data appropriate to the audience and to the question being asked.
DR. SCANLON: along those lines, I mean it struck me listening to all of you, it seems like there’s a lot of issues in terms of Brent’s framework of the measurement chain about the science and the measure selection, I mean we have real problems there in virtually every area and when you talk about reporting to the Congress it’s like we report what we can but we don’t necessarily report what we need to report.
And there’s a question, this would be I think maybe a little different for this committee, I mean a lot of times this committee is focused later on down the chain in terms of how we get things to flow, and these are sort of further up the chain and they’re more conceptual kinds of issues and I think one of the things we should be talking about is what’s our capacity relative to other people’s, other organization’s capacity to deal with those kinds of issues. But they’re critical because they are going to ultimately sort of determine what we want further down the chain and whether we’re going to be successful in implementing the measurement that we want is going to be a function of what we build into our capacity.
But it was really just striking, I think every one of you came back to that and you talked about the cost of developing these, I mean that may be as a potential role for the committee to talk to the department about the lack, the importance, and the need for development in these areas.
DR. CARR: Help me understand that a little bit more, so you’re saying more often we’re focused down the chain and we should be more high --
DR. SCANLON: I think we at times focus in looking sort of back at different things, we focus on, we’d like data to flow and we identify data that we would like to flow and --
DR. CARR: So the UB-04 with the condition modifier, that kind of thing, so that’s at a detailed level and you’re saying just at a higher conceptual level?
DR. SCANLON: Higher level, why do we want that information in the UB-04, I mean we want it because we have certain measures in mind that we think we can use to measure quality, that’s a very different type of activity to decide what those measures are then it is to say let’s get this information from all providers on all claims and those kinds of things.
DR. CARR: Well, not to perseverate but it gets back to my thought about the tremendous national attention on the electronic health record and thinking about once we have that and we can go into it and we can learn something will we have configured what’s in it in a way to answer the questions that we feel are the greatest priority.
DR. SCANLON: and I think what I’m saying is consistent with that, we really need to be thinking ahead in terms of the science and the measure selection to try to influence that, but then we also had this discussion this morning about the idea of flexibility in that record, we’re going to be doomed --
DR. CARR: So just to reiterate it, at a high level is where does the information go, what do we do with it, and another question is how good is the information, not that we pick and choose what data element ought to go but rather what’s the level of rigor looking at those data elements. And then the other is the sort of evolutionary process, I mean we’ve seen this with ace inhibitors, people are giving ARBs, equally good treatment but points off on the final report card.
DR. HAYWOOD: Can I just maybe throw on the table maybe a different, maybe it’s not different but I think Steve had just mentioned in passing or he will, another approach may be to take a look out as to what you think that idealized or design would be and then you actually start to chart our process towards that. And so in other words instead of necessarily looking, getting to quagmire of all these individual quality measurement activities, that you may actually consider whatever working with external stakeholders and talk about what that infrastructure should be and what the systems should be on these processes like Brent was talking about and then consider what the right framework or the matrix should be around measuring our progress towards that idealized design.
So that may be another construct because to be honest I don’t know who’s doing that activity, who’s taking time out, Institute of Medicine is a little different as far as that construct so I don’t necessarily think that’s the same construct and I know we’re focusing kind of again, we would like to stimulate people to go towards that but I can’t say we’re necessarily the vehicle either so that may be, I don’t know if that’s not concrete enough or not for you, it may --
DR. CARR: Let me elaborate again and see if I understand this correctly but that we are, in some ways we’re looking for a definition of quality and safety, I mean we talked about this at one of our earlier meetings, if we take the IOM construct of safe, effective, efficient, patient centered, equitable, timely, and we think about what, getting back to our matrix here, what are the system level initiatives towards safety, timeliness, effectiveness, what are the specific disease elements, because that’s really what we have. We have great information on heart surgery and interventional cardiac, we’ve got that part done. We have other conditions that don’t have and so you’re saying have an overall scorecard to say we forgot about populations, or we forgot about functional status or we forgot about equitable, and we’ve made enough progress on cardiac surgery or something like that.
DR. HAYWOOD: Well I’m guess I’m taking it more abstract then that for now by saying as you continue to have others come forward think about whatever that idealized model would be, in other words how do we know when we get to the end of the road that we’ve gotten there. And so it’s not just, to me it’s not a matter of saying well we’ve covered it for cardiac and we haven’t for this, instead to me it’s a matter of looking at the system as a whole and saying how far are we along that road as far as safety in the system, and there may be a multiple set of things on safety if you’re going to go into the six IOM aims. And then what you end up doing is kind of reporting out, to me at a macro level where our progress is, and so I don’t know how it ties in a little bit to NHQR because you’re using some of those, I think four of the six maybe. But that’s kind of the conceptual thinking around it was allowing for us to say okay, a way to really show that we’re still on target, that we’re moving towards some end point where we would actually have the system that we’re hoping to accomplish.
DR. ROSKI: Yeah, as I’m thinking about what would be great hearings that would really benefit NCQA and hopefully the American public at large there are four topics that maybe come to mind. One is you mentioned this notion of well if we want all this information to come forward who’s going to be in charge of developing these measures and how do we finance that, that’s one question.
A second question in my mind is should we be talking about a paradigm shift in terms of what we’re actually asking for, should we be talking about value as opposed to quality. Because I’m afraid that if we continue to only talk about quality and not talk about what it costs the American taxpayer and the American public to finance the system we may just lose the American public in this debate. And would it not be useful to come up with a definition of what value is, how you might define it, and how then subsequently we might be able to measure it at various different levels.
Number three, one of the greatest activities or sort of the greatest energy at the moment in the performance measurement arena I see in measuring ambulatory care performance and in particular physician performance. There are a lot of unanswered questions relative to well how exactly should that be implemented, is it really reasonable to hold single doctors accountable for issues that may have been contributed to by many, many doctors, and how should the American public, or how should the system be thinking about who’s accountable for what. And how can we promote through a measurement system if that’s possible the delivery of systematic ordinated care that leads to effective and efficient outcomes. So for example may it make more sense to try to focus on measuring groupness or coordination between doctors as opposed to trying to put report cards on individual doctors, really a lot of unanswered questions in my mind.
And then number four, a great deal of the responsibility for measurement will continue to rest with the private sector in coming forward with information to deliver it to the marketplace. I don’t know that there has been a very intensive debate about well what would a high integrity measure implementation system actually look like and what would the characteristics of such be. For example, this whole notion of is anything actually being verified by what’s being put forth as performance information in my mind is something that’s not particularly well understood and the sort of pros and cons of that. And I think we’ve spent so much energy on just coming to consensus on the measures when in fact then you actually have to ask the question well how in the world would we ever implement that and would know that the information that’s coming to us has any integrity at all.
So those are four topics in my mind that we are wrestling with on a daily basis trying to talk to various experts and stakeholders about what a reasonable position might be and the authority that your committee might have in setting some definitive direction here would be very useful to us.
MS. GREENBERG: A few thoughts. I just found this very, very useful and fascinating and stimulating and so I thank everyone for that. But not necessarily in defense of the UB-02 or the UB-04 but what this, it’s sort of like there’s nothing new under the sun, what you were saying Bill really reminds me of is that old fashioned camera image, remember that, that Lisa Izoni(?) and Cathy Coltin came up with and is I think immortalized in at least one of our annual reports, but this idea of all these, and I think it was really inspired by the HIPAA onslaught of minutia and detail on standards, and that is that all of these are just a means to hopefully or should be a means to answering certain questions, what are the questions we really want to ask about the health of the population of which the health care of the population is a subset really of the health of the population. And that that’s the focus that the committee is supposed, should have, and that’s partly what John really, Lumpkin, tried to get more of a population focus in the committee, etc., and I think that has been put into, I think we’ve made some success on that but it’s an ever going challenge.
But that’s why I’m really intrigues by Ernie’s suggestion although I’m not exactly sure what actionable things come out of that. But getting some clarity, and it’s really been throughout the whole morning, you start with Brent saying about well what are we trying, what is our aim. Well, that’s kind of what you were saying, there are these different aims, the different goals, and then of course to meet those you need different approaches, different data, different measures, and in fact getting to what Richard was saying some of these actually not only compete with other they’re in opposition to each other.
This certainly would benefit from clarity and explication and it might be a very educational to all the different groups who are struggling in this area. And it is potentially an area where the committee could shed some light or get some taxonomy and clarity, etc. I mean I can see that coming out of possibly hearings and really in a structured way trying to identify that, that could lead to people realizing in some areas that they’re going down the wrong path or that what they’re doing is actually conflicting with what’s going on over here and is there some way we can partner better.
But to really come down to what would be actionable out of that I’m not sure but it certainly is intriguing to me and particularly to get this focus back on what you said, Bob, was Simon’s charge to us and that is what questions are we trying to answer. I think the bigger question, what questions is the country trying to answer in this area and then how can we contribute to that. So I’m quite intrigued by that, on the other hand everything that Dr. Roski mentioned are all very interesting too and I’m intrigued by the possibility of the workgroup partnering with one of these organizations that has these issues and they’re not really in a position to hold hearings and to gather information. We certainly partner with parts of the department and I don’t know if we’ve, and we partner with the whole health industry, certainly with the standards development organizations, so those are interesting too.
And the only final point I would make is that as Justine reminded us we are, this workgroup is a workgroup of the populations committee and there are definitely things that have come out this morning that might be better addressed, or could be addressed in parallel while the quality group was addressing some things by the Populations Subcommittee and I guess we can bring those to Don tomorrow, of course in many ways you’re the same people but they do have somewhat different resources and certainly additional staff resources, each have different staffs. So those are just some of the thoughts I had.
DR. CARR: Bob has reminded me that we’ve probably talked through lunch and we’re supposed to get lunch and so I’m going to ask Don and then Simon and then Carol and then we will take a break.
DR. DETMER: I just wanted to pick up on the comment about are we maybe using the right construct and you mentioned value. Ten years ago I started a think tank called the Blue Ridge Group and we put out a report every year and our most recent set of reports has really built an idea of value driven health care and what does that mean and how do we look at that because in fact we really feel like that is where we ought to be looking. I think increasingly as health care becomes both paradoxically more effective and also more wasteful, but also more dangerous as well as potentially wasteful, we have I think a more and more of a need to look at return on investment and not in terms of narrow definitions of that, it can be also quality of life, it can be a variety of kinds of dimensions on that.
But at any rate, I can get these reports to you, one of them that really has I think fleshed this out. It’s not the last word but I hope it’s a pretty serious first of words on this because I honestly do think this is the way to go and to give you an example, if we look at the equity dimension because I must admit I do think that IOM set of six aims is important, I mean what we really want is health and at some point we’ll want health care to help us get that.
Now in a funny sort of way however social systems also seem to be making an important contribution to that, it’s fascinating that the UK puts half as much money into health care then we do and they clearly get more value in terms of the health status for the money they don’t put in compared to what we do put in. And it looks like part of that actually interestingly enough is just the assumption of a population that health care is there for them, whether in actual reality, whether actually when they get sick it necessarily it does show up or not.
And I think some of our, our Medicare program for example does in fact give people a true safety net in that sense so it’s not like our system doesn’t have elements of that. But it’s really fascinating and it seems like this stuff has a level of value at a spiritual kind of level as well as it necessarily does at sort of a nuts and bolts kind of care sort of way. We’re very, very still caught up in a medical model kind of approach to things and we’re still also sadly oftentimes caught up in a public health separated totally from the clinical side.
I think Carr(?) White spoke to that chasm, another one of these chasms, between the schism actually as he used the term between public health and clinical work and we do need to put these pieces together and I think value is one of the ways to start moving in that direction. So you triggered some comments that I think I might be able to feed in some things that could help you in that regard because I do think it’s a quite useful construct.
DR. CARR: Could we get a copy of the report today?
DR. DETMER: I can email the PDF the same way I downloaded his comments.
DR. COHN: Well, this is just in time information, I’m going to be very brief, I was just sort of fascinated by the presenters and their comments. I was sort of struck and reminded that the whole quality initiative and quality concepts really have not captured the imagination of the public and I think that’s what we were all talking about is that we’ve got a zillion measures but sort of nobody cares except if you’re a consultant or something like that. And it’s just something to think about.
I do have to say that like Don I was sort of taken with Dr. Roski’s comments, we’ll have to rename you the values subcommittee workgroup or God knows what, but I do think that value is a very important concept. Given that I hear time and time again that the chief issues for the public are cost, sort of convenience and service and then quality, to has us only look at quality metrics is sort of almost like staring at your navel and sort of missing the rest of the body.
Once again I’m not here tomorrow, you’ll be prioritizing things and perspectives and how you put this together but I think it is an important framework.
MS. MCCALL: A couple of comments just to play off of Don’s comments earlier, and in particular, two things, the first is I think there’s another paradigm, there’s a mental that we’re working with and I don’t know if you folks see it or not but it is that the physician is our unit of analysis. And I would submit that that’s a fairly narrow unit of analysis because if I am, if I think about Justine and if she’s my physician and if I have cancer or whatever it is, it’s not about her, it’s about me, the person. And so that’s really it all kind of goes to that point right and so even in all of that is to serve. So I think that we could do some things to expand our units of analysis, not to shift necessarily but to expand, and I think it also gets to what you were talking about Joachim in terms of the individual physician and I think the more we anchor on an individual physician the more we get pulled toward the accountability end immediately and we’re going to miss our aim in terms of trying to build something that can satisfy a number of things along the spectrum that Brent talked about. So I think that that’s one thing.
And I think bringing value into the equation as opposed to quality, quality is a concept that kind of brings with it patient and the entomology there is patient as a passive, suffering, silent thing, so I have no role in quality as a patient. But in a person or as a person I have a role in value, I actually have something to bring to that, it’s not about health care, it’s about health, there’s things that I do outside the health care system that may go into a PHR some day, that are about what I do in between visits to Justine and sometimes those are the most valuable things.
So bringing value in, giving it a definition, looking at that definition from constituencies and through different lenses helps get to how are different constituencies going to use the metrics, how are they going to think about them, what is a common denominator to get us going so that we can kind of get this train out of the station, or the boat away from the dock, whichever metaphor suits your fancy. But those were a couple of paradigms that I see running through there and I think extending and bringing in the definition of value can help break that paradigm and give us --
MR. HUNGATE: I’d like the option of having the first question after lunch to continue this excellent discussion.
DR. CARR: Bob, what time do you want us back here?
MR. HUNGATE: 2:00, we need enough of a break to make it a reasonable break.
[Whereupon at 1:10 p.m., the meeting was recessed, to reconvene the same day, June 2, 2005, at 2:15 p.m.]
MR. HUNGATE: Let’s reconvene. I’m ready.
DR. CARR: Okay, we’re ready. I think what we would like to do is, I tried to summarize the themes that we heard in our earlier discussion, invite any comment, amplification or additions but not just try to briefly help summarize, and then I believe John and Don I guess --
MR. HUNGATE: I really need to ask my question --
DR. CARR: Oh, absolutely, no, I fully intend, why don’t we take, ask your question, Bob.
MR. HUNGATE: Thank you. The discussion this morning, especially the last part, made me think about some history. I was on the board at Washington Business Group on Health when the precursor of the HEDIS measurement set got talked about, health employer information data set is the strict interpretation, the whole thing was done to help employers rationalize which plan they dealt with. And I say rationalize, it was meant to be a measure but it was acknowledged that it was only part of the things that were important so it became a way of judging, a tool. It has not been useful for employees. We’ve tried to use it at Group Insurance Commission which I chair in Massachusetts and it’s of no value.
We’ve shifted a lot of the emphasis in terms of measure development from plan measurement now to provider measurement, so there’s a great drive for accountability at the provider level. I am fearful that we’re going down the same path again and that we’re helping the employers to make decisions about which provider in some way but that we’re really not serving individual needs. And so I think we’re going to spend a lot of energy in accountability in picking these are the places you ought to go, but I don’t think we’re going to end up with people believing that their interests have necessarily been served.
And so my suspicion is that we’re going to have to move to procedure as the point of focus of accountability and I argue that from the standpoint that when a person has a non-health condition dealing with that condition usually asks for action, whether it’s a physician action in a pharmaceutical or a surgical action in a hospital, but it’s a condition driven interest. And the intensity of interest and attention to information by an individual then is vastly different then it is when it’s picking a plan.
I think that in general the critical decision is what’s going to be done and then the where it’s to be done is more likely to be driven by knowledge from other people’s recommendations of where did you have yours done, that used to be called the grandmother factor in obstetrics, that what the grandmother said was more important maybe even then what the OB said.
And so I’d like comment in that context in terms of as I listened to Brent’s focus on improvement I had the feeling that focus on procedure would be a better base for improvement then our focus on provider, that we might be able to stimulate better provider cooperation in data sharing if that were where the emphasis were, that it might improve the way in which the if you will political climate exists on the ground. So that’s my compelling question for me anyhow.
DR. CARR: Well, I might ask you to elaborate a little bit, so if we talk about what we heard this morning I think there was substantial discussion around the concept of value as opposed to quality and as Carol pointed out that really it’s not about the provider, it’s about the patient.
MR. HUNGATE: Well and let me pick up on that. My sense is that when we start talking about value you start getting into concepts like cost effectiveness, quallies(?) with the various measures of health status, functional status which the committee is on record as believing are missing and important measurements. Cost effectiveness is going to be a key piece of any value determination, cost effectiveness might be considered as a marginal cost to the individual over some threshold of cost effectiveness. So when you start talking about value I think it’s going to have to come down to the individual again and it’s likely to be procedure related.
DR. CARR: Well, I heard it also that what’s the cost to society to have all these report cards and these matrix and how we are better for it and is it what we want to pay for. And I heard Bill saying that we would look, maybe not go down but look up at the more macro as opposed to the micro --
MR. HUNGATE: I think this does become macro because it philosophically changes from an accountability model to an expectations model I think. So that you can, if you focus on the intervention and procedure and the achievement of appropriate expectations that’s different, it’s not accountability.
DR. CARR: So you’re saying a medical procedure?
MR. HUNGATE: Yes.
DR. CARR: So like heart surgery, your mortality will be less then one percent if you have no risk factors, that kind of accountability? I’m still struggling with when you say focus on procedure what does that mean. Just help me understand a little bit more.
MR. HUNGATE: Well, let me go back to - we didn’t talk about definitions of quality from various viewpoints. For a patient health is the expectation, interventions are directed at restoring health, that’s the hope. One of the definitions of quality is meeting or exceeding expectations. We don’t do a very good job in the health care system of setting appropriate expectations, oftentimes because the information base on which that’s done is not particularly good, it could be better, that’s kind of what I’m trying to get at is that improving that set of looking toward expectations is an improvement model as opposed to an accountability model. That that’s where the value in health care is at the procedure level.
DR. CARR: Alright, let’s see, okay, Bill, go ahead.
DR. SCANLON: I agree in terms of if we’re trying to influence things we really do need to influence them at the procedure level but I think it was Joachim who talked about sort of overuse, under use, and misuse, in some respects if we had information in what I think of as real time about sort of where this recommendations fits compared to those standards we may be able to influence utilization sort of in a positive way. Now I do know that there are people that are very skeptical though about patient’s willingness to make decisions in the face of having an actual condition and sort of in the kind of psychology of oh my God I need treatment, I’m going to turn this decision over to somebody and so the question is will this information be effective then.
When we were trying to rate plans and to some extent we were trying to rate providers, I think we were in a situation where we were instead of operating in a world where there was relative freedom in choosing what kind of treatment you got, we were thinking that we were going to be locked into someone who had tendencies to do certain things and we wanted to make choices on the basis of that. Now there’s been a lot of movement back away from that, I mean there is more flexibility now and so information of the kind that you’re talking about has more potential today then it may have had a number of years ago but there still is that issue of will patients, you started off by saying patients weren’t using the HEDIS kinds of measures, if given this kind of information which is going to be a more challenging task will they use it or will they, the psychological mindset be such that they’re not going to and they’re simply going to defer to physicians and we’re going to get the same misuse, under use and overuse that we had in the past.
MR. HUNGATE: My sense is that if the information base were good that I would want my personal primary care physician to be familiar with that information base and I would use a trusted source for that judgment. If I think about our system health parameter and compare the U.S. health care delivery system to other systems we under use primary care vis-à-vis other systems, percentage concentration in primary care and others is greater. If that information were in primary care it would be stronger and would be a better part of the system. So that’s a related system direction.
DR. JENCKS: I’m like perhaps some other people judging by expressions still trying to get my arms around this, I’m still trying to get my arms around what it is that you’re pointing us to. One thing that comes to my mind as you talk about this is the following, the appropriate level of morbidity from an unnecessary procedure is zero to an infinite number of decimal places and there’s no doubt that the fastest way to reduce morbidity from a class of procedures is to stop doing them on people who don’t need them.
I think though that you may be creating an unnecessary dichotomy if you see this as a consumer strategy rather then a broader strategy if I’m understanding where you’re going. We have shied away very much from looking at whether procedures were necessary because we thought that we could get buy-in from the medical profession and elsewhere on things not delivered that should have been delivered a lot more then on things delivered that shouldn’t have been delivered. But I think we’re beginning to emerge from that dark wood and we ought to be looking more broadly at whether it is possible to measure the appropriateness of various procedural activities.
MR. HUNGATE: I don’t disagree, the observation that I have made as a payer is that the HMO model which was expected to restrain inappropriate procedures has had difficulty doing so, that individuals take denial as you’re just trying to save money, you’re not giving me what I need and deserve and the data has not supported the denial adequately for it to occur. So yes, I think it’s part and parcel of the same piece. Now I should shut up.
DR. CARR: Okay, I’m going to just sort of summarize again to say that I think we’re going macro and micro and this is what we’re struggling with, where what this, what we’ll talk about tomorrow is of all the things that we’ve heard where do we fit in, so thank you for this approach to think in terms of the procedures. I would say also at the micro, well, maybe not at the micro level but the data integrity we heard about this morning from Brent James and others in the subsequent discussion so in some ways that data integrity thing as we heard from Brent is very micro, make a toolkit of the measures and help make them strong. But we heard also a broader theme which is as all these measures develop what is the process for making, maybe the toolkit is a means or whatever but the concept that the data measures have to have integrity is I think a very important one.
I think a third thing that we heard is that the same information will have different meaning to different audiences, or the same data elements would be presented in a different way to different audiences, in some cases it would be a yes/no, in other case a high/medium/low, in another place a percentage or percentile. And I think that is one of the things that has created tremendous struggle that we have consumers worrying about 99 versus 100 percent and so, and then another thing that we heard was about systemness, are we focusing on physicians or are we focusing on systems, and where will we get the greatest bang for the buck, and then bang for the buck speaking about value, that can we only talk about a quality process measure or outcome without understanding the expense that goes into building the infrastructure and putting it in terms that decisions can be made.
I think also Don was saying that, you said it very eloquently here, that we have, what was it, we have so much, we have tremendous technical expertise and we have tremendous waste and what was the third thing that you said there, that we have a lot of things, some good and some, because we have high technology and we have some great outcomes but we also have some waste and understanding how to use that is important.
So I guess sort of concluding, bringing together what we heard this morning, we’re hearing about are we looking at macro or micro, are we looking at definitions or tools or applications, I think I’ll open it up, and I want to say a little bit more about the populations, I know we started to touch on that and was there more that we wanted to say about populations or functional status.
MS. MCCALL: I have one more comment on needs if you’re taking a list, earlier today and I can’t remember Richard if it was you said basically infrastructure is a need okay when you talk a lot about systems and things like that. I would add on to that by saying that not just any infrastructure will do, what we’re talking about is an infrastructure that has built into its very fabric, its very essence, a kind of adaptability that is not usually historically built into our systems. And it’s almost like the technological equivalent of kind of the paradigm of consensus that we walk around with where we all have to get to agreement before we can do anything, oh my God, but we’ve got to get all the systems built before we can do anything.
And so there’s a different type of need I think we’re going to have for architectures that are open, for architectures that are not as tightly coupled as the systems that we design today, they have to be robust, they have to be sound, but there are different types of architectures and technologies out there that could enable a more loosely coupled, a more open to adaptation innovation. And I think we need to ask that a new mindset be brought because we will only be able to deliver what that system or those systems are capable of handing. So I see that as a need that kind of amplifies the general need of infrastructure.
DR. CARR: Again getting back to Simon’s mandate that we think in terms of this committee doing something that is value added, actionable and timely, I think that we’re thinking about big picture maybe definitions or applications or some of the more at the data level management and I guess as we keep through this I think we need to just think about us and where, is it a new horizon that we’re looking at or is it a gap in the current state, figuring out where it is and then what’s feasible for us to do to meet a particular need.
I don’t know, is this a good stopping point?
MS. MCCALL: Just to kind of wrap those up, I also heard a need for some organizing mechanisms, I want to make sure that those are captured so that you have those down.
MR. HUNGATE: [Comment off microphone.]
MS. GREENBERG: I’d also, I think we’ll probably be asking everybody who presents what you asked them, Justine, so I think we should be keeping a running list too of when we have asked people well what do you think we could contribute to or given our resources and what we’re good at and then we got a number of suggestions and I would think we would want to kind of keep all those on the blackboard and then revisit them and maybe group them or something, tonight or whatever, and then come back and revisit them tomorrow because we did get some specific recommendations which were some macro, some micro. I mean I think, I know I wrote them down but maybe some of us --
DR. CARR: I wonder is it worth using, I mean we’re ahead of schedule, is it worth using some of this time to vet them a little bit further?
MS. GREENBERG: Actually we’re not ahead.
DR. CARR: Oh, we’re not?
MR. HUNGATE: We’re pretty close. We’re a little ahead --
MS. GREENBERG: Part of continuing the discussion.
MR. HUNGATE: Continuing discussion. I don’t, let’s ask the next two panelists, would you rather have us do more discussion now or after you’ve added your information in the next round?
PARTICIPANT: Add our information.
MR. HUNGATE: That’s my intuition as well. They want to speak now, so we’re going to move to the next panel, I am now Don Steinwachs introducing the two speakers who need no introduction, thank you Joachim and Richard for helping us and we look forward to further dialogue.
And so you’ve already been adequately introduced I believe so which of you is speaking first --
Agenda Item: Panel 2: Health Informatics: Timing of Changes and Limiting Factors - Dr. Halamka
DR. HALAMKA: I’m up first, hey, very good. So I’d like to give you a broad overview of what I’m doing in Massachusetts in the context of quality just to set a foundation, I’ll tell you some of the activities and some of the challenges that I face. I have multiple roles in Massachusetts, my primary role is that I’m chief information officer of CareGroup, CareGroup a $1.4 billion dollar integrated delivery network with 3,000 docs, 12,000 employees, serving about two and a half million patients. In the context of IT I oversee about 150 mission critical applications, moving 100 terabytes of health care data every day. To give you a context of that environment we haven’t had a handwritten or voice order since 2001, we haven’t use a piece of film since 1998, we’ve had a fully electronic medical record since the late ‘80s, so it is very much a culture that thrives on innovation and electronic systems.
Now as a brief comment I’ve created total dependency on these electronic systems so when I had the outage in 2002 for a day and a half we actually went from the hospital of 2002 to the hospital of 1972, the COO had to go down to staples and buy a 100,000 sheets of paper because we had none and we had to teach interns how to handwrite orders because they’d never done it. So for all this technology that I’m describing it’s wonderful, we believe it enhanced quality and value, but it also creates a fascinating dependency that I just want to make sure you’re aware of.
My other role just to give you a context is that I as now CEO of the RIO and responsible for the grid of connectivity for clinical data sharing in the state of Massachusetts, and that grid takes several forms. On the one hand we have financial transactions, HIPAA transactions, and since 1997 we’ve been doing HIPAA transactions meaning benefits eligibility, referrals, authorization, claims, enrollment, that kind of thing, as a community. So instead of having point to point solutions where the Mass General is sending transactions to Blue Cross and there’s a different implementation guides for Tufts to deal with Beth Israel Deaconess, we did HIPAA as a region together and big bang, the end result of all that is that we reduced the cost of clinical transactions from $5 dollars on average to ten cents on average, and this means that all the paper, all the faxing, all of the standard processes went away as a community. That gave us a foundation and built trust to now talk about something more controversial then HIPAA transactions, medication lists, allergy lists, problem lists, these things that obviously have far reaching privacy and security implications.
So the way we’re organized in Massachusetts is that we have a number of organizations, number of entities that serve what I’ll call sort of the RIO functions, Mass Share which I run is the grid, so as a community utility it’s responsible for connectivity and I’ll talk about connectivity further in a moment. We have the Mass Health Data Consortium, and many of you are familiar with that, the work of Elliott Stone over the last 25 years and Elliott recently passed away, but new leadership is being found for that organization. And that organization is the convener, that is as a community let’s get all the stakeholders together in a room and just come up with a privacy and security policy that works for everyone. It’s very much a business incubator.
We have the last mile, so if I’m doing the grid and we have policies, how do we get into the doctor’s office, and the Mass eHealth Collaborative recently received $50 million dollars from Blue Cross to do 900 physician office’s last mile, so they’re very complementary, if I’m building the grid something has got to plug into the grid, the Mass eHealth Collaborative does that.
Now you should know by the way the Mass eHealth Collaborative is giving hardware and software to doctors for free except they must give us their data for quality outcomes and performance measurement. So that’s the hook, you can’t have the hardware unless you give us the data and obviously there’s privacy and security around it.
So in terms of those various organizations functioning as a regional interchange organization we have felt that there are certain community utilities that need to be defined to make all this work. First community utility is who is the patient. One of the challenges of course and you guys know because you live this every day is we don’t have a national patient identifier and it’s unlikely that we’re going to have a national patient identifier any time soon. So this means that as a community, if John Halamka goes to a doctor’s office in Newton and then to Brigham and then visits the western suburbs and breaks his leg, somehow you need to tie together my medical record across all those various points of care.
So we came up with a notion of a record locater service, a community utility by which with patient consent a simple, not clinical data, but evidence of your visit to a provider would be forwarded to a statewide database, it wouldn’t be controlled by any government entity, it would be controlled by a collaborative of payers and providers. And the way this works is that if I go to Beth Israel Deaconess and I say you know I’m here and I consent that my visit, not details about it but my visit was forward to the national, excuse me, the central database, it would give my name, gender, date of birth, not SSN, don’t want to get too controversial there, and my medical record number and place I received care.
On the other hand you know my job is pretty stressful, thinking of checking into McLean Hospital, that’s the Betty Ford Clinic so to speak of Massachusetts, I’d rather not have my visit at the Betty Ford or the McLean Hospital forwarded to a statewide exchange because in fact the very notion that I have a McLean medical record number is disclosing of a clinical condition. So that’s my choice, I decide. So community utility, secure mechanism by which we have pointers to all the places where patient data lives.
The use case by the way, imagine that a doctor who needs my information, who’s authorized to use the system, can now type in name, gender, date of birth and get a list of all the institutions that I had received care at and hence records would be available at.
Second community utility is a clinical data exchange, a secure transport mechanism by which standards based transactions about my medication list, my laboratory values, problem lists, allergy lists, clinical notes, can be exchanged. And this is really much more of a technical challenge, you can imagine just the vast aspects of privacy and data control issues, who gets to see what, what is disclosed under what circumstances. How do we deal with mental health and substance abuse notes, how do we deal with HIV status, how do we deal with sexual assault records, all of these kinds of data that certainly we never want to be compromised but even in fact we have to be careful that a doctor whose caring for a patient in the emergency room may need their medication list but really doesn’t need to know their mental health history. So a lot of the decision making on this kind of role based access control and auditing and security is all done around this clinical data exchange infrastructure.
And finally something that we’re planning on launching in the spring of 2006, an e-prescribing infrastructure that creates a statewide mechanism for any clinician to have a standard NCPDP transaction go to a community utility where eligibility of the patient is checked, formulary is checked against the payer, and then a prescription is routed to the PBM, the mail order pharmacy and the retail pharmacy as necessary.
Now you’ll notice in describing those community utilities I haven’t yet described the community utility for quality measurement and the answer is I don’t know how to build one. And the reason I say this is it is very challenging for me to define the exact use case of what am I going to measure, what’s going to be actionable about that measurement, what are the national standards I’m going to use to record the data and transmit the data. So to date I’ll tell you what we’ve done is we’ve said David Bates you probably know more about this topic then anyone else and we’re going to fund a working group in the context of all of this data exchange activity for you over the next several years to in the state of Massachusetts get all this clinical data and figure out what process measures make sense to help us deliver higher value care.
But we need your guidance, I today can go to HL7 and get standards for exchange of problems and allergies and I can go to SNOMED for vocabularies, NCPDP for script history, but HEDIS measures as you’ve described, I’m not sure those are that helpful and boy when I share them with doctors or patients they say well what about risk adjustment and outcomes, process measures like DOQIT. Well actually those seem a little bit more reasonable, did the patient receive the ace inhibitor after they had the diagnosis of coronary artery disease or CHF, well, that’s sort of a yes/no question, did you receive the med, okay, it’s a good process measures, well that seems better and maybe I can do that because I have access to medication lists at a community level so maybe that’s the direction we’ll go.
But the problem I think you’re going to see is that as these RIOs spring up nationwide they’re all going to be focused on immediate clinical transactional care and the quality question is going to be asked in a secondary way. They’ll say well we’re improving quality because now we have medication lists and therefore we aren’t prescribing drugs that have drug/drug interactions or drug allergy interactions, oh and by the way we’re reducing the costs of care because we’re not ordering redundant labs, we’re able to share MRI results across the state so therefore since we’re making fewer errors and we’re ordering fewer tests the value of health care has gone up. And that’s about the hand waving argument we’re at today and so if I spent 50 percent of my life as I do defining standards and policies for clinical data exchange I look to your leadership to help me define what is quality, how do I measure it, how do I exchange it, and then how do we use those measurements once I’ve built them for you.
MS. GREENBERG: That’s all?
DR. HALAMKA: That’s all. That’s my brief introduction, I was asked to do no prepared remarks so that was the brief one --
MS. GREENBERG: No, no, I mean that’s all you’re asking from us?
MS. MCCALL: He’s the farthest along, you’ve got the grid.
MS. GREENBERG: It’s fantastic, in fact my reaction is those who know Dan Friedman, we always kid him that he’s in love with Canada but he’s from Massachusetts, he should just be in love with Massachusetts.
One more comment on that in fact, I’m sorry that Simon isn’t here but the Standards and Security Subcommittee held a hearing about two months ago about the ROI, return on investment from HIPAA, and nobody could come forward with any statistics like you’ve brought forward because nobody had a community solution I think and that is something that we haven’t pointed out but I think really we need to take back to them because we kept hearing about was all the companion guides and well we haven’t seen a return on investment yet and I’ll have to believe you that you really have.
DR. HALAMKA: So I’ll just give you the numbers, so as a community we’re doing six million transactions a month and so this is real, and you can see we’ve gone from $5 dollars to a dime. But even at a hospital level my AR days at Beth Israel Deaconess Medical Center are now 56 down from the 70s, why, because we don’t have all these administrative denials and we’re not doing all of this ping pong of claims back and forth because we just get the clean claim in electronically and it’s paid. So I think our return on investment thus far doing knee, hand, and HIPAA transaction, the state of Massachusetts is easily $100 million plus.
MS. GREENBERG: We need to factor that in, what they’re collecting. But then I mean didn’t I also hear that Massachusetts has a basically not for profit health care?
DR. HALAMKA: Right, so we had one for profit health care entity, Tenant, and they’ve just left, so we are a regional payer state, five major payers, all non- profit health care.
MR. HUNGATE: Okay, Don, go ahead.
Agenda Item: Panel 2: Health Informatics: Timing of Changes and Limiting Factors - Dr. Detmer
DR. DETMER: I guess my question would be is what are we going to do until an American health care system arrives? I mean that’s really what we’re talking about and in a way as I see it, and I’ll build on his comments, I sit on a commission on systemic interoperability and we’re meeting, struggling on how do we make just these connectivity things happen for data transfer let alone create a system that can do a serious job of managing the knowledge base on this exploding investment we’re making in science at NIH as well as AHRQ and all the other kinds of good things and what’s going on in the private sector.
So we have a lot of tigers by the tail and in a way I like to see, I like to think about health care being roughly where the American financial system was after the crash, we haven’t quite had a visible crash but we got a lot of crashes all the time every day. And the thing that’s interesting is that we haven’t built really the structure that we will have I believe, I don’t know how long it will be before we’ll have it, similar to a Federal Reserve bank, similar to an arms link kind of thing that we have the closest metaphor we probably have to date of what will emerge is the Institute of Medicine but even then I’m not sure that that’s the right organization for what I’m thinking about.
But in the meanwhile what do we do until that arrives, it strikes me that Secretary Shalala when she did ask me to help reinvent I guess the traditional mission of the NCVHS from really being a dead statistics kind of discussion group and that’s not fair but that’s sort of what we were talking about in terms of vital statistics, how do we change it into an information, policy, advisory committee for the U.S. government related to health. And I think that transformation occurred and that’s why you’re talking about these things, thank goodness, going really forward and looking at it.
So in a way I think that’s kind of our situation so I see your role as being two fold, one being guardians of a vision and also guardians of gaps as you see gaps in that vision. So it’s wonderful that you’re thinking through I think this vision and it’s interesting because I think a few years ago when I was, had the pleasure and honor of chairing this group we really hadn’t quite, you know we were seeing almost all this in this big blob and I think what’s happening, we’re already starting to tease out the discussion in a way which it didn’t happen in the past.
In other words it is interesting, the Canadians talk about the info way, if you look at the NHII Workgroup from NCVHS it talks about knowledge and infrastructure and what we really are talking about is two related but different ends in view, you’ve got to have the if you will the track with a sufficiently gauged railroad, the cars can change and move on it, so a lot of the standards stuff, just going back to almost the 1870s when they’re deciding to decide what’s the size of a thread in a bolt, or doing that sort of stuff.
But then we’re also needing to figure out a way to look at the knowledge management side of this, what ends are we trying to see this whole massive thing achieve for individuals, for populations as well. So goodness knows, and it was beautiful, I think Martin Harris is a friend of both of ours, makes this wonderful statement, he says the interesting thing about IT systems is that you’ve got to make these changes while the train is moving, you’ve got to do the mechanical work on the care while it’s actually driving, you can’t put it off to the side of the road and put it up on the jacks for a few days.
So it’s really an interesting kind of challenge, just trying to deal with the mechanics of the data exchange kinds of issues, ignoring such wonderful things as the 330 million Europeans think it’s fine to have a unique personal identifier because they think that’s probably safer, have a picture of you because then I know what, I’m not going to give Gail this because I see her, I see what she looks like. We’ve put in a lot of additional kinds of strange sorts of things if you look at it from some world views. But at any rate, I don’t want to digress too much.
The point is I think that the NCVHS is one of the few sites that really can be I think a guardian of a vision that’s robust and makes sense as well as I say keeping an eye on some of the gaps. And so obviously we’re all creatures of our experiences and our biases and so where I come from, just so you’ll know where those of are, catalogue a few of them, had the great pleasure to chair the Institute of Medicine study in ’91 on the computer based patient record. And the subtitle of that report is an essential technology for health care, it’s still essential technology for health care, I mean we thought that in ’91, thought that before that, but every day it becomes more and more essential.
The thing that I find peculiar is how slow we are when we know through work like David Bates and others have done that the only way we’re going to make changes in quality is if we move to these system as fast as we can. We literally know that now so it’s fascinating to me how slow we are to try to make that transition. At least we’re in a different space then we were even three or four years ago and I think the internet is one of the key reasons for that, used to be we were pushing this issue and now we’re in a pull situation at least thank goodness. We still don’t have all the resources coming to it that we need but I mean that’s one of the issues is I think focusing on again, not the cost but the value that ultimately we’re talking about. So that’s part of that thing and that is an issue that we can’t afford not to do this, there’s no way you can come close to delivering care even today let alone tomorrow when we get all of this new science coming into play.
AMIA has now three major subdivisions as it looks at the field of informatics from the perspective of people who are in this game seriously and we have one area we call translational bioinformatics and that’s moving the knowledge that comes from the human genome research, proteomics, all of that business, the molecules up to clinical care, how do you move it into the care situation. And that’s a whole new domain of care and such as we get to personalized, individualized, kind of genetically relevant pharmacogenomics, pharmacogenetics, my drug just for me kinds of things, that’s going to start coming along so we got to think about, worry about that.
Clinical informatics, which obviously is both the decision support kinds of things that I think Brent was talking about as well as the infrastructure issues that John was talking about and then public health informatics, which looks at how does all this play out in terms of population health and as we think about this going forward with an internet rich environment where citizens increasingly will I think have the power, I think we’re going to finally see some of that, we’ve gone from a very massive mom and pop shop kind of organizational mode for health care in this country to something that I actually think ultimately will have the consumer or the patient clearly much more in charge and their loved ones. So those are sort of the domains.
I also have been shaped by the NHII Workgroup at the committee here and this idea of these three interlocking kinds of records and kinds of data exchange, the patient or their family or loved one, whoever is looking after them as well because that’s the information caregiver is usually important in our world, our situation. The clinician, whoever that clinician happens to be, or that system of clinicians, as well as then how do we get population data on different levels and different domains so that we can kind of play with that. So that structure’s important.
But I think Carol was talking about, I think we know this is a hugely complex environment and only biological kinds of ways of thinking about this will in fact work because that’s what we’re talking about, we’re dealing with macro biology in all of this and we have to have enough flexibility and creativity to allow some of these things to flow. So part of your task in maintaining that vision is to not let it rigidify to the point that it obviously, and that’s very tough in a bureaucracy, bureaucracies typically like Newtonian thinking, you lock it in, you agree on it, you go with it, instead of obviously complex adaptive systems kind of systems where it evolves over time.
So anyway, I think because of this complexity I do think that IOM set of variables, aims, is really a smart place to start. And in a matrix approach, try to look at how do you roll forward on that, moving toward this in that complexity zone, where you can try to put clarity to it but at least be tracking on things you think are really important, and they are important, this is safe for crying out loud. I think equity should be first because if you can’t get to it and it works all the rest of the stuff doesn’t matter, you didn’t get to it. I’ll have to admit I was on the committee and it was my five years in Europe that kind of got me to that way of thinking but at any rate, but I think equity is hugely important and I’m glad it’s on the list so that’s okay.
But at any rate, I would track on those variables, that’s influenced my thinking, and I think that finally they’re the wild cards, the gaps. One of the things that’s been fascinating about the way we’ve talked about the NHII to date is it’s really are, we talk about RIOs and we’re talking about how do we move these things forward, one thing that I don’t want this to take you off of your central focus but I think it is important, I heard at really one of those life changing talks by William Ho(?), he’s a friend of mine who’s now in charge of the hospital authority in Hong Kong. And I’d advised them on their system or their IT structure for Hong Kong a few years ago so we became friends.
Well at any rate, he actually was in charge of their IT system when Hong Kong got shut down with the SARS and he got SARS, he was actually running their IT system from an ICU bed and he had a bunch of people dying and so forth and it is an amazing story to hear this guy talking about this but at any rate. The point is one of the really interesting things, I thought these things were important to us as an IT issue and how do we run a health care system.
What became very clear listening to the William Ho talk was that yeah, that’s all true, it’s doing all those things, but what their IT system and their health system allowed them to do is reopen the port of Hong Kong. If they didn’t have, they built this thing to try to look after their IT system and their health system but it turns out when SARS came out of nowhere, and keep in mind when it came out they didn’t know what they were dealing with, and this became a totally political deal, just amazingly interesting to the media, to the whole society was terrified and how it should deal with this. But at any rate ultimately the upshot was if they hadn’t had an IT system that was quite robust they couldn’t have told Hong Kong when they could open their ports. That’s a lot of money every day folks as we know.
So one of the things that’s sort of fascinating to me is that we haven’t really thought about this NHII from the perspective of the world in which we live with things like Marburg(?), with things like Ebola, with things like bioterrorism, with things like the possibility of avian flu skipping from really birds to folks. And we I think really should be looking at this more also from a national security point of view.
I really do think that this is very much in the information super highway kind of issue, much like Eisenhower and the roads system post World War II, believe me if Seattle, Los Angeles, Chicago, Miami, any of our major cities got caught in a SARS deal, and these things are possible today, it’s totally possible today, if you keep those cities closed for about four or more days you’ve more then paid for a third of your state’s infrastructure. So I don’t think we’ve quite looked at this realistically in the context of what we’re exposed to because I’m absolutely convinced that the media will have us all for lunch afterwards, that why didn’t you tell us this, that in fact these systems could have really helped us in this sort of way. Of course hindsight is always terrific but I think part of what, and I don’t know the easiest way for NCVHS toward this issue, but I don’t think that that quite has gotten on the scanner in the way it deserves to be on the scanner.
And then I think I had one other little point, oh yeah, and it’s another little area where I think there’s some opportunity as you look at this interface because really what we’re talking now is how do we bring quality to IT, how do we bring safety to IT. In the past for about ten years after the ’91 to 2001, after the electronic computer based patient record, whatever you want to call it, report came out and went forward, we were having a hard time selling that frankly and it was when the errors report and the chasm report and we started saying well so what, the so what question, well why is this important. Then we started making some purchase and so the challenge we are at in part of this right now in terms of vision is how do we hook the values, how do we hook the return if you will to this and use IT as a tool because it’s ultimately a tool, it’s not an end in itself, it’s a means to ends. And I think that’s part of the reason that it’s now starting to capture our imaginations in moving forward.
So I think that we do have some things like the DOQIT program, the QIOs, there are ways that I think we are starting to see ways to try to tie the quality and the IT together. And I think that would be another one perhaps where you might consider having some hearings, trying to help, again, you want to think in the abstract but live in the concrete, how can you get some concrete things that will help you get some quick wins if you will and move some of this forward. There’s money out there, there are people out there, but they don’t typically have the informatics education that they need and such, AMIA is starting a program to help try to at least contribute to some of that educational need, but the point is that’s another little piece.
So I’ve talked very generally and tried to make a couple of little specific recommendations but why don’t I stop at this point then we can just ping back and forth --
DR. HALAMKA: I’ll just add to your specific recommendation, we did build a bioterrorism surveillance network in Massachusetts, it’s up and running today, and that’s a good concrete approach. Specifically at the point of registration in any of our emergency departments the chief complaint, the zip code where you live, the zip code where you work, not any names of course and the age of the individual is forward to the DPA in a real time fashion where then we do a control chart and say we expect in mid-winter, here’s the baseline level of respiratory complaints, ah, there’s not a peak of presentation of respiratory complaints to Watertown that we can’t explain.
Now this is an imperfect system admittedly, just to tell you because you guys live with data, during the Democratic National Convention a child presented to the Children’s Hospital Emergency Department with a nosebleed at the same moment that a gentleman was stabbed and had an abdominal wound at Boston Medical Center, at the same point a woman went into premature labor and had placental abruption at the Brigham, all this bleeding was going on everywhere and so oh my God, is this Ebola, but the nonetheless, the principle of syndromic surveillance with BPH central monitoring in real time is very concrete, we’ve done it and it works.
DR. JANES: Point of clarification, was that syndromic surveillance done with CDC funds? Is it the same syndromic surveillance that we mean when we talk about it at CDC?
DR. HALAMKA: Correct.
MR. HUNGATE: Good, discussion, overwhelming discussion. Go ahead.
DR. CARR: As you know I continue to be intrigued about ensuring that we have thought about how the electronic health record can capture quality so are you saying, John, that David Bates is working on this? I mean is it work that’s already being done and we’ll have a product soon, is it work that needs, that will be local or needs national attention?
DR. HALAMKA: So it’s work that’s purely local and it’s work that will be done over the next 18 months, and what will come out of it is the following, so we’ve implemented electronic health records to 900 doctors and interconnected three communities, didn’t make a difference, and so he’s attempting to with his team define certain process measures and today it’s so early I’m not certain what those are going to be, is it going to be diabetics have retinal exams more often or hemoglobin A1Cs fall or that mortality and morbidity are changing, don’t know yet. What I worry about, the problem that David Brailer has is that until David adjudicates national implementation guides for exchange of clinical data you’re going to have Massachusetts and Indiana and all these other places doing it on their own and in fact in the short term will lead to more chaos then order. But we’ll learn something along the way and what I will hope is that yes, we’ll take whatever David discovers, share it broadly with you, but boy, if there was a national mandate that said here are the measures, here are the data standards, here are the policies, I’d much rather use yours then reinvent something locally.
DR. DETMER: I think to get back to Brent James’ comments this morning, I think the Northern New England Cardiovascular Group is probably the best working knowledge management outfit going now that relates to cardiovascular disease, not everything, but I think as a model of what the future will look like and I’m convinced that that’s a model, if you haven’t had them in to talk to you about what they do have them in. But I think Brent is talking about also how do we move forward on some of these national protocols that aren’t inflexible but at least allow a protocol that allows you to be adaptive to it over time, using evidence, and moving it forward. So I would look at how do you put in the right kind of processes to see truth emerge because nothing, in nature nothing is free as Emerson said, you have to pay for it and you’ve got to pull it out of there.
But I think the point is, I think the Northern New England Cardiovascular Group is probably the best model on the fly. I think as far as these electronic personal health records, John didn’t talk about that, I think he should talk about that, because I think the link, the clicks and mortar kind of care that can occur between patients and their clinicians, because right now most of what Brent’s talking about and the Northern New England group is working with clinicians on decision support. I think it’s just as likely for simple depression, for panic disorders, for phobias, what Aaron Beck’s work in cognitive psychology has shown, that we can have clinical decision support systems for citizens to use themselves off the web and in fact be as effective as seeing most GPs for that, I mean good family practice people, there are randomized trials that show this --
DR. CARR: Can you talk about that, John? I mean what the capabilities are today?
DR. HALAMKA: So in 2001 we implemented patient site, you’ll find it at www.patientsite.org, and made it available to the two and a half million patients of CareGroup. Patientsite is a combination of a secure doctor/patient clinical messaging system so that doctors and patients can now use real time electronic means to send information back and forth, it becomes part of the permanent medical record so for 30 years we’ll archive all those transactions. It opens up the entire electronic health care record system of all of our inpatient and outpatient facilities to the patient so that the patient now sees the same data that the doctor sees with a few small caveats, which is we don’t release psychology and pathology results to patients for two weeks so that there’s a chance that the patient and doctor can have a private conversations. You don’t want to go to the web to find out your cancer diagnosis. Similarly we withhold HIV results, first time HIV results, thereafter we do show CD4, viral loads or other kinds of measures. And certain tests that are used in the staging for progression of cancer we typically hold for a week before the patient sees them. But beyond that every bit of clinical data is available to the patient.
Additionally each patient gets a personal health record and what that means is that they have the capacity to not change their record but add to their problem list, add to their medication list, add to their allergy list, and part of that is decision support. So there is a drug/drug interaction check that is the entire inpatient, outpatient, ED, over the counter, any medication that you may have had and/or taking, and by the way it’s the St. Johns Wort that you bought at CVS, all of that you can now manage and do drug/drug interaction checking yourself.
What’s been fascinating I’ll tell you is that this has truly implemented a shared decision making approach between our doctors and our patients. A story, if you want to talk to the patient I’m sure they’d be happy to testify to this committee, but one of our patients went in and discovered the doctors had made a mistake in the documentation for reporting the size of her tumor because she has all her radiology reports and all of her ultrasounds and all of the clinical notes. She was scheduled for surgery to deal with this rapidly expanding tumor that in fact was rapidly expanding only because of typographical errors, it had not changed in size. And this is what happens when you empower patients and their doctors to share their data in a seamless fashion and implement a personal health record.
DR. MCCALL: What’s been the reaction on the part of the different constituencies? Just kind of the run the usual suspects.
DR. HALAMKA: So from a patient perspective of course they love it because now I can talk to my doctor, I can see my data, I can refill my prescriptions, etc. The doctors like any change management fall into sort of three categories, the early adopters, wow, this is cool, I want it wirelessly, I love it. The well, I’m a little reluctant but I’ll try it and the I’ll never do that because it’s going to in fact bury me, I’m now going to get emails from patients at 3:00 in the morning describing every medical problem they’ve had since the age of four.
In fact we have not seen cyber-chondria, that is something that in the millions of transactions we’ve done since 2001, appropriate interaction between doctor and patient, the statistics, on average a patient does 1.2 transactions per month when they have a completely e-enabled doctor and 90 percent of those transactions are triagable to medication refill clerks, appointment clerks, those types of folks. This means that an average doc with a busy practice has five to ten clinical messages a day which now replace phone calls, and so the doctors actually describe because it’s now asynchronous, I don’t have to try to play telephone tag at lunchtime, that they actually save time everyday, it has not at all been that worrisome my practice will change kind of event that was forecasted.
MS. POKER: How many people have access to respond? In other words are you having a lot of users or are you just having a certain --
DR. HALAMKA: On average 40,000 unique patients use it every month.
DR. DETMER: And across the country we’re approaching, we’re in the probably low millions or will be shortly I mean if you look across the country in these things. So this is not, by all American standards we’re not there but I mean this isn’t just a couple, the VA is starting a system on this, there’s some of the vendor systems actually are pretty robust models of this and I guess one question that I want to ask you, are you able to look at some of the patient data to get insights for population management as opposed to just doing the clinical, the doctor’s data, because that’s another piece of this three way circle diagram and I’d be interested in your response to that.
DR. HALAMKA: So all the data, there’s quantitative and qualitative data so for example for depression we may have a LIKERT Scale on mood or migraine, a pain scale, these kinds of things. And then there’s home glucometers where we’re actually getting quantitative data from the health buddy, health your own networks, interfaces, you actually can look at glucose control and I have some interesting anecdotal evidence where you can see that there are populations that are really engaged with the doctors in glucose control, the cycle time for drug adjustment when they see variations in glucose is now rapidly reduced because it’s done via email in near real time as opposed to can I have an appointment in three weeks. So yeah, the technical capacity is there, the only other caveat is obviously the data has mixed quality.
MS. MCCALL: Tell me what do you mean, it’s mixed quality --
DR. HALAMKA: Well, meaning that a quantitative bit of data that comes out of a health hero device, that’s good. But where you’re asking the patient to type in a problem list or a medical it’s a little --
DR. DETMER: But we’re still early days.
MS. MCCALL: That’s fine, I just wanted to make sure I understood which one was causing the problem.
Now what about RIOs in other parts of the country? These things are tough to grow --
DR. DETMER: Well, we went through this CHINS(?) some years ago and some of them took it on the chin and some, I mean I think it is a real challenge, a lot of states are getting into this now, but the thing that’s fascinating about states getting into it, about half the U.S. population lives near a state border and so again if you’re in the Washington area for example you may live in Virginia, get your care in Maryland and your insurer is in the District. So unless you do deal with these interoperability issues you really can have problems.
So I do think that thank goodness we are finally getting, and I will say this, Secretary Leavitt really understands this stuff, Secretary Thompson was very supportive, so at least we now have that and some of the governors are starting to clearly, and states are starting to get into this. But the point is is that it does need to come together, I mean you have a lot of touch with New Hampshire and Connecticut and Maine and ultimately health care is local but it’s not totally local, it depends on where you live and so it’s one of these issues that clearly is going to take some care and feeding. On the other hand how much of that is really, I would rather frankly have you worry about the bigger vision and the intersective quality in IT then really being somebody that’s primarily worrying about how the RIOs are rolling out.
MS. MCCALL: I would agree with you. This is absolutely fascinating, equally so to what we talked about this morning. Just in reaction to some of your comments, Don, about basically being the guardians of the vision and I believe you’re right, and so I think that in order to have an impact, and when I look at all the different, kind of the full committee and the different subcommittees, that there is a unique opportunity here, we are at an inflection point, and I think that we can have an influence and I think that the choice is ours to have that, nobody is stopping us.
The other thing is that this is now the time for us to do that, the entire thing is at an inflection point and when things are and they’re new and emerging that is a time that we, to actually, if we were in private industry we’d say that’s the time to actually go rub against some policy makers, because that’s when the rules of the game get set. But we are the policy makers, in fact we influence the policy makers, and so now is the time for those rules to get set to enable it to come into being but to come into being in a way that actually takes a shape that we had intended. So now is the time and so I’m very, I aspire to do what Simon had charged us to do which is to come up with something that is actionable and I think that we should.
I also don’t want us to strive for actionability and strain for it so greatly that we miss what the real opportunity is here so I think we need to look for both. And so you started out Don by asking your question which is what do we do in the meantime and I actually think that is a very important question for us to say how can we get from A to B, I can see B, I just don’t know how to get there from here in a way that actually has meaning along the way. So I think that’s how we can make it actionable for today, I just loved your comments about us being the guardian, I think we can.
DR. CARR: Could I go back to your comments about Northern New England? We actually were members of that and I have great appreciation for it but I’m not sure which aspect you’re saying should be emulated.
DR. DETMER: There are a variety of issues, I mean I think one set of issues is also looking at how did it get to be because part of what you’re talking in his situation this is a huge amount of culture change and change management that continues to this hour and so part of what I think what you’ve got to look at is also how do you look at the change management dimension of all this, that’s part of that complex adaptive system kind of issue and there are lessons there. So there’s a set of issues that relates to sociology of change in the direction you’re trying to go that I think it’s an important example of.
But what is it now, it’s a multi-institution, large and small, that shares set data amongst these people to track very specific interventions and processes, giving you protocols to guide your thinking but allowing you to override that in a particular instance where you think your clinical judgment warrants that. But then it tracks how the patient did if you did override it or if you didn’t override it, so that then later you can apply statistical analysis to the dataset but also clinical analysis to the dataset and look at it and say well, we had X number of overrides in five different sites and it looks like from the data that they were right to override it two thirds of the time.
Now the issue is the difference is a difference if it makes a difference, how do we make it clear when it was, should be overridden and set up a sub-protocol to deal with those should be overridden kinds of situations and how you develop a flag where you tell, help these folks realize in this circumstance you shouldn’t override it because actually you did override it but you shouldn’t have, actually the outcome wasn’t there. So what I’m saying is you’re really wanting triple loop learning kinds of situations, where somebody can do something, you learn from it, but the system learns too, we’re not all just doing this in our little boxes, the system is able to grow over time too and you’re able to build more and more robust things.
And it’s like Brent James says, well, it’s not like every delivery is unique in the whole country, we’ve been having deliveries for millennia, not millennia but a long time, 20,000 years or whatever. At any rate, the point I’m making is it’s hardly new events. Now on the other hand we do know that in Southern Utah a delivery isn’t just necessarily a delivery, there’s also some spins in that, so there is going to be a local tweaking that also does need to occur, some of that will teach you some things you need to learn about population management but the point is the architecture needs to be set and set up in such a way that you can share the data interoperably, try to get as fine grain as you can, one of the tensions that we still get into these arcane debates about is free text versus structured data entry, and the terminology and all that business, but it’s making progress, SNOMED has helped and some things are coming along. So I guess does that help?
DR. CARR: Yeah, I mean, it’s big --
DR. DETMER: Well, that’s the vision, but it’s understandable --
DR. CARR: No, no, no, I agree with you, so I’m just trying to pull together the theme, so we’re saying that, well let’s say we can look at outcomes for cardiac surgery today but that’s all we can say, we don’t know why they’re there --
DR. DETMER: Fair enough, what I’m saying is what I’d like to do is get us from if you look at the old style, you see what I’m saying, you look at vital statistics and see who died, well it’s a little late to help them. So what we’re trying to do is move further upstream increasingly, is try to go to the point of decision. Well, it turns out the point of decision actually was whether, in the Southern Utah situation is whether I live in a certain kind of cultural context may actually have a lot to do with my health and that has an education issue. So what I’m saying is we’d like to move where the end is not health care, the end is health, and how do we build better and better intelligence and decision support systems to have the person in play, the person, the citizen in play, the citizen at times, when they’re not sick, stay healthy, but the citizen, there’s this great quote of Jefferson’s, if you don’t like what people think inform their discretion, we’re really not informing their discretion as much as we could and I think IT and the internet gives us the capability of sort of doing that, but a lot of this as Brent was saying, we don’t know what we know and what we don’t know, we haven’t even systematically sort of done that. So that’s part of this too, that’s what that --
DR. CARR: But if we were going to break this down into steps we would say it’s one thing okay we have an outcome but it isn’t very helpful to us unless we understand what contributed to the outcome, so when we do risk adjustment you say well a real sick patient but you’re saying there are other factors in terms of did you give an aspirin, did you give a beta blocker, and that’s a part of it. And then a final part is the examination of whether the assumption is a valid assumption.
DR. DETMER: Sure, and then I guess one way I think for example, one of these come to a fork in the road and take it kind of questions, one debate that always goes on when I was in NCVHS at least, I’m sure it still goes on, is do we go with a “core data set” or not. Well, in a way a core data set assumes you know what ought to be in that core data and by definition in a complex of systems most of us don’t have that Solomonic knowledge, it gives out of an extremely complex kind of thing and unless you really have that infrastructure, I’m not against to the end of time certain small datasets but let’s derive them off of what in fact science suggests to us we are warranted to do that for. And so these people that argue for a certain kind of summary, clinical summary or something.
Well, it may be extremely useful for a lot of patients but as a specialist, I practiced vascular surgery for 25 years and I did a little boutique kind of care called chronic compartment syndrome, I can tell you those data summaries would not help me a bit for the kind of patients that I literally took care of and if I was going to help them I had to get this other arcane stuff. So all I guess I’m saying is is that we have to have a robust infrastructure, we may pull pieces out of that but let’s do it because we know it from actually work itself if that’s helpful.
DR. CARR: I’m just trying to put it into concrete steps, I’m looking at what, like if we were to move this forward would we --
DR. DETMER: Well, I think Brent was giving you some very good advice.
MR. HUNGATE: I thought so, too.
DR. DETMER: I think you got some very good suggestions that I’ve heard today on some very workable, arguable kinds of things, to work with AHRQ and I think some things again with CMS and so forth on a variety of these issues.
MS. MCCALL: Just to add on and I think like Justine, so what do we actually have to do, what do we need to do, and I think there are some very clear things to do and I do think they come from what we’ve heard today and one of them from Brent’s talk is, and you actually had some slides as well, Don, and I don’t know if we have copies of them --
DR. DETMER: Yeah, I’ll get them to you.
MS. MCCALL: And I think they have to do with these outcome chains, I think with yours about starting with what you know and mapping to the more complex, we start with what we know and I think we start with what is clear and I think we actually have some very good guidance from the IOM on where are the greatest burdens, to people, by condition, so basically we’re going to go where the light is bright and we’re going to start with not core metrics, let’s just call it a starter set, I feel like I need a little starter kit, and then we need an infrastructure that can actually also grow.
There’s this great story, I don’t know how many of you have actually heard of an author, Kevin Kelly, he wrote a book called Out of Control and it’s about complex adaptive systems and things like that and he has a great story in there about a gentleman who tried to reconstruct a prairie, and in trying to reconstruct the prairie he gathered from different sources of knowledge all the stuff that was going to grow there and he kind of took it all out and scattered it around and some things grew but he didn’t get the prairie, he got something else. And what he found out among many things was that the order in which these various species were introduced mattered greatly because something that actually required three or four things to already be there so it could survive wouldn’t survive if it was introduced first.
And so I think we also need to pay attention to the order of things, we may see what we ultimately have as a vision but that’s what I think from RIO and other examples, there’s a culture, you had to some change and probably what you learned and may not have known that you learned but at the time is what the order of things had to be in. And so I think order is going to matter, I think we can get specific with a starter set, I think the infrastructure is absolutely critical, there’s a whole host of areas but I think, Justine, I think we can get really specific, it’s going to be delightful.
DR. HALAMKA: And I’ll get specific in that what you had in the past was administrative data so you had your UB, your diagnosis procedure kind of stuff. Well I know where the patient has been, I know their problem list, I know their medication list, I know their allergy history, I have free text notes to the extent that’s helpful, it helps you and I in clinical care, whether you can do anything with that data from an analytical standpoint hard to know, I have linked and coded labs, microbiology data, and again this is probably free text but say radiology interpretation. So if you now had instead of just administration data, the whole medication/allergy visits, radiology, lab, what do you think we should do with that, is there interesting low hanging fruit that we could --
MS. MCCALL: There’s tons.
DR. DETMER: Actually there’s an NIH study, a meeting, two days, at the end of next week, I’ll be talking about my own views on what the research agenda is for informatics, applied informatics in some of this stuff. So I mean there are other things going on that you can tap into that I think could help you as well.
MS. MCCALL: Right, I was thinking not so much about what we can do with the data that you’ve got, the question is how to get where you are, and what was the order in which things needed to happen --
DR. DETMER: Well, in fact that’s what I was speaking to, is how do you get these things moving, because clearly we need to also get this in the water supply of the whole country and we’re going to be forever fiddling around getting there otherwise. We need to put some movement to it.
MS. MCCALL: Is there anyone who has a strong picture of what we can do, even prior to having that, is there something we can do with the administrative data that we have that’s better then what we have today, that doesn’t move us in a direction that’s different from our ultimate vision, I at least want us to stay in the same hemisphere, so that’s another question around actionability, is there something that we could do with what we’ve got to either enable the growing of whatever these infrastructures are or to enable the creation of some metric sets and some standards, or both.
DR. CARR: I think that’s what CMS and AHRQ are doing, I think they’re really pushing it to really sort of aggregating data and at least on the AHRQ thing saying what’s the signal to noise ratio. But it’s very granular, it’s something but I mean I’m not sure if they’re exhausting the capabilities of that.
DR. SCANLON: There’s also another strain which has come up in other context and which goes to the dissemination, and it’s the idea that if we are thinking about pay for performance and we want sort of measures for pay for performance, sort of how do we get them and we most likely get them through administrative data. But then how do we get there, do we go and then we fund the development of IT or the dissemination of IT, or do we just demand this information and that the easiest thing for providers to do is to adopt the IT, to recognize that its become a routine part of business and that they need, if they want to participate in this system they need to be able to submit information and that they take this path of least resistance. It’s not clear what’s going to be the best way for dissemination.
DR. DETMER: The commission is talking through that one right now.
DR. HALAMKA: A couple of concrete steps that we’ve taken, we recognize that the only way we could start getting the data in is to have a little bit more homogeneity in the data capture mechanism so as a state we went through 133 functional criteria and these were pulled off CCHIT and HL7 and other organizations that have been thinking about how electronic health records should capture data. And we then submitted this RFP to 133 medical record vendors and picked seven that we felt could adequately capture all the data elements that would be necessary to do good clinical care and ultimately quality and performance analysis.
And so companies like NextGen and GE and EPIC and the various Allscripts, those sorts of folks, are now part of our Good Housekeeping seal of approval, EMR, so that when the doctor says hmm, I want to get into this EMR thing, where do I start, it’s well, you need to capture this data if you’re going to play and here are the vendors who can do it for you and by the way we have statewide special deals so you can now start participating in this. Some of these are ASP models so all you need is a web browser, so we’ve really reduced the barrier to entry. Other thing, statewide privacy policies and working groups, so these, exactly as you described the development of the prairie, if you build the grid, you give people help with the last mile, you build incentives to do it, these are all necessary steps.
DR. DETMER: Predictable cost, I mean there are a handful of these variables depending on which part of the system you’re in, if you’re a hospital you have one set of them, if you’re typically a very small office you have another. Now there’s still going to be unique bells and whistles that are also there but having said that there are some common things that we do know now that really do help in what order and those sorts of things.
Actually just as I’m saying, one of the things the commission is moving probably to, is thinking pretty seriously about, is looking at e-prescribing as a possible place to try to move the system generally across the country because it is something that is reasonably complicated although it’s very important and it does give you the leverage of some of the return on investment savings from drug/drug interaction kinds of things and quality/safety kinds of issues. So it lines up on a number of variables that you’d like to optimize against.
DR. HALAMKA: And so to that point, just in terms of the prairie again, when we tell our doctors you should go do e-prescribing they say well let’s see, Zicks(?) Corp, tells me I can go buy this product and it will work on my Blackberry and then Dr. First tells me I can use the web and Rx-Hub is telling me the solution for formulary and then SureScrips has the retail pharmacies, I have no idea what you mean by e-prescribing. And so as a community we’ve created this utility and as I said we’ll have this up in the spring where all of those various functions that constitute the world of e-prescribing can be consumed in one place.
DR. DETMER: It might be good for the group actually to visit John, I mean he does it, I talk about it, but I mean I go to some places, I think that really could be useful, Kaiser, VA, whatever, but I mean I really think that unfortunately these are not things you can imagine and if you just go to a place that has one piece of it you don’t see what it is if you go to a place where it’s really a filled out thing. It’s a different way of living, it’s a different way of thinking, it’s a different way of doing care and it’s sort of something you can’t kind of imagine if you haven’t actually seen.
MS. POKER: I just have one question from especially you, the panelists, and that is with the excitement of the value that we talked about earlier and John especially when you presented that the cost of a transaction or community RHIO that you’re doing, went down from $5 dollars per transactions to ten cents, I mean that kind of blew me away. So it is something that we’d want to look at or is this something that you would recommend for us to even consider to look at the value of the RHIO because, it made you laugh but I mean, I mean that is really very impressive, if we could impact and get the same outcome --
DR. DETMER: Well it costs a lot to run what he’s doing too, I think the key issue is it’s the only way you can run it going forward, I mean I don’t, it’s a question of you can’t be in the business any other way if you’re going to truly be really doing what in fact the population deserves and we can now do. So in a way I hear you, personally I ultimately I guess I’d rather spend your time and energy making sure you got a good vision and you’re tracking it and you’re worrying about some of the important side issues.
DR. HALAMKA: To get started to your point, we had to actually do this, we studied the health care environment for Massachusetts and determined that roughly 15 percent of the health care delivered in our state is redundancy and waste and so $30 billion is spent per year so $4.5 billion dollars is a redundant testing and inappropriate care. So we went to Blue Cross and said you as our primary insurer are probably spending about $2.5 million dollars on us on unnecessary care, would you give us a billion to build out the entire medical record infrastructure for the whole state. They said no, we won’t give you the billion, but we’ll give you $50 million so you can prove that it truly does create value. And so we’ll see, it’s a microcosm of the whole country but with Newberry Port, North Adams and Brockton as three communities with 900 doctors over the course of the next 18 months you will have very detailed analysis on what the RIO was for those communities.
DR. DETMER: Blackford Middleton has done some interesting work in Boston related to Partners work and one of his studies, and again, I think there is some value to have certainly some of that research going on, I don’t want you to mis-hear me. One of the interesting things that I think he pointed out is that you can do EHR light and you can do EHR with an infrastructure, and EHR light, okay, versus EHR done right, you have to spend four times the money to do it right but you get 12 times the return on investment. Now even if his numbers are off a little bit it argues that you got to start somewhere so it’s smart to start someplace but start with the vision that’s going to build the thing out to get you your return on investment from the knowledge support because the knowledge management piece is where you really start getting your leverage, that’s where you really start getting your leverage, that’s where you really start saving lives and saving money.
MS. MCCALL: Going back to some of what Brent talked about this morning, kind of that picture of some of that spectrum, I guess my question for folks is how much of the type of grid, and the analogy is perfect, it is a grid just like the electrical grid to plug into, how much of that grid is necessary in order to work in some of these different areas along the spectrum? That’s the question I think that we need to know the answer to in order for us to push forward looking at definitions of quality and value and things like that so that we know how much a prerequisite that type of system is in order to do certain types of things.
DR. DETMER: Which slide are you on?
MS. MCCALL: Any one that has that little continuum, it could be the first one or the second to the last --
DR. HALAMKA: Certainly I think it’s very challenging to do any analysis across a region without knowing who the patient is because if you find that they had a treatment at this clinic and then they went to the hospital and this bad stuff happened and they had other treatments, so you really need to track all events around an individual so I would argue that some type of master/patient index is a prerequisite. And then certainly at least some method --
DR. DETMER: Which is patient centered, that’s patient centered care, it meets that IOM criteria.
DR. HALAMKA: And certainly whether it’s real time data exchange as we do for clinical care or a batch every month where we get some data elements that can be analyzed, that’s obviously --
DR. DETMER: Plus looking for effectiveness and efficiency --
MS. MCCALL: Well, I would assume that you could do almost all of this, you guys are really poised because you’ve built it from right to left, are poised to actually add some kind of new analytics and all different kinds of fun things to get into that learning phase, you’re ready for those second and third order gains.
I guess what I’m thinking about is this, you’re head and shoulders above everybody else, leave out the rest of the country, think about what is happening with if you’re familiar with Care Focus Purchasing and some of what AHIP is doing now, imagine that all administrative data could theoretically be put in one pot and it had, you could actually tell that a doc was a doc was a doc was a doc and a person was a person was a person, how much of this could you do? How much of this could we do, could we use it for payment? Could we go that far? Does it get it EHR light? Can we produce decent metrics around quality?
DR. COHN: Carol? Let me just ask because what is your definition of the dataset that you’re talking about, are you talking about just diagnosis and procedure or are you talking about diagnosis, procedure, lab and pharmacy? What is your definition --
MS. MCCALL: Diagnosis, procedure, lab tests but probably not results or it’d be Swiss cheese, pharmacy, so claims data --
DR. COHN: So medications in other words --
MS. MCCALL: I beg your pardon?
DR. COHN: Talking about medications.
MS. MCCALL: I’m talking about medications, anything I can get on a claim.
DR. DETMER: I guess in the near term frankly in a pay for performance kind of thing I probably just as soon put on a rider that allows them to charge off getting into a computer based system for three years on one of his approved systems so that it definitely isn’t just my own backyard but really does link to a system. So I’d get them into play personally before, I don’t know if that’s responsive to your question but that’s, I’m not sure I’ve answered your question but if I were going to be doing things I’d try to get the system, but, but, according to some working guidelines, I mean not just sort of let them buy something, I mean let them buy something that’s going to get them something.
DR. HALAMKA: And I think it’s a phased continuum which is obviously if administrative data is what you’ve got then of course try to optimize, especially if you’ve got pharmacy claims that’s going to add richness. But in my case I run all the hospitals based on dashboards with metrics that are process driven and I’m looking at clinical data elements driving all of our decision making for which administrative data is almost useless so obviously deal with what you’ve got but go beyond administrative data as soon as you can.
MS. MCCALL: Just like you said, there’s a real EHR and then there’s EHR light, there’s a real metric set and then is there a metric set light because if there’s not then I would submit that our actions may not be focused, we’ll have to have a deeper discussion but it may be that our actions are not focused on dealing with the here and now. Because to put a buff and polish on administrative data may not add any value to the long term, all it does is detract time, money, energy from that. So I think we need to know the answer to that question so that we can decide where to put our priorities, where to spend our time, our money, our energy.
DR. SCANLON: I think one of the things that in again other contexts, I’m talking about pay for performance, I mean the issues have been that administrative data already sort of in terms of current measures needs some type of supplementation and that we need, we would need to think about working there. But then you get more to I think which is the more fundamental issue which is the measures that we have. And particularly for the physicians we have the problem of accountability, we have the problem of the appropriateness of the measures by specialty, and we’re not really very far along on either of those dimensions I think, I mean at least from what I have seen. And that’s a big issue in terms of if we think about trying to improve administrative data are we improving it in a way that we’re going to be able to use it when we do get those measures so I’m thinking metrics first.
DR. DETMER: Yeah, I guess to take one more run pass at that, if you had administrative data that relate to those things that where we really know evidence based processes that make a difference. In other words I wouldn’t just say administrative data, I mean I’d start it on things that may give you a better signal to noise ratio then that because we already know that we have so many wrong incentives, I mean we have a really huge procedural bias in our health system right now. So I think to just look at administrative data, I mean we’ve really got to really look at how we’re rewarding things based on what we know gives us an ROI in both satisfaction as well as --
MS. MCCALL: Imagine the following, imagine a world where people are paid in different, kind of different levels for different things, and the absolute highest is where I as a provider use the type I have available and I use and I honor the type of grid that you have and I use it appropriately and I’m doing the right things. So I use the system and I have good metrics, that’s like the top.
The next is I don’t have that kind of thing available but somebody has actually put in on a UB-04 a couple of enhanced measures and so from a reduced administrative set that are meaningful and pithy and have clear value I’m doing well on those and so I can kind of get paid maybe the middle level.
And the third level is I’m not using the grid, I’m not even performing up to standards on whatever those clear metrics are or I’m not submitting the data, and that could be kind of the lowest level in a P for P world. But I’m really concerned about only focusing on this future state because pay for performance is here, it’s going to need crisp and clear metrics, I think that that’s also part of our charge is to opine on what those can be or to put in place a recommended structure --
DR. DETMER: See I got to get more docs to cross the threshold into this way of practicing, particularly small clinics, I mean in the UK, in Europe, it’s the flip side of the U.S., about 97 percent of docs are using electronics in their offices. That’s very interesting, it’s just that’s the way they started, their hospital systems are really very weak compared to this country. But increasingly we’ve got a big challenge is to get people to cross into the threshold, once they’re into play I don’t think you have to keep rewarding them, that’s just the way I live now, but you’ve got to get them into that thing and that’s why I think some of it is how do we, because that’s a major challenge for us actually on the NHII.
DR. HALAMKA: Let me describe the barriers quantitatively that we’ve seen, on average before we started doing a lot of these community wide EHR efforts it cost a doctor $10,000 dollars to acquire an electronic medical record system. On average they have a 25 percent productivity loss for the first three to six months they use it because it’s just simply more time consuming to put documents and controlled vocabulary --
DR. DETMER: It’s a lot of change, it’s a huge change.
DR. HALAMKA: -- typing and these kinds of things. And then when you look at the downstream benefit, let’s see, they’re ordering less tests and gee, quality is going up, does the doctor get paid more for that? Well, actually no, in fact it’s Blue Cross whose now got record profits. So wait a minute, I’m investing $10,000 dollars, I’m losing 25 percent productivity and the payer is getting 90 percent of the benefits. Now why should I do this again?
DR. DETMER: That’s why the perversity in this thing.
DR. HALAMKA: And so hence doing this pay for performance where we’re now going to get the doctor the systems they need, subsidized, and incent them to use it, that’s the only way to cross --
DR. DETMER: But you don’t need to do it forever, I mean once they’re actually into it then I think they’re doing it.
MS. MCCALL: I would agree with you.
DR. COHN: First of all I want to apologize, I thought I was going to make it in time to hear them both present but obviously they went early so my apologies to both of you. Obviously I’m not someone who has to be persuaded about the significant value of electronic health records but I was actually just trying in some ways to think about Carol’s question which is sort of this, there’s this sort of, even if we were to move full speed ahead with full scale implementation knowing the sort of that the gap of the ability of people out there to actually implement, we know that there’s going to be X number of years, we know there’s a very real need for pay for performance. Also nothing that we have one single payer at least at the table, CMS, that is going to have actually a fair amount of information.
And so the question is what for people like that or people who have maybe more comprehensive use, what can you really do and then your term that might help us incent until we sort of move forward and obviously, I mean claims data, I used to make a big differentiation between claims and non-claims data, there is some value to claims data, I mean it has diagnosis and procedure and at the end of the day if I’m an emergency physician and it’s at night and that’s all I have I’m thankful. So I mean I think that there is some value there.
Interestingly enough obviously now our largest payer and certainly Humana and Kaiser and others, even for people that are not cared for by our physicians, we typically have access to pharmacy data on them and indeed outside as we’re talking about e-prescribing, there’s one group, all the MedCo PBMs and all that have tremendous data stores of pharmacy data and so suddenly geez there’s potential for leveraging pharmacy data.
Now the piece that I don’t understand and John might be able to help me with this one a little bit is that the one big piece in all of this stuff that I keep finding people are not somehow getting is lab data, and you may have mentioned that already and I will apologize if I’m being redundant, but people talk a lot about well geez HCVA 1500s where they ordered a test but you’d also have to get the value back. I know you’re actually getting the values through your network and all that and I guess for the life of me I can’t understand, well geez, it’s nice to know someone got a hemoglobin A1C but I really would like to know the value, if I’m talking about pay for performance or management, I mean Clem McDonald, bless his heart he’s not here today but as one of our chief lab visionaries would really be observing that lab data is in many ways a better surrogate of diagnoses then diagnoses. And now we can go down there for a whole lot of things and EHRs are critical for being able for decision support and all of that.
But in terms of a pretty reasonable dataset, I guess the one big piece I wonder why we’re not, what’s so hard about is this lab data. John can you comment on this one?
DR. HALAMKA: The reason it’s hard is linked vocabularies have not been universally adopted so it is a CBC varies at each lab and institution, reference ranges are highly variable across various labs and institutions. And just the nature of the data is very complex data, some of its quantitative, some of its qualitative, if you look at the way, it’s not that you have one lab test with one value, you have panels with multiple values, so when you look at even just the standards to represent it HL7 2.X and now HL7 3.0 reference implementation, it’s just one of the more complex kinds of data to transmit and because of the lack of vocabulary and difference of standards and reference ranges you haven’t seen its ubiquity yet.
DR. DETMER: What’s really bizarre to me in a way is that we haven’t stepped back and said look, why don’t we rethink this whole thing and instead of millimoles(?) percent and so forth, let’s put it in to what is the range of where there is 90 percent of normal or 50 percent and flip the whole thing into something so the person, you don’t have to go through all this interpretation, I mean it’s really sort of fascinating to me that we just haven’t made these things more usable, but it’s all back to the way it sort of got discovered in 1830 or 1920 or whenever, it’s just sort of fascinating to me. And all that having said I’m extremely respectful for what Clem and his group, LOINC and all his efforts to try to move it along. It is fascinating in the personal health record data that patients love lab results, they just love it, I mean any data they just absolutely love to get to, what their lab, they had their blood drawn, well then what do you make of it --
DR. HALAMKA: Right, well the caveat with that is we’ve had to wrap all that patient exposed lab data and turn it into knowledge, so we use a publicly available website called labtestsonline.org to say oh, your cholesterol is 200, why should you care.
DR. COHN: Can I ask one follow-on, and Trent, I hope you don’t mind if I put you on the spot here again but I guess as I look at your datasets, and you’re obviously an important player at the table, you’re talking about value based purchasing and all things like that, you’ve got diagnoses, you’ve got procedural data generally, you’re going to have pharmacy data very soon with the Part D benefit, you’re on the project plan either getting or doing something with lab data or is that part of the vision of moving forward with any of this?
DR. HAYWOOD: I mean there’s been, to say it briefly it’s on our radar screen, there have been MedPac recommendation regarding to providing that information, so we’ve taken that under consideration. Obviously what we’re struggling with is similar to our other kind of data collection vehicles that how do we actually start to link it up in a way that is actually usable. And so the short answer is yes, we’re looking at that particularly as you were talking about pay for performance, where you really want to be able to show some level of intermediary outcomes or performance along those lines. So we fully anticipate that we will be trying to get down the road and it’s just a matter of how and when we can actually do that but it’s definitely an area that we’re trying to pursue.
I should just quickly say just on this issue though, that I think Carol was raising, I couldn’t tell if she was saying that it was a dichotomy meaning that administrative versus clinical data, I think the reality is probably like word processing outcomes, that for the short term you’re going to have some overlap or a combination thereof because on the one hand you’re looking at it from the standpoint of really trying to provide enough information for clinicians to take care of patients, on the other hand you’re looking at it from the end point of having signal or sense of the services that are being provided, whether or not people are getting a particular service and how they’re being provided. And so I think the reality is in the short term that you’re going to have some administrative activity, that you really are going to have some of that administrative activity at least for the purposes of being able to know how well you’re utilizing some of your resources, or resources are being paid for.
But the key in doing that I think from our perspective is doing it in such a way that it doesn’t undermine our ideal of really being able to use clinical data and so to the extent, and keeping in mind one of the things, even though some clinicians would obviously argue you got to get clinical data, on the other hand when push comes to shove as to recognizing the burdens around that they equally come back and say well maybe it may not be bad since I don’t have the infrastructure in place to start with something along the lines of administrative data, recognizing that underneath that there is full or measurement sets and clinical data that actually can be obtained. And so I think internally what we’ve talked about is the notion of whether or not you do that in such a way that you may have to allow for some use of administrative data, short term but the longer term goals is to really move towards clinical data.
DR. JENCKS: It would be helpful I think, it would be very helpful to me to understand what your vision is of how soon recommendations coming out of this committee, starting with today as sort of the beginning of thinking about them, would take to influence the health care system. I ask this in a very practical way, I suspect we’ll talk more about this tomorrow, because I’m not convinced that you’re going to be able to have an impact on the health care system soon enough to make investing appreciable effort in claims based or quasi administrative data useful. And that doesn’t mean that Trent isn’t going to be building pay for performance systems around those data and a whole bunch of things but the question is how soon can this group have an impact.
MR. HUNGATE: That’s a good question.
MS. MCCALL: I think that’s the right question. If we were to take that next step further then, Trent you’re going to do a whole bunch of stuff around P for P, Godspeed to you. Is there a role that we have in opining on metrics, that when people do use P for P off of administrative data, is there a role for us in there?
DR. HAYWOOD: I’d be curious --
MS. MCCALL: Should we have a role, we, NCVHS, have a role in either recommending or having a point of view on metrics that would be used as a part of P for P programs that are based on administrative data?
DR. HAYWOOD: I don’t know, I would just say without, again I don’t have any informed opinion about whether or not you should have a role but I would say if you choose to have a role, so don’t take this negatively, if you choose to have a role you do it not being alone, I mean meaning that there’s other people in that space because earlier this morning we started out talking about, kind of Justine, where the gaps were, if you go down that route which may be appropriate, so I’m not saying it’s not appropriate, there are other people in the space so then you would also have the issue of what you’re bringing in comparison to people in that space and all the other people that are coming out with principles and guidelines and all that activity.
MR. HUNGATE: Let me follow-up on that with a related question/observation. Your comments, Don, about the Northern New England model lead me to believe that that’s a knowledge manager that fits over in the improvement model of Brent’s. Now suppose that we took that information system that they had and --
DR. DETMER: Well, it’s really the model.
MR. HUNGATE: The model. Well, I’m trying to think at the specifics of their model because there are pay for performance things that are pressures that they have to accommodate and I wonder what we’d learn by looking at the impact of the various pay for performance models of that organization. To what extent is it synergistic, is it the problem that we’re talking about. Now is that a reasonable thing to consider?
DR. DETMER: It’s a great idea, I think it’s a terrific idea.
DR. CARR: I think tying together what we’ve heard today, I mean what we have in most scenarios is just a summary of data, either from administrative or submitted data, and what NNE has is a reactive/interactive model so I think that’s what we’re saying, they’re unique in that they take the data, analyze it, refocus, test and bring back, a little bit of what Carol was saying too about take a hypothesis and then test it out and I think to what Brent was saying this morning, then you know. I mean because having sat through the meetings like a lot of times you hear about the things that didn’t pan out and they’re boring and a lot of effort went into them and then you find out it wasn’t worth it but it’s very valuable because then you stop working on something that isn’t helping you. And similarly sometimes you look at something that’s looking at you right in the face and you didn’t realize it because you never sat down to look at it. So I think really what we’re saying by that model is you start with whatever your dataset is but that you interact with it and make it better on a regular basis.
MR. HUNGATE: Marjorie?
MS. GREENBERG: I’m not sure I can pull this all together but I think Carol is putting before the group kind of the issue that, I mean this committee has been, not just this workgroup but walking around for years and maybe if we’d heeded Steve’s junction and done something ten years ago it wouldn’t have been too late but the question is is it now, and that is this whole, is there really anything that could be done in a reasonable period of time to enhance administrative data that really could make a difference in the pay for performance, quality environment.
We know that for, and is this, if so is this an area that the committee should pursue in any way, or should we just put this to rest now. Because I mean my little role in the beginning was to describe this report which basically, or the series of reports, pulled together recommendations over a period of many number of years on how administrative data might be enhanced to improve quality, and then all of the cons for doing it, the pros and the cons, etc.
I mean basically there were four things, the committee has recommended one, I mean from a real content point of view one of them which is this issue about secondary diagnoses which California and New York have been collecting for a long time. And I think Brent mentioned, I mean he didn’t mention that but he said something about if we just could tell whether this was something the patient had or whether it’s a complication that would be hugely helpful. So that’s been recommended now, I mean we’ll see what happens, if it gets implemented and how it gets implemented, but there are people working on not only enabling that but trying to come up with definitions and to make it as useful as data as possible.
So there are really three other areas, lab data which we’ve been talking about, vital signs which nobody has mentioned here which was one of the candidate recommendations but I think had considerably less support then lab tests in the two hearings that we held last year and partly because people kind of didn’t really know what to do with it, even lab data, though there was a lot of support for the importance of lab data it was more targeted lab data as Don I think suggested, not just all the values, what are you going to do with that, and functional status. Those are the three sort of areas that come out from the report that no recommendations have been made on an that consistently we receive input would be useful or are important for performance assessment.
So the issue, one of the issues is, and this is not in lieu of promoting the infrastructure, rounding out the vision, all of that, I mean nobody is saying it’s in lieu of that, clearly that’s where everyone wants to be, though I’m hearing from John that even though he’s there, I mean in a really impressive way, they really don’t know exactly what to do with it from a point of view of quality assessment. But in any event, that was maybe false modesty, that would be our, that should be our worst problem that everyone is there because I think most people aren’t.
So the question is really you want to get where Massachusetts is but in addition is there something that could be done, I mean, or that should be promoted, I mean we heard from, we hear from purchasers, we have a purchaser, a large purchaser here but we heard from other purchasers and employers, they want something. And if somebody doesn’t, they’re going to ask for stuff that maybe they don’t have a clue what to do with but they are the purchaser so it’s possible that they’ll, the people will have to produce it, I don’t know, and so is there any kind of light that can be shed here or can this be focused in a way, the best outcome it would seem that if you were going to try to enhance administrative data is that you would do it exactly as Bill suggested, that it would be worth your while to invest in these information system to get it because it’s just too much of a pain, too burdensome and too much of a pain in the neck to try to get it any other way and yet you’re under the gun to produce it.
So it would, the incentives would go in the right direction rather then now going in the wrong direction, or do we just kind of ignore what we heard from the purchasers and saying yes, they want stuff now but sorry, until we get can get this infrastructure in that is being described for clinical records it just isn’t going to happen. I mean I really think it behooves this committee, the workgroup and then the committee, to decide whether it’s going to try to pursue this or just put it to rest, say we had these candidate recommendations but there’s not consensus and there’s disconnect that we found between purchasers and providers and we’re just not going to pursue that anymore. Or say yeah, we want to do something and we want to do something smart and targeted and that puts the incentives in the right place.
But I think just kind of not addressing it is problematic because nobody else is going to. Well, I mean they will, but nobody else is going to do it in this broader kind of conceptual way that this committee has an opportunity to do.
MR. HUNGATE: Steve, response?
DR. JENCKS: Well, I wouldn’t guarantee that that was true, I wouldn’t guarantee it wasn’t true either, but I think you need, one of the things you do when you do a strategic plan is you try to figure out what your role is vis-à-vis the other actors in the process, NCQA and so on --
MS. GREENBERG: I’m not suggesting we do it by ourselves, we have to be with other partners obviously.
DR. JENCKS: But one of my questions would be is that a model that’s likely to work, is the nature of a FACA compliant middie(?) really adapted to doing partnerships with other types of entities?
DR. CARR: Stop giggling, Don.
DR. JENCKS: He’s been there.
DR. COHN: He’s been here.
DR. JENCKS: So I’m not trying to say one answer or the other but what I am saying is for example that for a lot of the measures and issues around measuring for the right now there are things like the Alliance for Ambulatory Quality and the Hospital Quality Alliance and the National Quality Forum and we could go on, but what is it particularly, this is something that Simon asked very forcefully earlier and I think you’ve got to come back to it, what is it that this group can do that others can’t do.
MR. HUNGATE: That’s right, that’s our question. Stan?
DR. EDINGER: I was just basically, I used to work where Trent did but he probably can’t say it now, but waiting for something to happen is usually not an impediment for something happening. Usually when the Congress or the administration wants to have something implemented whether or not you have sufficient data on which to act is not necessary going to impede you from being told that you must have something in place. So I think they are not going to wait for us or another group to do something, I think Trent and other people, whether it’s Trent or the Blues, whoever is doing it, because of the current problems with budgets and the Medicare program, Medicaid, I think something will have to be done and I think it will be done whether or not we can agree to what has to be done, it will just be told that somebody will have to do it and two years from now you will come up and have a system in place and hooray, I mean having been there that’s exactly how you usually get told sometimes. Say well we’re not ready, we don’t have studies, well that’s nice but do something.
DR. DETMER: I totally hear what Stan’s saying, I hear what you’re saying, I think on the other hand we do have all this group of people really staying up late working on the quality side, and we’ve got IT people staying up late working on the IT side, we don’t have huge numbers of people staying up late worrying about how do we hook quality to IT. And it strikes me that if there’s anything this group could really focus on as part of not losing sight of where you really want to go it’s lay out that matrix and start some hearings on how do we in fact incentivize some of those kinds of things, deal with some of this so that we start moving, because things will happen anyway, but on the other hand you’re in a position, you’re going to be here, been here 50 some years, so the point is you got an opportunity if you have the vision and you start worrying that issue and you drip, drip, drip, drip, drip away at it I think you’ll make some difference because you do have, a lot of people have to listen to you, they do.
MR. HUNGATE: Help me a little bit in terms of your separation of the two tasks of infrastructure and knowledge, where’s our role
DR. DETMER: Well, exactly, and I think, from my perspectives I think you have Brailer’s office, that is and a Secretary who clearly get the NHII thing. And that’s plenty to do, that’s a lot to do. But we also are talking about how do we put the info on the way and I think you’re the info folks and I think what you’re really trying to say, well where are we trying to go with this train and worry about how do we put that knowledge management piece on it and then how do we reform our system so that we do in fact line up some of these incentives so that they really do move us to where we know we want to be.
And I don’t know, I’m still speaking in abstractions but at least I think you hear kind of what I’m saying and it strikes me that that is something that the NCVHS, it could guard the vision as I say and then worry about some of these little other issues, making sure they get tracked. But part of that guarding that vision is building out that intersect, and as I say the number of people who really, Brent James is one of them, I think Jerry O’Connor in the Northern New England, and David Bates, you’re talking about probably ten souls in this country and 30 worldwide maybe that really have, that live and breath this sort of stuff and actually are nimble enough to think and talk about it from sort of different places. But it’s still an evolving thing, I mean no one has totally revealed wisdom on this, I mean that’s the nature of it. That’s how I see it, John, how do you see it?
DR. HALAMKA: This is really, it’s a critical juncture in history because we’re starting to see these standards being put in place, the tracks are being laid for interoperability, and they’re being laid driven purely by things like clinical transactions and not necessarily by quality so unless we have some guidance of what we need to put in, what we at least need to plan for, we may get a little far down the road and then have to go back and reengineer for quality, I’d rather --
DR. DETMER: If you’re thinking through it in a matrix sort of way, what’s most likely to help safety on this, what’s most likely to help efficient, the assessment of timeliness, the assessment of basic centeredness, those sorts of things, I think that kind of working conscious to try to lay some of this out and try to get some sensible markers that could sort of move you in that direction and hear from people who are doing some of that, AHRQ funds some of those things, try to see how that fits into policy as well as payment. It strikes me over time that really is helping but it’s not necessarily answering Marjorie’s question.
MS. MCCALL: Well, I’m going to go back and try to take a stab at Marjorie’s question, I’m going to actually take a point of view and I’m going to submit that we shouldn’t spend any time on enhancing administrative data and that, let me tell you why, and I’m from a payer and for all of those from Humana who are listening to me right now, I don’t know how they feel about that, but here’s why. I think that pay for performance is going to move forward, I think we’re going to learn some things from it, but I think it will move forward with what it has. I think to do anything in addition to enhance that I think is more pain then gain, the pain to gain ratio is just not where it would need to be. And not only that, we’ve robbed Peter to pay Paul and Peter, or in this case John, in all of that work needs us more. And so I think what we need to do is to try to speed up our evolution, that that is where our greatest value comes.
Said another way is that sometimes spending time over on another area actually kind of somehow strangely relieves the pressure and we need the pressure to be to come over to where we’re building. And so I just lay that out there as something for people to react to.
DR. HALAMKA: So I’ll totally agree with you, the IT agenda right now is such that my demand will always exceed my supply no matter what my budget is, and that if you give me a project enhancing administrative data it just means it’s going to delay all my clinical data interchange activities.
DR. EDINGER: Yeah, and I think if you argue the issue about lab data, I think the issue is do you need the lab data, but if you go down the point of how do you add it to the administrative dataset rather then if you need it in the first place, you’re probably going to spend a lot of time with a discussion of which codes, which we’ve already had, how many codes, which codes, you have numeric data, you just have percentage data, and then basically in an electronic system you basically have the information, you wouldn’t have the whole discussion but I think he’s talking about many years to do it on an administrative dataset.
MS. GREENBERG: Well, all right, I think, I just had an ah ha experience but I may be the only ah ha, it may be only my ah ha, but I would say let’s put that to rest as the way it’s phrased as enhancing administrative data because although I’m not convinced that we’re going to have as widespread of clinical data, I don’t see the money going into it that I think would need to be going into it to get any reasonable time of this happening. At the same time I’d put that to rest as saying enhance administrative data because I realize all the problems with it and I’m tired of hitting my head against the wall.
But I think the other side of looking at that though, a different way of looking at it rather then enhancing administrative data is all these different groups out there working on this and I don’t, I think we’ve agreed NCHS should not be looking at, or NCVHS should not be looking at measures, I don’t think that’s our goal, and Bill said well, if you, you shouldn’t be talking about the data then until you know what the measures are, well there’s something to be said for that but then we really can’t do anything to some degree except those measures that are agreed on.
But forget about whether it’s going to come from administrative or clinical records or whatever but to some, to address this vision of where we need to be and do what Ernie suggested of bringing in the key stakeholders, I mean we hear it from providers and payers but there are a lot of other people on that grid and finding out what it is that in this area, whether it’s to define value or quality or a combination of the two, people feel they really need and then the vision is that this will get produced in some way at some level, with some degree of privacy, and we really will be able to link records so you know that this person who got this here is the same person who got that there. And the vision doesn’t have to mention administrative data at all but at the end of the day people are going to have to respond to that vision and so they’ve got two choices.
I would say the administrative data already have the capacity to collect some of this because of the nature of the UB-04 and the claims attachment standards even though those haven’t been promulgated yet but they’re, I mean there are implementation guides for them, etc., so people can produce this information, whatever it is or whatever the vision is that we’ve sort of identified what people need and how that differs across these stakeholder groups, and then, the mechanisms are out there, one mechanism is to try to produce those data, work with the groups that have the measures, the data to produce those measures, from whatever cobbling together of administrative data is possible, enhancing, and this would not be the thing you send in as the claim because we don’t have that option to everyone send in a different claim but it’s what we’ve talked about before as a quality transaction.
But in any event, some type of reporting, which you could do through cobbling together administrative datasets, the claims attachment standard, the taking advantage of all the capacity on the UB-04 and maybe the 1500, whatever. Or you could do it through really accelerating your electronic health record systems. And actually that’s the American way anyway, we don’t tell people you have to do it this way, I mean you haven’t told them you have to use this system, there’s seven different systems. And at the end of the day people have to produce the information for these measures.
And one option frankly if they just can’t get these clinical data systems in place in a timely way to do it would be to cobble together these administrative systems which frankly could have the capacity but you don’t want to lay it on everybody. Or the option is to really speed up the clinical records. And probably good sense will tell people and will tell the decision makers they should maybe do the latter, and to me I’ve just decided that looking at administrative data, I now believe that the administrative mechanisms actually have the capacity to collect lab values, to collect functional status, and collect some vital signs too if you wanted it. Forget about talking about enhancing them, the capacity is there, it’s like putting in placeholders, the committee has talked about this for years, the placeholders frankly are there, are going to be there as soon as the UB-04 goes in and with the 837.
But I guess it’s a little different then what you’re saying, it’s not saying just forget about administrative data, recognize that administrative data could possibly for some people enable them to report information that could be used on the measures where there’s agreement but that’s not really the ideal way to do it. And that seems to me would be part of the vision, sort of conceptualizing that, and so I can now, I mean I’m happy because I think the way we were going was a dead end because you were never going to get agreement on how to, what to put, what to require in the administrative data. But it is one mechanism and I think through some of our work and the work of some of these other groups we’ve created the enablers but the rational person may say that’s not the way I want to do it. Does that make sense of the way to resolve --
DR. CARR: Marjorie, your fortitude is inspiring, I’m going to say that, and I think you’re right but the one thing as you were speaking about that reminds me of when we were at the hearings and I think what folks struggled with was which blood pressure, which temperature, which glucose and so on, and I think thinking about John’s architecture where he is looking for something manageable that will be information, maybe our efforts need to be how you configure the question so is the patient normotensive, uglycemic(?), normal cholesterol, so that it’s a yes/no that then can be downloaded and you can say, or not evaluated, and so whether there were 15 blood pressures taken in an hour because the patient was in pain in the first 15 minutes and settled down is not, that’s, you wouldn’t know how to do that but you leverage the clinician’s judgment by saying is the patient normotensive or not evaluated or hypertensive. And so I think not losing sight of the importance certainly of functional status and of the elements of vital signs and labs that we want to know, we would want to translate it into the language of the electronic health record and begin thinking about how do you ask those questions in a way that you get answers back that is usable and adds value.
The second thing that I was thinking about too in terms of again leverage, and I’m a big proponent of the IOM dimensions, but is to begin to say will our electronic health records have times in them so we know time to antibiotic, time door to balloon, whatever all those things are. And so maybe stepping back at a conceptual level and say that we needed or how we get it and so on so that we build the steps to get to, we can answer a question of was something timely or --
DR. HALAMKA: Just a comment, I do not capture vital signs electronically in non-monitored patients --
MS. GREENBERG: In what?
DR. HALAMKA: In patients who are on the ward I don’t capture it so I can’t give you that information. But I could certainly do it for telemetry patients, ICU patients, those kinds of folks, functional status I do capture. I definitely would argue please don’t require me to give you vital signs on your average ward patient, that would be a burden --
MS. GREENBERG: How do you capture functional status?
DR. HALAMKA: We have a multidisciplinary discharge product where a case manager, a nurse, and a doctor all contribute to the preparation of what I’ll care the transfer of care documentation, whether it’s going home, going to a post acute care facility where they have to document their ability to eat, their ability to do the activities of daily living, ability to ambulate and all that stuff --
MS. GREENBERG: You code it?
DR. HALAMKA: It’s all vocabulary controlled but its our own vocabulary.
MS. GREENBERG: Can I just make one more statement? Then I’ll stop, it refers to my fortitude, the only thing I have more fortitude on is the fervent wish that this country would recognize that the rest of the country is using ICD-10, but that’s, it’s really the same kind of solution, you could go to the clinical modification of ICD-10 two ways, the way that we went to the clinical modification of ICD-9 or by implementing SNOMED CT and putting the resources in a rural based mapping, which won’t be 100 percent but could be 80, 90 percent and then the rest would be a lot easier.
So again, I think it sort of changes the view but whereas clinicians may say we don’t need, and that’s why we need to hear from these different groups, we don’t need 10 CM, 9 CM is good enough because we’re not using it anyway, we’re using SNOMED or we’re using labs, whatever, you’re going to hear from the statisticians and the people who are dealing with, and the researchers, etc., but they really have a problem using a 30 year old classification system. But again, it’s sort of, the incentives could be towards moving people towards using controlled vocabulary if it was then, then it would make it considerably more painless to then move to the current classification system. So I think it’s moving these incentives in a way that yeah, you can do it the old way but it’s going to be a lot more costly, do it a way that you really get value out of. End of that.
MS. CRONIN: I just wanted to make a comment after spending the last year and a half working in the Secretary’s office on HIT and now coming over to CMS and seeing all the challenges and thinking through a lot of the operational aspects of P for P. One thing I think has really not been tackled is trying to come up with a strategic vision for clinical decision support and while it was talked a lot about in the Office of the national Coordinator and in Don’s association and a lot of other circles elsewhere no one has really tried to lay out what is the infrastructure we need, how do we make sure clinical practice guidelines get into a computable format, that it’s housed in a trusted place somewhere where it could be easily downloaded, that you have some type of way of automatically updating whatever kind of clearinghouse or infrastructure is created so that whatever is in the peer reviewed literature can then be put into this type of infrastructure. There needs to be a lot of thought given to how applications can not only be certified or developed in such a way that it will make it very user friendly to clinicians to use guidelines or whatever kind of logic is built into these tools. But I think there also needs to be incentives for these clinicians to actually use these features, otherwise there could be alert fatigue or a lot of other issues that would cause them not to even consider using this evidence that’s readily available at their fingertips.
So I think that really what Don said earlier is right on from my perspective for what it’s worth, the intersection between IT and quality is really what needs a lot of attention now and is what is probably really under resources. And if you can really focus on trying to figure out how do you realize that vision of having interoperable electronic health records that have robust clinical decision support and what infrastructure is necessary around that kind of system then it probably be value added and also would not be something that we need in the next six months whereas a lot of the other discussion I think earlier today was focused on issues that are being addressed by a lot of outside parties right now.
DR. DETMER: Could I trail on that? That was wonderful, nice to see you. Actually I think, when I was saying we’re in these early days it was structure that we will ultimately have, we have obviously came out of the genome project, the Gen Bank, and now we’re getting the molecule bank, small molecule bank, and we’re trying to get funding for that. But the point is we need also a global clinical trials bank and there’s some debate can we haul pharma into that sense of enlightenment, Eli Lilley has really certainly taken the lead right now toward that but that’s not another huge knowledge base to help us know what we know and kind of what we think we don’t know at the moment, the clinical guidelines bank if you will needs to be developed out. There’s a public health guidelines bank that needs to be built out. There’s a patient side of all of this knowledge base that needs to be built out, there’s a whole set of these knowledge structures actually that really do need to be sort of built out, that’s a piece of this vision writ large and I think that that is a set that I think NCVHS could really carry this next way because in a way NHII group succeeded, they have, we got that one, now let’s go to this next chapter, let’s capture that one and let’s drive it along because that’s ultimately where the lives are ultimately, the tears that don’t need to be are shed, that’s really where they’re going to be.
DR. HALAMKA: If it will help the committee I will donate because I have them here all the decision support guidelines, order sets and protocols used throughout the Harvard system, they’re yours.
MS. CRONIN: I just would like to say, I mean some of your colleagues are really sort of the leaders in this area, Blackford Middleton, Jonathon Tish(?), David Bates, Don Detmer, John Halamka’s colleagues would have an awful lot to share with this committee if you chose to go in that direction.
MR. HUNGATE: Steve did you have a comment and then we’re going to take a quick break.
DR. JENCKS: I am not going to stand between this committee and a break.
MR. HUNGATE: You may have the floor upon our return.
[Brief break.]
MR. HUNGATE: I’m very open to comment but my suspicion is that we’re all pretty tired, I know I am, I’ve listened hard and thought hard, and I have the feeling that the people that join us tomorrow would benefit by the wrap-up discussion, and so what I’d like to do is basically wrap-up today tomorrow morning. But that would leave you out of an opportunity to add to that --
DR. COHN: Can I make a comment then or two before --
MR. HUNGATE: I would appreciate that before we depart.
DR. COHN: Okay, well good. Actually we scheduled a meeting to end at 5:45 and you leave early, this is interesting --
MR. HUNGATE: I thought I had more stamina then I do.
DR. COHN: Expect to see some of this on the full committee now actually, we’ll have to do this at the full committee meeting.
I think it’s been a fascinating day and as I said the reason I’m commenting, others should free to comment also because I had it say it, my general view is you tend to forget things overnight but sometimes you get a little perspective on them.
We talked earlier in the morning and I was sort of commenting about, making sure that we had actionable recommendations and certainly I want to make sure everybody realizes that actionable recommendations does not mean that one has to have near term vision, that it’s appropriate to have a longer term vision and then the question gets to be well what step by step by step can we take that moves people in that direction and that really becomes part of the question. So I don’t think that there’s anything I heard today that was really off of the table.
Also I would comment and I can’t, now once again I think you all are going to need to look at each other and talk about your bandwidth and ability to work on things as well as priorities but other subcommittees have more then one issue, or more then one thing that they’re dealing with at a time so the fact there may be a number of things that you come out of these meetings with as important doesn’t mean that you have to come up with the one. Obviously you don’t want five, you don’t want ten, you probably don’t want more then three. And you may want to decide that you want to have three and talk about them to the executive subcommittee to see about senses of priorities or whatever, but you’re looking at one, two, or three but as I say I don’t think you’re by definition limited to one. But once again you have to reflect on your support, your bandwidth and what it is you think you can do.
I do have to say I can think of three things that I heard today that I thought were sort of fascinating and I’m hoping that you will opine on them and I’m sure that there’s going to be more. One of them of course is this, as Don Detmer was putting his hands together talking about the interface between IT and quality, actually I think John also made that point though I don’t know if you used your hands on that one --
DR. HALAMKA: Fingertips.
DR. COHN: Fingertips, okay. And I think that that’s a very important issue especially since I think I’m hearing general consensus that we don’t want to go back and do the quality information report again, which I would certainly second. So I think that’s an important area although exactly what that looks like, feels like, I mean it’s a long term thing where the question is what steps does one take to move us into that direction and that’s obviously a conversation for tomorrow.
The second piece which I was I wouldn’t say fascinated about but I need to thank the representative from the NCQA was to sort of bring up the issue of well, is it really just quality or is it really value that we need to be looking at as a NCVHS and as a workgroup. And I do know that that conversation got sort of lost as the afternoon went on and then we just went back to the quality moniker but I think that the, in a world where there’s so many people who are uninsured because they can’t afford insurance, and for other reasons, where generally you ask the person on the street what’s the biggest issue with health care and it isn’t quality, it’s cost, followed by other things and then quality. I think that value may be something that we want to explore more and is maybe something that we could add value to, something that we could provide important input on.
The third one is I think a request from Trent Haywood I just want to remind everyone about and I guess I just tend to always listen especially hard to CMS just because, I mean they’re not here but they’re the biggest payer, I think they need the most help. I’d also like to be able to take advantage of the program when I get to be of the right age and so I want to make sure that they do well and remain above the water. I know Bill thinks the same way but Trent was sort of asking in terms of a relatively specific piece of work that was looking at what the states were doing in terms of I think quality measures, I think we’d have to get a little better definition from him of what it was but I think he was getting feedback was is that well, you’ve got NCQA and you’ve got JCAHO and you’ve got things that the federal government may want to do but you’ve also got these various state requirements for tracking and measurement and all of that stuff and we’re being pulled in 12 different ways and there’s unreasonable burden there.
I don’t know if that’s true or not in this day and age but that was at least how I was interpreting that, Bill, and we probably want to get greater definition from you.
DR. SCANLON: We could talk to him because I actually interpreted it differently.
DR. COHN: Oh, did you, please.
DR. SCANLON: There’s a parallel sort of set of quality assurance activities that’s going on at the state level that’s not relying as much on information but is relying on inspection, and there’s a question of the disconnect between the two. I don’t know whether maybe Steve sort of could react to that
DR. JENCKS: Not without being caught up on what was said before I walked in the door.
DR. COHN: What I was commenting on was that Trent when he was here today and also in conversations with him had sort of indicated that he thought it was something we could do that would be useful would be to look at the states and what they’re doing around, and I get a little vague at this point, I don’t know if it’s performance measurement or pay for performance or just routine sort of requirements that are somehow different then where is happening nationally or being thought of by CMS. And I think one of the values of course and why I always like to have customers is you can sometimes ask then a second time well what exactly did you mean by that.
But those as I said from my view were I think the three things that I was hearing that sort of struck responsive chords, knowing that what you’re hearing is I can’t even faithfully reproduce what Trent was specifically asking. Steven, can you comment on this one?
DR. JENCKS: We have a considerable interest in understanding the kind of stuff that is going on at the state level, I don’t know whether you’re especially well equipped to get that information but the states really are in some ways are a potential laboratory here and that could be very helpful, I mean I think that what’s going on in Massachusetts is an example of one kind of laboratory and somebody said well, all right, how about the other places where a RIO might emerge. And the answer is maybe somebody up at Brailer’s shop knows the answer to this but I sure don’t, and understanding what the lessons are from those efforts might be very important to producing a smart report. So we’d be very interested I think in information about what’s going on there. Surveying total burden of that, is this, I mean it’s something we’d be quite interested in but is this something that you would be particularly good at doing?
MS. GREENBERG: Just to find out what states are doing and quality measurements? I think the National Association of Health Data Organizations is probably better equipped to do that, and they are certainly attempting to track this somewhat but, not a but, they probably would need additional resources to do it comprehensively. They work with AHRQ on the HCUP data which I think at this point most of what states are doing are around hospital data. But they’re building emergency department systems, they’re building to some degree ambulatory care systems, the states are. So that could be, I mean that could be a recommendation that the committee could make that this should be done and even a suggestion to look into working with NADO(?) and other appropriate groups rather then doing it ourselves.
DR. JANES: So you’re saying a recommendation that, so Marjorie you’re saying we would issue a recommendation that NADO would in fact be the appropriate organization to do the survey and then bring it back, because I don’t think that NADO, if it would just be some sort of discharging or responsibility to NADO in saying pursue this issue, see what you come up with and make a recommendation, I don’t think with all respect to NADO and Denise Love, I don’t think a recommendation from NADO would carry the same weight, at least with the administration, places like CMS, that a recommendation from this committee, so I guess I would say sort of subcontracting to them to collect the data. I agree with you, they’re the people who have this in their fingertips and then bringing it back if the workgroup decided that was an issue that they wanted to address.
MS. GREENBERG: Well, or it should involve AHRQ, too, I mean I think we should talk to people from AHRQ because I think it is primarily around hospital measures at this point and this is certainly an area they’ve been working on, quality measures for HCUP data. Do you know what HCUP is? Hospital Cost and Utilization Project, where they get hospital discharge data from the majority of states and then try to kind of normalize it.
MS. MCCALL: Never heard it called that before. Back to Simon your comment that you wanted to pay particular attention to the voice that CMS was bringing here, I’ll confess that it wasn’t clear to me precisely what it was that Trent was referring to, I think the first thing to do is to go back and get him to clarify it, whether it’s around measurement burden or inspection and kind of how state and other bodies that are going to be influencers here actually come together in some sort of form, kind of managing the silos as you would say. So I think we should clarify that first and then come back and see --
MS. GREENBERG: Oh, I agree, it’s premature to make a recommendation but I’m just suggesting that, you asked if it should be done by the committee and I think there might be a better group to do it.
MR. HUNGATE: And your reaction to the quality IT intersection, is that the vision thing in your mind? Is that the same is this idea of what is our ten year goal?
DR. COHN: Same as what?
MR. HUNGATE: As the idea of what’s the ten year objective, where are we trying to get to? There was a lot of enthusiasm conveyed for that by Steve --
DR. COHN: I guess I’d reflect the question back on you, I think it’s certainly a longer term set of issues, I don’t know that it’s a vision per se, that I guess you guys all have to talk about, I mean I don’t know whether value is the vision, it’s hard to know where the vision in this exactly lies.
MR. HUNGATE: I think it was highlighted by Don’s talk about the knowledge base management, that we’ve talked a lot about the infrastructure and there’s work on the infrastructure but there’s not attention to what’s the content and evolution of the knowledge base and where is it going and what does it mean in terms of coherent measurement systems. We have a lot of incoherence in the way performance measurement is being done and it seems to me the value would be in describing a coherent performance measurement system, which might need articulation.
MS. MCCALL: I’ve got a question for you, which is as I’ve sat through today and participated in today and done a lot of reading in advance, and I look at the charge for this group that you had put out there and I wish I actually had the charge for the NHII group because there were some specific words in there that struck me. So my question to you is this, does the charge for this group, does it change? Does its placement change? You, Marjorie, had talked earlier about its placement within the Population Subcommittee and because of what is now becoming a hand sign, this, and this intersection between IT and quality and they have a system without a purpose and a system without specific metrics and goals that map to those, it’s just another system, we won’t necessary get what we want. Does the alignment here change? And it’s as much food for thought as anything else.
MS. GREENBERG: That reminds me of discussions we’ve had about which subcommittee represents the big picture, I mean we’ve had these discussions I think in the executive subcommittee, is it the Population Subcommittee, is it the NHII Workgroup, is it really the Quality, which one is the organizing framework, well, it’s all of the above I guess. I don’t know if given now, I mean I think at this point, I don’t know, you’re the chair, I won’t even comment.
DR. COHN: What were you going to tell us?
MS. GREENBERG: Well, I was going to say at this point I don’t see any need to make a change, I think hopefully is Don going to be here tomorrow?
MR. HUNGATE: Yes he is.
MS. GREENBERG: Okay, I heard some things coming out today that are definitely more sort of population based then health care based but of course the population world has to benefit from this NHII too and has to be a player in it so it’s the same, to some degree the same systems or add-ons to the same systems. So I mean I think at the time that this workgroup kind of comes up with its charge, or comes up with what it feels it should be doing, its work plan for the next 18 months or whatever, it may then be obvious whether it makes more sense to be a workgroup under the executive subcommittee or to continue to be a workgroup under the Populations Subcommittee and since it’s mostly the same people and Don is on both I think, maybe it doesn’t matter that much.
MR. HUNGATE: My sense is is that the critical choices are between the agendas for all of the committees and how well they mesh in an NCVHS being effective, and we don’t know that yet. And so I think it may be a moot question but it may not be, I don’t know. I’ve often wondered about how these charges came about and so we decided, I decided for the committee that we’d kind of say what we thought it ought to be and we’d see if anybody else agreed, that that was kind of the vetting process through the executive subcommittee and the other heads and say okay --
MS. MCCALL: Does anybody happen to have the charge for NHII?
MR. HUNGATE: I looked and I don’t have it with me.
DR. COHN: Actually I don’t, I have the charge for Quality somewhere or other.
What can I say about this one?
[Multiple speakers.]
DR. COHN: Carol, I guess what I would say is that my general view is that form follows function and so the first step is to figure out what in the heck we’re doing and early I had all sorts of plans about reorganizing this that or whatever and I sort of came to the conclusion that there was no need to make quick decisions on all of this stuff but some of this may very well evolve and become obvious.
The other piece and I wanted to sort of remind you when I talk about goals and work plans, one of the things you’re all going to have to sort of think about tomorrow, there are things that you may come to that are, that are clearly what the workgroup needs to be doing, there are other things that are important enough that we need to be thinking about whether or not aspects of it need to be sort of more generalized and maybe a goal for the whole committee and then the question is is how things fall into it.
And I think that’s all to the good and I don’t know which may be which or how all this may play out but certainly at the end of the day I think Bob you were sort of commenting that part of what I see the executive subcommittee doing is sort of looking at all these things, trying to figure out synergies, trying to figure out, trying to provide some guidance on priorities, trying to sort of reflect a little bit so that we all do a better job as a result, most everyone is involved with more then one workgroup or subcommittee and we want to just make sure that everything sort of leverages off each other without a lot of redundancy.
MS. MCCALL: An example of one that we actually talked a little bit about today that probably don’t belong here, but it’s how do we create another Massachusetts experience, how can that be done, what do we need to do within NCVHS, what can we do to help prepare the soil for that, and it’s probably not within this workgroup, it may be within another --
DR. COHN: John, do you have any twins?
DR. HALAMKA: Well, there are 168 hours in a week. And to that point there are many, many lessons learned that I’m willing to share with the committee and a whole history of what we’ve done and how its come together. I actually just wrote a paper for JAMIA last week on that topic and we’ll hope it’s accepted.
MR. HUNGATE: I wanted to make one and then I’ll give it to you, Marjorie, my experience in getting the quality report approved has left a lasting impression on my psyche and I don’t want to repeat the difficult arduous approval process. And so I think that a lot of what this workgroup is going to do is going to have to be at full committee as well, in other words we could not come forth with any sort of a vision without having had an awful lot of exposure at full committee to the content that we’re dealing with. At least that’s my intuition of this, that it’s just not going to be possible to do it in that way.
DR. COHN: If I can make a comment, I think you’re right though, I do think that you should maybe take a different lesson out of your recent experience, and that’s that when reports drag on to the point where the tenures of the members who were involved in even hearing the stuff has gone there’s no way you can win, and that’s one of the reasons why I sort of pushed for timely reports, I mean this is sort of think really the bigger issue, it isn’t that every report is a painful experience but letters are a little easier to get through then full reports --
MR. HUNGATE: What I’m trying to articulate there is that I think our work plan must embody a mixture of hearings and testimony before the full committee, invited speakers that deal with the content that we’re bumping into, that we think is knowledge that ought to be at the full committee in order for the full committee to judge a work product later on. That’s what I’m trying --
DR. COHN: And actually I agree, I mean I think we all talk about the idea of bringing the full committee along on subcommittee and workgroup activities so that when things come forward it isn’t like as in e-prescribing like huh, I don’t know what that is, standards and e-prescribing, similar issues, you try to carefully script.
MR. HUNGATE: I think Brent’s presentation this morning would be a good subject for the full committee. That’s a personal bias, to me he’s dealing at a very fundamental level --
DR. COHN: Let’s see what the goals and the work plan of the workgroup are first and then from there you sort of figure out what sort of information needs to be shared that helps make everyone understand because I mean I think, to my view, I mean I really like Brent James and I really like his presentations, depending on where you go he may not be the most relevant presenter on all of this. So I would just hold the thought in abeyance for a moment.
MS. GREENBERG: I was back a chapter when Carol said how can we encourage what happened in Massachusetts, one thing that I hear repeatedly and heard last week certainly at the eHealth Initiative, and you always hear it, solutions are basically, or the problems are basically not technical, and one of the biggest issues is kind of building that web of trust in a community or in an area so I was going to say roll back 25 years and put an Elliott Stone, my late friend and colleague Elliott Stone in every areas. But really when HIPAA came out and he came to the national committee and he said we really, you really should be encouraging kind of regional solutions to HIPAA and everyone said thank you very much, or supporting regional groups that will implement HIPAA, and the HIPAA administrative transactions, and that’s what you did so whereas you have a return on investment all of these other places are telling us they have a thousand companion guides and they have no return on investment, I mean that’s telling.
But okay, so it’s too late to roll back the clock 25 years but I do think we should be thinking in terms of this model, there are other models too I’m sure but I think the models that have been most successful are regional models, models where you have been able to build that web of trust, and it’s not the technology solutions really that are escaping people, it’s the community is not ready to do this, I mean they’ve been working in Indianapolis for what, I don’t know how many years but they also have built that web of trust. Well, some communities may be more, maybe the Californians would never trust each other, I don’t know, but I do think we shouldn’t just say isn’t that nice and it’s too bad we don’t have this elsewhere and think in terms of what could encourage those kind of regional solutions, when now we’re talking about regional health information organizations.
But I think we shouldn’t just be, it shouldn’t just be the fad because I think there is some evidence that building that kind of web of trust in a community and exchange and policies and everything can have a real payoff.
MS. MCCALL: Couple things and the fact that it’s local, we have this piece of information here so I find myself kind of focusing on that first sentence, but back to what it takes to do another Massachusetts, somebody said earlier and I think it was Don, we were talking about the fact that health care is local, think about these systems. Think about the value, let’s say you’re the only person on the earth that owns a fax machine, how valuable is it, not very, the value of your system goes up in proportion to the number of people that have a fax machine and what these systems do, I mean it’s almost as though they’re defined by their ability to interoperate, I mean health care, it is a system and it is about coordination things because of the nature of the specialization that’s involved.
And so perhaps one of the greatest values in addition to some other things that were in play was the fact that you did do it, essentially in that area you did it everywhere, and so it was much easier for people to actually realize the value, as well as all the other things that went into it. And having worked for a company that I think is part of the gold standard there now, Allscripts, know what it’s like to go out into a local community and kind of try to pitch by the ones, it’s just getting to base camp is pretty tough in that climb let alone getting to the peak.
DR. HALAMKA: And there’s no question that we have a networked effect as you described which is now if you’re not connected to the network of interoperability you’re perceived as not a player.
MS. MCCALL: Second thing, thank you so much for going and getting the NHII charge, just the thing that struck me when I was reading this a couple weeks ago is that it’s described as a set of things, technologies, standards, and applications that serve a purpose so it’s not, you know when we think about NHII it’s not a system, it’s not a database, it’s not purely a technical architecture, that part of what we’ve been talking about here today is a part of that infrastructure, whether it’s knowledge, whether it’s the ability to begin to know what we know or even begin to know what we don’t know, so that we have some sort of infrastructure and architecture that can be more adaptive and more nimble. That somewhere in here is part of what we’ve been talking about today, at a very high level of course.
MS. GREENBERG: They spent a lot of time talking about this, I mean when they came up with it.
MS. MCCALL: And that’s why I promoted that question, not to be answered today.
DR. JANES: I guess two things come to mind as I listen to you Carol, one is I think what I’m hearing you say is the same issue that came up again and again, certainly John referred to it and Don referred to it, and that is for us to think about that, again the fingers, that intersection between quality and IT which I thought was really a fascinating issue and Marjorie and I had this kind of a bathroom conversation about this because I thought her sort of moment of personal epiphany was a great one when she essentially said why are we obsessing over mode of data collection, why don’t we just focus on the metric of quality and then let people work it out themselves, if they’re living in a sophisticated IT world that’s fine, if they’re living in administrative data, and I think that’s right, I was sort of sitting here and thinking along those lines as well.
The one thing that I think that that misses which I think is what Don and John were talking about is not just to free ourselves, free the quality world from that bond to administrative data but actually to rethink the paradigm as well, that there may be in fact we may have contorted ourselves for so many years in our thinking about quality by that constant issue, I mean you see it in NCQA when you want to develop a new measure, what’s the first thing they do, feasibility, can you measure this with administrative data, I mean I really think that it’s become a knee jerk for so many of us. If we tried to pull away from that and rethought about quality would we in fact find perhaps not a completely changed paradigm but one that at least moved in different directions and that’s really exciting.
And then if you sort of rethought quality and then put it out there and say we’re not going to make any statements about how you should, about the mode of data measurement, then I think that might move us in the direction that some of our speakers were urging us.
So then the only other point I wanted to make is if you buy into that vision and like I said I find that kind of an exciting idea, then it’s certainly from a very practical point of view then it also opens up the issue that a number of people, you Bob, had been talking about, and that is our relationship with other allied subcommittees, like NHII and standards to a certain extent, Standards and Security, folks that have dealt with some of these issues as well.
And I just remind the workgroup, particularly those of you who have been involved for a long time, and that is there have been times, I remember a time when Kepa was still on the committee when we used to desperately try to get people from the technical subcommittees to come in and sit with us because we were wrestling with technical issues. And we used to always run into this, what always struck me as just ridiculous but ridiculously unyielding problem that the two various subcommittees tended to be on meeting schedules which meant that the same people couldn’t sit on quality who also sat on standards and security or NHII, and I used to think are we going to really be blocked by this issue but it seemed to be really stubborn. We will perhaps have to address again some of these issues if we decide to bite off these purely cross cutting issues and try to figure out ways that we can get people to bring their expertise into other groups.
MS. GREENBERG: During those breakout sessions of the full committee that has been a problem.
MR. HUNGATE: It can be an issue, right. Okay, anybody want the last word? We’re adjourned until 8:00 tomorrow morning.
[Whereupon at 5:25 p.m. the meeting was adjourned.]