American Health Information Community
Population Health Workgroup #15
Friday, March 2, 2007
Disclaimer
The views expressed in written conference materials or publications and by speakers and moderators at DHHS-sponsored conferences do not necessarily reflect the official policies of the DHHS; nor does mention of trade names, commercial practices, or organizations imply endorsement by the U.S. Government.
>> Jennifer Macellaro
It looks like we're ready to get started. This is Jennifer. I'll first start off telling everybody it's a Federal Advisory Committee Act proceeding here today so it is in the public domain. We're being broadcast over the Internet and recorded and transcribed so if everybody can identify themselves before they speak, and speak clearly and distinctly. There will be an opportunity for public comment at the end of the meeting and it's really helpful if Workgroup members can keep their phones muted when they're not talking. I'll just go ahead and run down the list of the people on the phone and then you can introduce those in the room.
Today on the phone it looks like we have John Loonsk from the ONC. Jim Craver from the CDC. John Lumpkin from the Robert Wood Johnson Foundation. Leah Devlin from ASTHO. Scott Becker from APHL. Dave Parramore from DOD. Laura Conn from the ONC. William Brandt is in for Marty LaVenture from the Minnesota Department of Health. Lisa Rovin with the FDA. Angela Fix from ASTHO. Tom Savel from the CDC. Farzad Mostashari from the New York City Department of Health. Brian Keaton from the American College of Emergency Physicians. Mark Carroll from the Indian Health Service. Michelle Meigs from APHL. Michael Barr from the American College of Emergency Physicians. And Paula Soper from NACCHO.
>> Michael Barr:
Just one correction. This is Michael Barr from the American College of Physicians.
>> Jennifer Macellaro:
I'm sorry.
>>
We gave you a promotion.
>> Jennifer Macellaro:
Did I get everybody else?
>> Michael Barr:
We can talk about that offline.
>>
It's 3:00 in the morning.
>> Sunanda McGarvey:
Shu McGarvey is also on the line.
>> Jennifer Macellaro:
Okay. So I'll turn it over to you there, Vernette.
>> John Lumpkin:
I'd like to welcome everyone to the call. We're in the midst of a very busy time for our Workgroup. As we all remember, we got our name changed not too recently, not too long ago. We are proceeding to have a series of discussions working through our agenda that was set last fall. Starting off with case reporting, we have a document that -- was it sent out with this? I think it was, yes --that we have hammered out and I think we are ready now to send it off to AHIC. We have two other areas that we've begun to work in --
>> Chip Kahn:
Hi, Chip Kahn here.
>> John Lumpkin:
Hi, Chip. I'm just doing the opening remarks and walking through the agenda.
>> Chip Kahn:
Okay.
>> John Lumpkin:
The first area that we've already conducted a series of hearings on has been adverse events detection and reporting, and we'll, we have a rough outline of a series of recommendations, and then we're going to talk about the next area, which is response management.
So it's an important time just a reminder, we'll get to this later but I thought I would do it at the beginning because everyone, sometimes people drop off. Our next meeting is March 29, and that will be an all-day meeting in D.C, and we will be having testimony on response management and we'll be talking later on in the meeting about what that's going, what that should look like.
Chip, any comments, to open up?
>> Chip Kahn:
I guess in a moment I'm going to do -- we’ll just review the outline one more time. So I think I'm fixed.
>> John Lumpkin:
Okay. Is Steve on yet?
>> Kelly Cronin:
No, we're still waiting for him.
>> Chip Kahn:
Actually, I had lunch with him.
>> Kelly Cronin:
Oh, okay. He's in transit.
>> Chip Kahn:
So I'm sure he's at the building by now, and he may be having trouble getting up, but he was --
>> Kelly Cronin:
I'm just sending a note to see if we can get him.
>> Chip Kahn:
I'm confident he's downstairs. If -- by now.
>> Amy Helwig:
And this is Amy Helwig from AHRQ, I've joined the room here, at ONC.
>> John Lumpkin:
Great.
>> Chip Kahn:
But I think the next one, number 4 is pretty easy, I assume, approving our draft meeting notes and testimony summary from second. John, should we proceed to that just for a moment?
>> John Lumpkin:
Sure.
>> Chip Kahn:
And ask if there's any objections or specifics that would change anything that is in that. Or we can -- Kelly, is there anything that needs to be said about the testimony summary?
>> Kelly Cronin:
No, I think we've disseminated it and we haven't received any corrections or responses, so I'm assuming everyone's okay with it. We just need to confirm it today.
>> Chip Kahn:
Okay. Well, it sounds like we can move on to the next item.
>> John Lumpkin:
Great. I think we have a set of slides, do we not?
>> Kelly Cronin:
Yes, they're just coming up now.
>> John Lumpkin:
I guess if I had logged on to the website, I would know that.
Let's see. Let's start off with again remembering that we identified some priority areas to work for as a Workgroup, and adverse events reporting was the second area after our first Workgroup effort, on biosurveillance. After the name change we did the report which is going out on case reporting, and now we're addressing adverse event reporting.
As you can see, our broad charge of the committee, which is to make recommendations to the Community to facilitate the flow of reliable health information among population health and clinical care necessary to protect and improve the public's health.
So within that context, you can see on the third slide that we're discussing now adverse event reporting. We split that into basically into two groups. And the first group is us that adverse events that are related to medical products including drugs, biologicals and medical devices, which tends to follow a different pathway, with significant involvement of the FDA in monitoring the use of these devices. The process of reporting on those are fairly clear-cut as to the purpose, and the importance of that is not controversial to most sectors of the provider community who would be the data. the ones who would be providing the data.
The second area are other adverse events, including drug and devices, but also things such as, things that may be called sentinel indicators of lack of performance of the healthcare system, or potential lack of care of the healthcare system, events related to falls, anesthesia events, near-misses, the National Quality Forum “never events”. All of these sort of things are required in some states to be reported, and other states they're voluntarily reported. But the issues related to that we felt ought to be separated from those where there's clear agreement and requirements for reporting on those kind of things, which are the drugs and biological devices.
So having gone to there, and as we're going through the slides, I think it would be useful if you would be walking through the document called Recommendations to AHIC for AE, for adverse event detection and reporting. Does everyone have that? I'm assuming that if you didn't, you would be complaining. So at this point --
>> Lisa Rovin:
I'm sorry, this is Lisa Rovin. I'm confused about, are we going over the recommendation letter that's going to be presented at the 13th, or is this now something different?
>> Chip Kahn:
This is something different.
>> John Lumpkin:
This is something different.
>> Lisa Rovin:
Okay, we'll be going over that letter later?
>> John Lumpkin:
We've gone over that a whole lot, so unless there's still someone that has something left on that letter.
>> Lisa Rovin:
Yeah, our comments were not taken into account and as it currently reads we're not comfortable with recommendation three point zero, for the reasons we put into our comments last time around, and the time before that.
>> Chip Kahn:
John --
>> John Lumpkin:
How do you want to deal with that procedurally? I think that --
>> Lisa Rovin:
Tell me what you want me to do and you can continue with this if this is different and we can take it up later, or another time, whatever you'd like.
>> John Lumpkin:
Okay, why don't we -- Chip, what do you want to do? We need to get that done.
>> Chip Kahn:
I guess my suggestion would be for you, I think you should continue through what you're doing now just so we finish the whole because this just leads to a discussion because this is in a sense just a discussion item.
>> John Lumpkin:
Right.
>> Chip Kahn:
And then the next item ought to be, I guess, we should put the letter in as the next item if there's an issue to be raised.
>> John Lumpkin:
Okay.
>> Chip Kahn:
Particularly since I've got to leave at 2:00, so I guess we should --
>> John Lumpkin:
So we will come back to that letter.
>> Lisa Rovin:
Great, thank you very much.
>> John Lumpkin:
Okay. The next portion of the letter will then talk about the Patient Safety and Quality Improvement Act. And I think someone was going to talk a little about that. Isn't that correct, Kelly?
>>
I think we were hoping that Amy Helwig could expand, perhaps, on that a little bit.
>> John Lumpkin:
Is Amy on?
>> Kelly Cronin:
Just to say how it’s relevant to the scope of what we're interested in and what was in the background.
>> Amy Helwig:
Sure. The next section of the text really covers some broad background information on the Patient Safety and Quality Improvement Act, and that is currently under draft in terms of regulations. But AHRQ has been delegated by the Department to coordinate the efforts to help with the implementation of this act.
>> John Lumpkin:
Okay.
>> Kelly Cronin:
Since I worked on this act on the Hill -- this is Kelly -- I can expand a little bit and say it does create new organizations called patient safety organizations that will be qualified and they will be receiving adverse event reports of various types, that fall mostly into the second category, John, that you referred to up front. And AHRQ has been working with the other agencies and trying to plan not only in response to the IOM patient safety report, the most recent one that we've talked about as an ad hoc group, but they are also in the middle of their plans in implementing the Act as Amy said. Which involves some work around standardization and common formats of adverse event reporting of different types and in addition to taking on taxonomy.
>> John Lumpkin:
Right, and we also noted in the text, going on from there -- well, are there any questions about that portion? So we're talking about the rough outline for the document under subheading 3, Patient Safety and Quality Improvement Act, and there's some text, what we picked up from the hearings and subsequent discussions. Are there any comments on that section before we move into the recommendations?
>> Michael Barr:
This is Michael Barr from ACP. Just a question, is there a copy of the draft that's available for a review? Of the actual act?
>> Amy Helwig:
Yes. The act was from 2005. So yeah, that's --
>> Michael Barr:
It's on the Web?
>> Kelly Cronin:
Yes, there is various versions passed in the House and the Senate. It finally went into law about a year and a half ago. So we're just much more -- we're in the implementation phase now.
>> Michael Barr:
Thank you.
>> John Lumpkin:
Okay. Any other questions?
Okay, so the recommendations on slide 5 -- let me just say one other comment about the states on page C. There are a number of states where public reporting is already beginning to occur. And so as we look at the recommendations, it will make sense to make recommendations that will rationalize the process of reporting so that they're consistent, while the content may change from state to state as to what's required, the form and standards engaged, involved in reporting ought to be harmonized.
>> Laura Conn:
John, this is Laura. ASTHO is also working on pulling together for us about what is reported and required in states, and I don't know if they want to say a couple words about that process.
>> Angela Fix:
This is Angela. Actually, what we're working on is just a couple different things. We were in conversations with Amy as well as with some people over at NASHP, the National Academy for State Health Policy, looking at their toolbox and learning about what they're doing and doing some minor research on our own as well, but not going too in-depth because we think a lot the stuff that is going to come out of NASHP is going to be helpful for everyone. And they're compiling this matrix, really, of 27 states and, to see what state legislation requires for adverse reporting. So we didn't want to duplicate their efforts too much.
>> Kelly Cronin:
And when will that be available, Angela?
>> Angela Fix:
They're saying that they will have it sometime in April. But Joe Rosenthal indicated that it might be possible to share a draft of the grid sometime in March.
>> Kelly Cronin:
That would be great.
>> John Lumpkin:
Great. Other questions? Okay. Let's move into the first round of recommendations. And these relate to overarching issues. The first one is support -- and obviously this is not the exact wording, but this is the general gist of what this recommendation would be -- support for ongoing processes and activities resulting from the IOM recommendations. So for instance they recommended that a common reporting format to serve as a common language for patient safety reporting. The act that we mentioned, the Patient safety act, Safety and Quality Improvement Act of 2005, authorizes the secretary to develop common formats for reporting patient safety events. So that would be what would be embodied in recommendation number one.
>>
Just a quick note. Should we be specifying which IOM report? Because the text mentions “To Err Is Human” but I think a lot of this is coming out of the “Achieving Standards” report.
>>
Right.
>>
We'll definitely be refining the language and including the specifics on this.
>> John Lumpkin:
So we'll include the specific report and specific recommendations from those reports. Other comments?
Moving on to 1.1. This recommendation would be to support within the Department that work by AHRQ in coordinating efforts for common formats for patient safety event reporting in conjunction with CDC, CMS, FDA, HRSA, IHS, and NIH. And Amy, if you might give a brief --
>> Amy Helwig:
Sure. The work that AHRQ is coordinating, we've been delegated by the Secretary, as John mentioned with recommendation 1.0, to work on developing some standard procedures for patient safety event reporting. We are looking to our federal partners for their expertise. As you know we've heard from some of these agencies at our last Workgroup on the reporting systems that they have developed. So we're coordinating the efforts of trying to use their expertise in the areas where they've already developed some reporting systems, and also working on filling in the gaps where other reporting systems or subject areas need to be developed further. This work has been ongoing for about the past six months and we expect it to continue forward.
>> John Lumpkin:
Okay. Any questions? Okay. Moving on. Recommendation 1.2, overarching recommendation 1.2. To do an evaluation of the VA HAI-ISS program, once it's implemented, to understand its effectiveness and how similar programs to be pursued by other healthcare providers and how health information technology can be developed and implemented to enable such programs. So the question is, is there someone on from VA?
>> Sunanda McGarvey:
This is Shu, I just got a message back from Shawn Fultz that he could not join. He has a conflicting meeting. I'm in the process of e-mailing to see if he has an alternate who could join us.
>> John Lumpkin:
Okay.
>> Kelly Cronin:
I think it would be important to get his input on this, John. They are in just an early implementation phase, but they also I think had one of the most impressive testimonies that we heard in our all-day meeting, and they have such an innovative program that there will be a lot to learn from that. And it seems that when it's appropriate, a program evaluation could really be useful to not only other federal systems, but other hospitals or like healthcare delivery systems.
>> John Lumpkin:
Do we have anyone else on the call who has any suggestions for --
>> Dave Parramore:
This is Dave Parramore from the DOD. Perhaps it might be helpful to the group, I don't know if you can find it helpful but we could send from the DOD also our efforts in this regard. We gave a demonstration actually to the AHIC members on behalf of Dr. Winkenwerder, of our efforts with the surveillance tools and I can send you some paragraphs on that if you think it would be helpful.
>> John Lumpkin:
Great, that would be --
>> Kelly Cronin:
Dave, have you done any evaluation of how effective they are for nosocomial infection surveillance, for example?
>> Dave Parramore:
Absolutely. You know, we have a globally available electronic health record, and particularly our surveillance tools coming out of Iraq, Afghanistan, and Kuwait, the members of AHIC thought, particularly the first responder tools, were very helpful in this effort.
>> Kelly Cronin:
But specifically for patient safety, have you looked --
>> Dave Parramore:
Yes, there are some efforts ongoing with patient safety and I could get you that information as well.
>> Kelly Cronin:
That would be great. Because if there's something you think is particularly important that would be good for others to learn from, then maybe we can figure out how that could be disseminated or maybe formally evaluated.
>> Dave Parramore:
Okay. I can have something for you probably by Monday if that's okay.
>> Kelly Cronin:
We're going to be working on this for quite some time but when you can, that would be great.
>> Dave Parramore:
Okay, thanks.
>> Jonathan Einbinder:
This is Jonathan Einbinder. I guess a quick question, when I think of sort of patient safety surveillance for adverse events, nosocomial infections are one dimension of it. Is there sort of a assumption or is it taken as done the work that's already happened in other areas of adverse event surveillance, been a fair amount of work from places like Intermountain Healthcare and Brigham and Women's Hospital, among others, on adverse events and more on medication adverse events than about hospital-acquired infections.
>> Amy Helwig:
This is Amy from AHRQ. Just to comment on what Jon has said, with the work that AHRQ is doing in terms of developing common formats, not only are we looking to the work that's been done by the federal partners, but we also have done a survey of the private, or the private and other hospitals patient safety event reporting systems, and that would include the programs that you've mentioned, and it kind of looked to see what the different formats and standards are. So we're taking those into account as well.
>> Jonathan Einbinder:
Very good.
>> John Lumpkin:
Okay, other suggestions of things that can be included in this recommendation?
>> Kelly Cronin:
I think, John, it might be helpful to ask to have a little bit more of a discussion on what people think they heard from Ron Valdiserri, from the VA, who I think also has a lot of experience from CDC, and, you know, the other folks that are really familiar with these other hospital systems. I know Premier has recently come out with some announcements in the hospital systems that their hospitals have implemented, as being quite successful. So I’m just wondering if we could stimulate a little more conversation around this and find out what specifically should we be interested in, if evaluation is the right kind of recommendation to be making, what specifically should we be looking at, or recommend that they should be looking at?
>>
Well, who would be evaluating what? I mean, what would be the goal of evaluation? Would we be evaluating to see which is the most effective in terms of identifying events? Is that what the goal of the evaluation --
>> Kelly Cronin:
I guess one basic assumption, too, is are we talking about hospitals that have a full-blown EHR implemented? Or are we talking about that have some level of automation?
>>
Having an EHR or any level of automation is not --
>> Kelly Cronin:
But -- right, but I think our Workgroup interest is how do we connect the dots when it comes to health IT. So how can health IT be an enabler in these overall reporting systems or surveillance systems in the in-patient setting?
>>
I don't think that you'd want to limit it only to the in-patient setting. So that's one comment. And then you're asking should the Workgroup focus on only events that would be tracked by HIT? Or reporting systems that would be using -- is that what you're asking?
>> Kelly Cronin:
Actually I'm just asking for everyone to contribute their ideas on what they think they heard, because we had a lot of rich testimony which we tried to summarize, and some of the concepts that were presented were very new and innovative and I'm wondering, are there any in particular we should have interest in that will really advance this area, could potentially.
>>
You have a little bit of a moving target because what, has in some of the other areas, what EHR does is change some of the inherent principles that are in play. For example, one thing that's been talked about quite a bit is prevention by identifying or screening patients in advance. For example, why wait until a patient comes into the hospital and then falls? Wouldn't it better if you had an EHR for an elderly patient come from a nursing home or even from home, if you had the doctor's record that said this patient has trouble with balance or this patient is at risk for falls? That's something that it's very, very difficult to do in the absence of an EHR, because you're getting a very thick chart. And even where you have, for example, on the front of the chart you got that sticker that says allergies. The physical chart doesn't always move, so you have a patient comes in, the family is not there, that penicillin allergy, as we saw in the NHIN prototypes, but that begins to change the way you actually do patient safety. And the way -- because we have built, and a lot of what we do in patient safety is built on the foundation of a non-EHR environment. So it becomes a little bit of a moving target.
And so you have the requirements that you have at AHRQ to do something now in the very proto-EHR environment. But we also have to be building for a future environment that may give us a lot of different options. So you really need both. You need to be doing what you're doing now, because we have to deal with the facts on the ground as of 2007. But the question is what can we be doing, and the VA gives us a window into that. Because of the accessibility of the data from -- so that's the kind of evaluation that's helpful. It's not only how is the VA using their data to prevent or to deal with patient safety issues now, but what is in the VA system that enables them to do other levels of patient safety prevention. And there are other people who have those kinds of systems.
>> Kelly Cronin:
Chip, I don't know if you know about the Premier hospitals, but I know about a week and a half ago there was a big announcement around that.
>> Chip Kahn:
Yes.
>> Kelly Cronin:
Perhaps some of your members might also have some innovative systems.
>> Chip Kahn:
But when you say -- when you look at the Premier demo, though, is that what you're referring to?
>> Kelly Cronin:
PMS demo. This was really specific to surveillance for nosocomial infections.
>> Chip Kahn:
But we're sort of only at the very beginning of that. And that in a sense is, I mean, that, there's no real-time here. This is all stuff that is then collated and collected and reported on a, I don't know what all the time frames are, but on a much broader time frame. So I don't know, but, and we're only going to begin to do the SKIP measures in the next sort of two or three cycles. So I guess the answer is there will be measures, there will be some measures in the next, I don't know -- and let me also say, I don't know were nosocomial infections in the agreement for the second round of Premier? Is that what you’re saying? I haven't looked at the items that are going to be in those --
>> Kelly Cronin:
I wasn't even really thinking about that. I think the announcement was more global about --
>> Chip Kahn:
Yeah, because they said they were going to add measures. The measures that are now being used are not measures that I think would be directly relevant to what we're talking about. But there are 33 measures now I think being used in Premier. But they're going to expand those measures so this may be, I'd have to check and see whether this is one of the areas they're getting into. And I think there may be some areas that our guys are tracking for their systems, but I have to check on that. And there are a lot of states out there that are, and actually, I got a note the another day, I mean the states are asking for inconsistent data with inconsistent criteria, and that's actually sort of, to take the Secretary's language, that's a hair on fire issue right now for the hospitals. So actually this is not a happy issue for hospitals.
>> Kelly Cronin:
Right. I mean, one of the things we might want to be mindful of is the adoption rate of EHRs in larger hospitals is higher than in small and rural hospitals. And it's also significantly higher than in ambulatory care right now. And if there is, you know, sort of continuing investment in these large hospital systems over time, and more robust interoperable electronic health records as they become more interoperable; there an opportunity for us to be thinking about patient safety type functionalities in those records? And could we be looking at the VA system and their approach and perhaps the approach of what Brigham & Women's or InterMountain Healthcare or other like systems are doing, even though theirs are more homegrown, well, they're all homegrown, we could be looking at a couple of select systems and really be perhaps, you know, helping some of the other hospital systems understand what they could do to decrease their infection rates or prevent errors.
>> Chip Kahn:
Well, I think, you know, the more you talk, as I'm listening to you, I wonder whether this is an area that we might want to have testimony from specific hospital systems and from some states that now have new programs on reporting infections.
>>
So it also sounds like we may be talking about not just doing evaluation for evaluation’s sake, but evaluation for the purpose of looking at scalability, of looking at transferability of systems from places where they have grown up to places where they may not exist or may be starting, and seeing what systems are transferable, what systems are scalable in all these homegrown systems. Some may be better able to be adapted for broader use than others. So maybe talking about demos in a sense, or demos of --
>> Jonathan Einbinder:
This is Jon Einbinder again, but just a comment. I think that infection surveillance is a sort of order of magnitude harder problem than, say medication adverse events surveillance. And you know, at least from my perspective, places like the VA will be on the cutting edge showing that there's value there and how it might be done. But to get to the point about EMR adoption, the VA has infrastructure that really very few places in the country -- I'm not sure how generalizable that experience will be to other places in the country and I still, kind of to reiterate my other comment, wonder if the technology that is ready for generalizable is more of the medication adverse event or surveillance or data triggers if you will. I know it's in the list but I do want to make sure that's not lost sort of in the recommendations. I think there's real value to trying to get that more broadly adopted.
>> Chip Kahn:
If I could read from an e-mail I just, I was looking at my e-mails here. If I understand it, 13 or 14 states have signed up with CDC's new version of, see if I get this right, NNINS, as a mandatory way to submit to the states for view and public disclosure hospital-acquired infection rates?
>>
Yes.
>> Chip Kahn:
And New York and South Carolina first followed by California and many others, and the problem is -- and it starts -- the abstraction -- let's see. Abstraction has to begin by July of this year. First of all, the definitions are a little squirrelly, I guess from the hospital standpoint but in addition they're requiring patient level data, multivariates, up to 30 per case including microbacteriological data and bug sensitivity on every surgical procedure performed in a hospital, as well as ICU and general floor surveillance. Not just the surgical cases with infections, all of them, and then the person goes on to say this is absurd and truly unachievable without four to six data abstractors in big hospitals like MGH, and one to two in small community hospitals. Now, I suppose that the same information could necessarily be put, be done electronically, and I assume at MGH it probably is electronic, but that doesn't mean you won't have the data abstractors on the other side.
>>
MGH is one of my institutions that I work at, and the data is not electronically available. It is available for review in text form. But in terms of parsing a microbiology report for sensitivities or, you know, automating the reporting processes, that's nowhere in our probably several-year horizon, at a place like MGH, which really is farther down the adoption pathway than many other places. My guess it is in the horizon for a place like the VA but I'm not sure how many other institutions would fall into that category.
>> Chip Kahn:
Well, anyway, but I think then what we're seeing here is this is a theoretical discussion, to some extent.
>>
I mean, the reality is to comply with the programs, it is going to be manual abstractors for the foreseeable future for most institutions. And you know, it would be wonderful to build out the informatics structure to automate it, but it really is much harder to get all the things you just mentioned. The cultural results, sensitivities, et cetera, for every procedure, compared with, say, medication administrations and lab results, which are what you need to do medication adverse event triggers.
>>
And that helps to define the gap. I mean, as you say, and everybody's experience is similar, in any one hospital, trying to automate those systems and bring all the systems into some level of interoperability is a very time-consuming and very expensive task. And that, I mean, the gap that we've been talking about is the demand for quality and patient safety data right now today, which the states are legislating in many cases, and our long-term vision of having all this available for the EHR, and I think that one of the questions that we're talking about right now is are there ways that AHIC can help facilitate closing that gap, and what are the things that AHIC could do in terms of figuring out how to close that gap. It may not be bridgeable. I mean, it may that be every single hospital has got to go through this on their own. But I don't think that that's proven. I think it may be worth some effort to see if there are scalable or transferable or generalizable models in some places that can be more easily adapted than having to have each hospital reinvent the wheel.
>> Kelly Cronin:
Yeah, this is Kelly. Just to add on to that. I think there might be some near-term opportunities in trying to identify what might be able to be accomplished in the next few years as the adoption rate increases in hospitals, and you know, what are some incremental steps that might be more feasible? And we do have a medication management use case that's being put for the now to the Health IT Standards Panel, to the Certification Commission, so that when the in-patient EHR certification process is in place, I think there will be a fair amount of specificity around medication management and the module for computerized physician order entry and John (indiscernible) may want to speak to that but -- and thinking about Jon Einbinder’s comments, there is perhaps more of a near-term opportunity to do something around medication management where ADEs could be triggered, and perhaps that is something we could consider, given that some of the microbiology and other kinds of lab results needed for nosocomial infections would be more complex and require much more sophisticated interoperable systems in the hospital.
>>
I think you want to do both, but it is sort of a portfolio and one is kind of near-term maturity and one is a longer-term maturity. I think the third area, which is, it is in the document, is going to be dealing with text. You know, whether you call that Natural Language Processing or looking for triggers in text, but I think that's the other really rich area of work.
>> John Lumpkin:
John Lumpkin here. As we're moving through this conversation and thinking about recommendations, Kelly or John Loonsk, any thoughts to what extent this issue can be addressed within the quality reporting use case that was identified by AHIC as being one of the next set of priorities?
>> Kelly Cronin:
And what specifically around the quality use case? I'm sorry, John.
>> John Lumpkin:
This concept of identifying, you know, if you're going to do quality reporting you have to identify what it is you're going it report.
>> Kelly Cronin:
Yeah, and certainly the data harmonization around all of the data elements that would be in the numerators and the denominators of a core set of measures is hopefully going to get done over the next several months and that will be fed into the Health IT Standards Panel. So the priority is, really we're specifying that we start with the core set of measures, building off the work that the Ambulatory Care Quality Alliance and Hospital Quality Alliance has done and that will be sort of a starting point.
>> John Lumpkin:
Okay. That will be would be a little tough to piggyback off.
>> Kelly Cronin:
I don't think they're getting into the sphere patient safety right now. But as both of our groups proceed, we're obviously going to be overlapping in this area potentially and we’ll have to coordinate that. But I think at this point they're not clearly taking on patient safety measures. At least not adverse drug events and I don't think nosocomial infections are in the HQA set.
>> John Lumpkin:
Okay. Any other avenues we should go into as we're looking at this particular recommendation? It seems to me that we need to think about the emerging systems that are able to track this, but also recognize the unevenness of development of electronic health records and their capacity to be able to detect these kinds of events.
>> Dave Parramore:
This is Dave Parramore again from the DOD. I heard it mentioned earlier that there might be a possibility for demo. I'd like to reiterate I can certainly arrange for that if the group would be willing to bring in a group to do an AHLTA demo and a surveillance demo relative to this topic.
>>
Yeah, just to reinforce, there is certainly a space where the DOD systems and the VA and things like the National Healthcare Safety Network, these are systems that all address somewhat different parts of a problem. What they, each system does what it does extremely well, and there may be ways of building on those systems and using at least parts of those systems to adapt to other settings. I mean, that, I think, are some of the demos that we're talking about.
>> Kelly Cronin:
And this is perhaps maybe a bit of a controversial system but are there surveillance modules in AHLTA or VistA that would be open source that the vendor community might be able to take on and innovate around?
>> Dave Parramore:
Yeah, this is Dave again. We've certainly had those discussions at the DOD, and that possibility exists. You know, we're already doing a lot of interagency work with the CDC. We've also explored the Department of State and of course we're interested, particularly with Tricare, sharing data with local, state and national level to our beneficiaries in terms of electronic health records. So there's certainly that potential.
>>
I mean that's, you know, that could be a potentially dramatic step forward, is talking about open source systems for public, or any kind of adverse event reporting. It is controversial, obviously.
>>
Yeah.
>>
It's controversial but it would be a very powerful step even to engage in that dialogue.
>>
I think, you know, along the same open source sharing of the work that the AHLTA and VistA, I think it would be interesting if you have any demos relating back to us in a prospective identification of potential -- you had mentioned it, John had mentioned it, Kelly mentioned it, triggers which may alert -- and Steve even you on the inpatient, triggers that may alert to an event, not necessarily having to wait until the event has occurred and then working on generating a report. So any collaboration around that area in terms of working with the vendor community as well would advance quite a bit.
>> Dave Parramore:
Okay, this is Dave again. I'd be glad to do, or at least organize that demo. Perhaps I could take this discussion offline and send the group some information for perhaps follow-on demo.
>> Kelly Cronin:
One aspect that might be of particular interest that might be of less interest to the vendor community, of course I think we'd have to talk to them to know this, but if there's a way to be figuring out how to automate the report to go from a given provider to a patient safety organization or to a public health agency in such a way that it wouldn't require a lot of unique features or functions on a standard EHR. That could be something that would be publicly available and just add it on as a module to any interoperable EHR.
>> Jonathan Einbinder:
This is Jon Einbinder again. But that functionality that has been demonstrated at some of the HIMSS, I don't know what they're called but this interoperability demonstration things that they have, the ability of an EHR to detect or report an adverse event and automatically generate, for example, an FDA report. A lot of that has to do with HL7 version 3 stuff.
>> Kelly Cronin:
Well, John and Chip, is this something you think we should explore more, or do you want, do you think it's premature?
>> John Lumpkin:
You know, I'm getting the, I'm just wondering if this recommendation and these ideas will be on the same track as the rest of the recommendations, given the state of the art. So the question is, if we want to explore this, do we want to explore this on a separate but parallel timeframe? And make something in a little bit more detail here as a subsequent recommendation to the rest of them?
>>
I would endorse that, John. I wouldn't drag it out over a long period of time, certainly. But some more offline discussions about some of the potentials there I think would be very valuable, and I think as long as we set a timeframe for doing that, so you know, we're all kind of forced to bring it to closure. Because I think that this is, it's potentially very powerful but it is somewhat controversial and I think we need to think through what we're suggesting before we make it too public.
>> John Lumpkin:
Other thoughts on that? Do we have any volunteers on the Workgroup to continue to play around with this area? Tease it out a little bit more?
>> Kelly Cronin:
I could talk to some of the folks at the VA since they've already been through working with WorldVistA, a lot of the issues around making their EHR open source, and it's been with enormous controversy and, you know, I think given that the concerns that have been expressed by industry over the last two years with respect to the VistA-Office EHR, which was intended to be an open source EHR that other vendors could innovate around and get certified. It still, it was just considerable concern in industry. So I think if we're doing this for the good of public health, and it would just be a distinct potentially one module that would be trying to connect to public health, then maybe that's something that is discrete enough that would be of less concern. But I think that given the significant controversy over VistA-Office EHR, we do need to really look into this carefully. So I think, you know, perhaps staff will just get some more well thought out opinions about what would be involved with this, and perhaps even get some input from industry before we would take this on in a more comprehensive way.
>> John Lumpkin:
Okay. Sounds good. Shall we move on?
>> Chip Kahn:
Yeah.
>> John Lumpkin:
Okay. The next slide moves into the second area, which is adverse event detection -- I'm just shifting my papers around -- which is to facilitate public access to clinical data from HIEs and RHIOs. This could involve memorandums of understanding or standardized data sharing agreements to support sharing of data with authorized users from public health. This may require a definition of roles and authorizations required for stakeholders outside of clinical care to receive and analyze data for purposes of adverse event detection. This would create the policy and procedure infrastructure for the data sharing.
>>
John, just to let you know, on the policy end of this, I think the NCVHS is likely going to take on more of a, both a conceptual framework around appropriate secondary uses, and also think through some of the specific issues with individual use cases like quality measurement and reporting. And we're just exploring it with them, but I think they have a lot of depth in this area, as you know. I think you probably took it on in part when you were chairing. So I think it's likely that this more global issue will probably be dealt with elsewhere.
>> John Lumpkin:
Okay. So do you think we can just perhaps note that in the text and take the recommendation out?
>> Kelly Cronin:
Yeah, I guess it would be good to get feedback from some others, because it's an important issue, but I don't know if there's any other more practical considerations around MOUs or data sharing agreements that would be worthy of taking on. But I think the policy aspects of this can be dealt with elsewhere.
>> Sunanda McGarvey:
And Kelly, this is Shu, we did get feedback from some that we didn't have time to incorporate into all of this that suggested that the biggest obstacle was not so much the MOUs or data sharing agreements, but more the resources to actually, once the data was available, to be able to analyze and determine, you know, determine what to do with it.
>> Kelly Cronin:
Right, and I think that's probably really going to be most applicable in 2.2 and other pilot-oriented recommendations.
>> Lisa Rovin:
This is Lisa at FDA. I'd like to support Kelly's suggestion. I think the NCVHS has a wealth of experience that we could benefit from and it would be good to get them on this and see what happens.
>> Leah Devlin:
And this is Leah Devlin. I guess I'd just like to say that we need to keep that resource issue in the back of our minds as we move throughout this whole technology piece. Because if there's nobody with some smarts there working it and receiving it, and making some sense of it, you know, we haven't really done too much.
>> Lisa Rovin:
That's right. This is Lisa at FDA again. We were one of the people that put in the comment that Shu just shared. We have a very recently executed data sharing agreement with the VA and the obstacle now is figuring out what people actually have time to do.
>> Brian Carnes:
This is Brian Carnes from the health department in Indiana. We're exploring research institutions such as Purdue to help us disseminate that information, and actually do research projects and engineering projects on how we can better our processes.
>> John Lumpkin:
Other thoughts around this issue of whether we include this as a recommendation or just include it in the text outlining in some of the work that's been mentioned?
>>
Well, what's the recommendation specifically that --
>> John Lumpkin:
Well, the outline of the recommendation would be that we facilitate public health access to clinical data and through the HIEs or RHIOs and that this would be done through MOUs or standardized data sharing agreements.
>>
And so really the comments apply to the recommendations, I think that are really on the next slide, because we need proof of concept, we need pilots, and what several people have talked about, Natural Language Processing, data-driven triggers. I mean those are coming up on the next few, so this recommendation maybe it is more for the narrative, that this is the reason that the rest of those recommendations are in there, that we need to facilitate this and the following recommendations suggest ways of facilitating it. Is that another way of looking at it, John?
>> John Lumpkin:
Yep.
>> Lisa Rovin:
I think that's a great idea. This is Lisa at FDA. I think that's a really good way to handle it.
>> Kelly Cronin:
I guess I'm also wondering, from a CDC perspective, if in fact -- I guess one of the things we are going to have to specify as we start drafting this letter; what would be the scope of adverse events that you would be interested in trying to detect in a pilot. And will we be going beyond medications realistically. And if we were, and the data could be flowing to more than one public health agency, would there be any utility with having just a model data sharing agreement that could be more widely used. You wouldn't have to be replicating it with every jurisdiction. So it's really much more of a practical tool that could be used by a variety of public health agencies clearly for the purposes of public health.
>>
That's a little bit of a different recommendation.
>> Kelly Cronin:
Yeah, it is. I think we need to think about what would be useful and not --
>>
I mean, but it's a very relevant comment. That changes the recommendation to say that AHIC should look at the development of model data sharing agreements based on what's out there now. And some of that, and I'm not, there may be people on the phone who are experts. I mean, we got some testimony about that. Some of the state regulations, they dictate what state agencies do or do not have access to. So there's that, you have a continuum there. Some of it is statutory and regulatory at the state level. And some of it may be non-regulatory or voluntary. I mean, you mentioned Premier as an example. I mean, there are different models.
>> Art Davidson:
This is Art in Denver. I was just wondering, it seemed like the focus of the discussion was earlier on figuring out maybe some prototype or some pilot to show that there's a way to make this happen. And I think Kelly's last comment about working on model data sharing goes away from that theme that I thought we were talking about, which was proving the value of this. I mean, it would be great if we could do both, but I think just proving the value and how to do it is probably a first step. And given just the last commenter said, given the local and state jurisdictional rules that may determine how data is shared may be too big a step to try to work on a standard at this point, in terms of data sharing agreement.
>> Kelly Cronin:
That's a really good point, Art. And I think getting back to Nancy Foster's testimony and what some other people said in our last meeting, there's so much complexity now where it's not feasible to determine data requirements right now. So the model, a model data sharing arrangement may be difficult to even craft if we're not clearly understanding what the scope of data would be that we'd be interested in.
>> Paula Soper:
This is Paula from NACCHO, for informational purposes, I do want to let you know that we just launched, actually today, the survey for RHIOs and state and local health departments asking them about their relationships and if data is being exchanged or not. One of the questions we did ask is, is there a data sharing agreement in place and if they were willing to share that data sharing agreement with us. So hopefully we will at least be getting these data sharing agreements in and maybe at a later date we can talk about how we want to look at those and review those for this purpose.
>> Chip Kahn:
John, this is Chip. I'm going to need to go in a minute. I wonder whether we could approach it, the other issue differently than I had suggested because this looks like it's going to go on for a while. And break from this and go to the other issue if you want me here. I guess I feel like I should be here for this discussion since I would be presenting with Steve at the AHIC meeting on this.
>> John Lumpkin:
Okay, if that's agreeable, we'll then move to Lisa.
>> Lisa Rovin:
Yeah, this will be very fast, Chip, I won't hold you up.
>> Chip Kahn:
Okay.
>> Lisa Rovin:
The same thing we discussed at the last meeting. Recommendation 3.0. It’s the question of risk communications. Our risk communication, it's a science and people figure out what kind of format and content are best for particular subaudiences. So I had asked that this recommendation acknowledge that, which it does with respect to content but does not with respect to format. So I would like to see in the first, sentence health alerts including broad categories of content, meta data about the content, and format. And similarly in the last sentence, should not impede risk communication needs to optimize alert content, and format. You could even just add it in the last sentence, it looks like, if you just said, should not impede risk communication needs to optimize alert content and format, just adding those two words at the end now that I look at, it may do the trick.
>> Chip Kahn:
Let me ask Kelly. Is that okay? Is there an issue?
>> Kelly Cronin:
Yeah, it sounds fine.
>> Lisa Rovin:
Great.
>> Chip Kahn:
Okay, well, then that solves the problem.
>> Lisa Rovin:
Yep, we're done.
>> Chip Kahn:
So John, unfortunately, I need to go. But I'll check in with you and Steve later and figure out where we are. But it's been a very helpful discussion, and I look forward to our next.
>> Lisa Rovin:
Thank you, Kelly and Chip.
>> Chip Kahn:
Thanks, bye.
>> John Lumpkin:
Now, if I can sort of describe where, what I'm hearing, it sounds to me like we need to, that this area of data exchange is not mature enough for us to suggest that standards be identified. But that it is maturing as we speak. And perhaps what we ought to recommend is that the HHS closely monitor developments in this kind of data sharing, so that at the point in which it seems appropriate, that standard MOUs and standard data sharing agreements can be developed.
>> Kelly Cronin:
John, I think with one caveat. There are more standards around, in the medication area, and I think also FDA has a standard for adverse event reporting. So I'm wondering if based upon Marc Overhage’s testimony and some of the work that we'll be hearing more about in Boston, the Center for Bioinformatics at MIT that both industry and FDA are involved with, I think that it, there might be a ripe enough opportunity when it comes to drug-related surveillance. Perhaps not for some of the another categories of adverse events that we're talking about, but for drugs it may be more tangible.
>> Farzad Mostashari:
This is Farzad Mostashari, New York City Department of Health. There were a number of health information exchanges, RHIOs, in New York City that are under development that we're partnering with around various public use cases and one thing I want to mention in our legal discussions has been the concern from public health that we, we may, where we are authorized to receive and particularly where it’s mandated reporting, it may set a bad precedent to even engage in kind of MOUs or data sharing agreements, in terms of the precedent that might set. And the position of our general counsel has been that if it's mandated reporting, to public health definitely, then, you know, it's mandated into public health and you don't need a MOU for that.
>> John Lumpkin:
Well, --
>> Farzad Mostashari:
I just want to put that --
>> John Lumpkin:
Yeah, let me play with that a little bit. Who is mandated to be reporting?
>> Farzad Mostashari:
The hospitals.
>> John Lumpkin:
And so a, an MOU may be for a hospital to have an intermediary, in other words, the RHIO, or --
>> Farzad Mostashari:
Yeah, a lot of the --
>> John Lumpkin:
The RHIO is itself is not a mandatory reporter.
>> Farzad Mostashari:
Correct. And at least the structures that have been commonly deployed in New York City have maintained the reporting relationship of the hospital through the RHIO infrastructure, but the group, the data source reporting is still the hospital, rather than deferred that to an intermediary role, which introduces all sorts of other potential headaches for the RHIO. So our RHIOs have tended to say you can report from the infrastructure that is developed by the health information exchange, but you own that infrastructure, you own the data on it, you do the reporting. If there's an inappropriate disclosure or whatever, you're responsible, the hospital retains that responsibility and right.
So there may be different models out there. But in many cases, the major issue I want to raise was I want to agree that it in many ways it is not cooked yet and maybe we should kind of watchful, do some watchful waiting.
>> John Lumpkin:
With the caveat that for drug-related adverse events, that there may be opportunities that should be, to be taken?
>> Farzad Mostashari:
Sure.
>> John Lumpkin:
So does staff have enough maybe to try a whack at some language on that?
>>
Yes.
>> Kelly Cronin:
Yeah, I think we can do that.
>> John Lumpkin;
Okay. Maybe we can move on to 2.1.
>> Amy Helwig:
John, this is Amy. I've got a comment about 2.1. And that’s just, essentially, in looking at the layout of the outline, I really think that recommendation 2.1 is better served in the next category, which is probably going to be the 3 series, adverse event reporting. So if we could just move it to that section, I think it's more appropriate there.
>> John Lumpkin:
As I was walking through, I was thinking similar. Anyone opposed to doing that? Okay, we'll come back to it, then when we get down to the 3 series. And unless I'm wrong, at least on my notes, we did skip 3.1, didn't we? There is no 3.1?
>> Sunanda McGarvey:
John, this is Shu. The numbering is just off. There were only two recommendations we had within the outline.
>> John Lumpkin:
So you knew we were going to move 2.1 down to 3.1?
>> Sunanda McGarvey:
That's right, I knew, that exactly.
>> John Lumpkin:
See, I'm trying to give you credit for all of this good stuff.
>> Sunanda McGarvey:
Forethought.
>> John Lumpkin:
So let's move on to 2.2. Which is support proof of concept pilot for adverse event detection, which can be leveraged to the RHIOs, HIEs and possible NHIN implementations.
>> Kelly Cronin:
John, I think Shawn Grannis is our Workgroup member with the most experience, or at least Regenstrief and the folks from Boston are the ones with the most experience related to this recommendation. So we can try to get, I think, more of their input offline. Unless John Einbinder, you might have some insights here as well.
>> Jonathan Einbinder:
Yeah, I think those are the, I think I’d add InterMountain Healthcare to the list, but I think those are the folks that we'd want to get input from. I know certainly on the Boston side and we roped them, in David Bates and Tejal Gandhi but I think that's right on.
>>
Just a couple other groups I'm aware of that I can get you information on as well that are doing work in that area.
>> Jonathan Einbinder:
I think it also depends on, some of it goes to what our definition of a adverse event is. And are we really talking about adverse event detection or prevention or mitigation? I know from the detecting point of view you want to detect them, but kind of with an eye towards what also comes into play. Real-time, retrospective, et cetera. I'm not sure if that needs to be specified or not. But presumably we want these to be not just for sort of aggregation and analysis, but for actual improvement and mitigation.
>>
Uh-huh.
>> Farzad Mostashari:
This is Farzad Mostashari. I also want to recommend the work that George Hripcsak at Columbia and CBMI has done with the adverse reporting using Natural Language Processing.
>> John Lumpkin:
Now, do we want to, is the concept of this wording including not only doing proof of concepts but evaluating them?
>> Kelly Cronin:
Yeah, I think that's a great idea. And I think the other thing we heard from Marc and Shawn is that the definitions around these, as John has indicated, are quite challenging, and Lisa, perhaps it would be helpful to also work with FDA to figure out what would be of most interest and how might we advance on more standardized ways of defining adverse events of interest for the purposes of these pilots? I think -- FDA went through this when they collaborated with several, I think five or more hospitals over the last few years, on a pilot project that, where they obtained data on some rare serious adverse events. I think Columbia was one of the participating hospitals. And they've been through this exercise before where they've tried to sort of carve out some SAEs of interest and defining them but I think it's likely going to have to be a collaborative effort to first identify what would be of most interest to measure and then how would you actually define and measure that.
>>
Sure, absolutely.
>>
I also think on the evaluation side one aspect of this which is very important is kind of the business case, if you will, or the ROI, and I'm thinking of even in a very operational level, one of our community hospitals where we had a relatively modest proposal to put an adverse event detection commercial product for doing surveillance for med adverse events, would have been about an FTE, less than 100 K in capital and ultimately did not fly. So it's kind of what is the business case, or the ROI. There's been some literature on that but I think that's an important piece of the evaluation.
>> John Lumpkin:
Other thoughts on this one?
>> Kelly Cronin:
So in other words, just to expand on that thought, because this happened in that pilot with FDA a while back. It's really, these, participating in these kinds of efforts would be on a voluntary basis with providers. There's no obligation to work with public health on these kinds of projects. And if you'd want something to be more scalable over time, there would have to be a real reason for them to participate. So whether it's for mitigating adverse events or preventing them, or somehow they're able to improve their care, or their patient’s outcome, and that would be something tangible that they could justify their involvement.
>>
Yeah, there's, this is old literature at this point but David Classen and David Bates papers on kind of the attributable, preventable, post-adverse event costs, these are for in-patient medication related adverse events and the numbers are compelling, but I don't get the feeling that those are really, sort of healthcare institutions believe those are real dollars. I think this is one of those places where there's some literature and evidence to be developed.
>> John Lumpkin:
Yeah, they're saving dollars, but for who?
>>
Exactly.
>>
Should there be a recommendation toward a business case here?
>> John Lumpkin:
I think that we can, here's kind of my read on there. It sounds like we already have, that the issue isn't supporting the proof of concept pilots. But that there already are pilots. And that we should evaluate them, including looking at the business case.
>> Kelly Cronin:
You know, this is one area, John, where I think we know of two places in the country, maybe three.
>> John Lumpkin:
Or four.
>> Kelly Cronin:
Yeah. It's just, and they're very different projects. Like the Natural Language Processing in Columbia is probably very different than what CBI is doing in Boston and Marc Overhage is doing in Indianapolis. So I think it would be really helpful to maybe think a little bit more about what would be of most value to public health, as we try to figure out what is going to be more readily available through RHIOs or HIEs as NHIN gets deployed and then maybe build on the proof of concepts that are already out there, but given their variability or different purposes, that we try to at least understand what is the most valuable way we can move forward in the next, say, five years. And it may not be through Natural Language Processing alone. It may be through just data aggregation of all beds and all labs to try to look for a suspected serious adverse event.
>> John Lumpkin:
It seems to me when you sort of take this within context of recommendation 2.3, how much of 2.2 is what you do and how much of 2.3 is how you do it? And it seems to me we have to kind of do those in parallel.
>>
I think that's very fair, John. Just to the question of a business case, do you need a business case for adverse event detection? I think that's the point that you're making. We're past a business case for adverse event detection. That's already happening. The business case is for the effectiveness of electronic health records in, or who has access to that data. It's the same, it parallels the discussion of develop a business case for the RHIOs. I don't think we need a business case for adverse event reporting, do we?
>> John Lumpkin:
You know, I think to the extent that much of it is required, either legally or based upon business agreements.
>>
I'm just supporting the point that you just made, John.
>> John Lumpkin:
Yes, and I think I was agreeing with you.
>>
Okay.
>> Kelly Cronin:
I mean, I think we need to be mindful of the added complexity when you're thinking about pharmaceutical companies getting access to this data. If you know, they do appropriately at some future date start accessing aggregated data for their own surveillance purposes, it adds in a different dynamic. I mean, certainly there's revenues to be generated perhaps from a RHIO perspective if it's appropriate and there's policies and procedures in place to enable that. But we may not be talking about just public health.
>> John Lumpkin:
Right. But I think the issue is is that the business case would be the use of electronic reporting and the use of RHIOs as part of that, as opposed to the business case for doing the reporting at all.
>> Kelly Cronin:
Right.
>> John Lumpkin:
And then, so the way that we're addressing these between 2.2 and 2.3 is conceptually how, is 2.3 is really the methodology. 2.4 gets to that also a little bit. And then the models that would be in 2.2 would be if you do have a methodology that works, does it make sense to do that in an electronic environment. Is that a way to parse out these recommendations? Does that work for anybody?
>>
Could you say that again, please, John.
>> John Lumpkin:
I'm not sure I can.
>> Kelly Cronin:
Yeah, John, I think it might be helpful to, since David Bates contributed to a lot of this and I think we all know he's really one of the foremost experts in this area, I think his idea was that this would be done more within a clinical setting, so it's not so much related to a regional implementation of a health information exchange. It's more sort of what would be done, you know, in a given clinical care setting and how might you evaluate these kinds of tools where an EHR has already been implemented and you're trying to evaluate some new software that would enable this.
>> John Lumpkin:
Are you referring to which recommendation?
>> Kelly Cronin:
2.3, mostly 2.3. I think 2.4 could be, the data mining could be more with aggregated data.
>> John Lumpkin:
Right.
>>
I would just, I would confirm that impression. I think that David is really focused on the, sort of within the institutional electronic health record as kind of a locus of this work.
>> John Lumpkin:
So 2.2, you know, again not looking at the specific words, 2.2 is how to do it at the network level, the regional level. 2.3 is how to do that within an institution or a EHR level. And then 2.4 is looking at it in the concept of data mining.
>>
I think there's an interesting, something I've never thought about before, but the notion of what's the right level of organization for the detection to occur, and then at what level is the data aggregated. In other words, do you aggregate data and then apply detection, or do you apply detection and then aggregate the results, are actually two alternative approaches this. One is more EHR-centric. One is more RHIO, or aggregate database-centric. There's a, in that IOM report, in the patient safety data one, there's a figure that talks about an informatics infrastructure for aggregating this kind of data and it kind of, the model is more the work done at the institution and the results are aggregated using data standards applied to near misses and adverse events. I'm not sure if we need to weigh in on that or say that that's something that kind of, either of those approaches need to be kind of thought about. They are different.
>> Kelly Cronin:
Yeah, I know there's, in partners that I've talked to, John, I'm sure you've worked with, that are really keen on this concept of a more sophisticated approach to Natural Language Processing and trying to be able to query and tease out some of the text-based data that really would be quite meaningful to detect adverse events. And I think David's thinking around this is that there's a couple of tools out there, but there's a lot of room for improvement.
>>
Most definitely. And it sort of addresses a fundamental challenge which is the lack, there can be data standards for capturing things like diagnoses and lab results and medication but a lot of healthcare data is still unstructured. Which it's a problem beyond the scope of what we can solve but you can apply detection tools against that unstructured data. That's kind of the Willie Sutton principle, it's going where the data is, at the same time trying to structure what you can.
>> Kelly Cronin:
And I think FDA has already been interested in some pilots early on with the DOD. Or perhaps that's not something that has been fully implemented. But with the arrangement with VA there might be an opportunity to be doing this kind of data mining and in an EHR environment that wouldn't necessarily require data aggregation. But I think there's also been an awful lot of work done in the last several years by industry and some folks at FDA on data mining techniques for aggregate data.
>>
Yeah.
>> Kelly Cronin:
I think David's only point it has to be clinical data. We're not, I think there's sort of less interest in mining existing adverse event databases and a lot more interest moving forward in mining clinical data, whether it's from an EHR system or it's from an aggregated source like a RHIO.
>>
Well put.
>> John Lumpkin:
So I think that we have three recommendations that are looking at different scopes of, or different levels of where the data is going to be evaluated to detect, and how. I think they're pretty close to right, then, is what I'm hearing.
>> Kelly Cronin:
I think one of the big issues is feasibility. And David has indicated, and I think even the IOM report sort of pointed these toward AHRQ, and I think it's perhaps a question to Amy and her colleagues in terms of what's in your portfolios that might support this kind of work. And could it be done in partnership with FDA, VA, and other organizations that we've talked about.
>> Amy Helwig:
Yes, and if you would like, I can enhance these further with some of the work that AHRQ is doing in terms of adverse event detection within the portfolios and if you want to shape these.
>> Kelly Cronin:
Yeah, and in particular we don't really want to be advancing recommendations that are just not feasible to implement. So if there's really no resources at all available for the agencies or these other outside institutions, let's be reality-based about it. We can recognize the importance, but perhaps not put a recommendation out there if we're not really able to act on it.
>> John Loonsk:
This is John Loonsk, I would second that. These are very broad recommendations and if there aren't resources to pursue them, then I think it's not necessarily that fruitful.
Let me introduce another concept just to complicate things a little further. Which is that one of the opportunities here in making recommendations is to move, is to work toward moving something from an unstructured text-oriented status to a structured vocabulary-specific status through the overall process of harmonization. So the discussion you all just had sort of was one versus the other, but indeed some of the crux of the issue is what's the cost benefit of trying to advance, and what data elements, from the unstructured to the structured, and which ones to focus on, either because, well, maybe they exist or they're well recorded or the workflow for recording them is conducive to structured data entry. And which ones are just not worth the pain in the near term because they're not, they're either going to be harder to structure or they're just not worth it from an adverse outcomes standpoint. So just to complicate that further, I think looking at specific data elements, or specific kinds of data, in the context of whether they could be effectively move from the unstructured to the structured with an eye on helping adverse events would be also a very useful thing to do.
>>
Right.
>> Kelly Cronin:
So say we did end up and in fleshing out what a pilot might look out for adverse drug event detection, if there were some definition around priority adverse drug events in addition to the terminologies that would be most commonly associated with those so it could be nausea and vomiting or rhabdomyolysis or --
>> John Loonsk:
Simply allergies, having allergies in structured form, will be helpful to adverse events detection --
>> Kelly Cronin:
Even more globally than.
>> John Loonsk:
Just it's a, just a simple example.
>> Kelly Cronin:
Right.
>> John Loonsk:
And I don't know, that's sort of a meta-analysis of the different approaches, the different pilot approaches with an eye toward what would likely be desirable to do with text mining for some time to come, and what would potentially be responsive to being well structured in the context of workflow and accomplishing specific outcomes.
>> Kelly Cronin:
So -- and I think it might also be helpful to talk offline more. Do we want to be looking specifically at LOINC or certain parts of SNOMED that would be more feasible and how might we do that evaluation?
>> John Loonsk:
Yeah, I think the starting point is to look at sort of a coarse level at what the data are. And what the, it's like the quality activity. You know, you have to look at what the measure is in a general way, start to drill down on it. And then there's also, you know, to be honest, there's a process there that looks at surrogates for it, because you may at times, and this is prominent I think in the quality area, the data you get from an EHR may be a surrogate for the data you actually would like, really want, but it may be okay and good enough, and more accessible. And that's one of the complicating factors, but starting at a more granular way of looking at what are the data you would want and you know, obviously a lot of that has been done. But then --
>> John Lumpkin:
So let me see if I can try to pull this piece together with our recommendations. It seems to me that we have one recommendation to do pilots, and these pilots would look at one of three things. They would look at the data at the RHIO level. The second is it would look at it -- and that's kinds of what we have in 2.2. We could make that sort of a sub under the pilot recommendation. And then 2.3 would be looking at it at the EHR level. 2.4 would then be looking at it in looking at a data mining environment. So we've got it looking at supporting pilots and that ties in to the issue of resources. If there are resources to do pilots, these are the three areas that we want to do. And then the issue I heard John just raise is that at the same time, we need to identify where, where we can structurize, if I can use that word, the use of data elements that are important to be able to determine quality or to determine adverse events.
>>
Right.
>>
That's my understanding. Does that sound right, John? Yeah, I mean the way I would frame it is right now it's easier to look for your keys under the lamppost and what we want to do is make sure in the future people only lose their keys under the lamppost.
>> Kelly Cronin:
Yeah, the tricky part is what's your core set of adverse drug events or what are you are you going to start with? Just like AQA and HQA, went through an enormous amount of work to figure out what are the implementable measures. It puts some burden on the FDA or another organization to identify what are the adverse drug events of interest that we want to start with.
>> John Lumpkin:
Right. And then to structurize from there.
>> John Loonsk:
Since we're on a roll, let me introduce one further concept, which is the data triggers, the possibility exists and this is being looked at in quality as well, of where a data trigger can initiate augmentation of data, or information. And so you may have some indication that something, an event has occurred, but you need supplemental information to record it and that's an augmenting of the data.
>> Kelly Cronin:
I would think in particular looking at the MedWatch form and getting some of the epidemiologists’ input on the determination of causality and what's needed to make the determination on the causal relationship would be in particular important.
>> John Loonsk:
Or trigger. I mean, in terms of you won't necessarily have a full causality. But you may have an indication that something may have occurred to initiate a further investigation and that some of the technologies here are to potentially initiate that at the EHR level and that's partly what, you know, is being worked on in other settings.
>> John Lumpkin:
Not to disagree with you, but perhaps if I could suggest that we would enhance the term methods for employing data triggers. The implication that we originally started off was you have a trigger that says you have an adverse event. But with the enhancement, I think I heard John say, was that it may be that trigger will initiate further data collection to then determine if there's an adverse event. Or it may, because of itself, let's say a trigger says they did surgery on the left leg and the lesion was on the right leg, could, they know will be identified as an adverse event.
>>
John, you're correct in that a trigger only is an alert that something may occur or could potentially occur. It's not an adverse event itself. So it just leads to, a trigger brings an alert to do the further investigation. And it may be that an event has occurred and it may be that an event is about to occur or could potentially occur.
>> John Lumpkin:
Or there may just need to be more information.
>>
Correct, yes.
>> John Loonsk:
So it's really to detect, it could be --
>>
Potential.
>> John Loonsk:
Potential or -- yeah. I mean, or --
>>
And you can detect them prospectively, real-time or --
>> John Loonsk:
I would make it very specific in the language to make sure that it's inclusive of both because I think both are important.
>> John Lumpkin:
Right, so we can enhance that.
>>
Uh-huh.
>>
There may be varying thresholds. There may be a threshold for making a phone call, a threshold for sending a team out, and I suppose at some point in the future the data may be good enough that you have a threshold that you may want to suspend distribution of a drug based on that. I don't think we're at a point where we have any data that's good enough to do that now, but theoretically you might have good enough data in the future that you might be able to act on it that definitively.
>> John Lumpkin:
Okay. Have we hacked at these enough?
[laughter]
I think we can move on. So the first one under that, is to harmonize standards for adverse event reporting between ICSR and VAERS, and that work is being coordinated by AHRQ.
>> Amy Helwig:
Well, AHRQ is not directly coordinating harmonization between ICSR and VAERS. We are coordinating the development of the common formats to be used by the patient safety organizations and we are working with FDA and they are bringing their knowledge to the systems that they are using.
>>
I should say, ICSR, and VAERS are kind of pre-harmonized. You can think of the ICSR as a subset, a type of subset of the VAERS, but they were developed in tandem so I guess I'm not sure what you mean by harmonize.
>> Amy Helwig:
You may want to take out that, just change number one ti say recommendation to harmonize standards for AE reporting and then remove that hyphen ICSR.
>> Kelly Cronin:
This is in no way meant to be anywhere close to final language. These are just concepts at this point so I don't think we need to be, we'll have lots of opportunities to refine language in the next couple of months.
>>
And I think originally the question was whether as so many of the standards have gone through HITSP for harmonization if there were adverse events relating standards that needed to follow that same path.
>> Amy Helwig:
Well, the ICSR has been through HL7 and you know, I understand you guys are way backed up on use cases, but a use case to get the ICSR blessed by HITSP would be wonderful for us. I hadn't suggested it because I understand from many talks John has given that the on-deck circle for this use case is fairly jam-packed, but we certainly would be delighted to see something like that.
>> Kelly Cronin:
Well, I think, you know, we can revisit our ranking during the next priority setting process, and I think we actually didn't specify much around the adverse event reporting priority when we advanced it last time. So we may want to think about how to specify what's already out there, what could be built upon, like the ICSR.
>> Amy Helwig:
And one thing I'd like to hear a little bit about, maybe not at this meeting but at some meeting, is what other levers the Office of the National Coordinator might have to encourage a broad use of the ICSR, now that it's been finally HITSP-balloted. I've lost track of everything that you guys can do.
>> Kelly Cronin:
Yeah, I mean we're really relying on these primary mechanisms, but we can talk more about it.
>> Amy Helwig:
Yeah, through the RHIOs or whatever. Maybe the answer is nothing. But I know I'm not sufficiently knowledgeable about that to make a good judgment.
>>
Regarding the third subpoint under 3. John and Kelly, I would like to just work offline with FDA to, since this is such a very early stage to craft, to explore that further.
>> Kelly Cronin:
Sure. Yeah, it's actually something that Lisa and I talked about but hadn't had a chance to talk to you about it. We can change it however you feel we need to or is necessary, or strike it.
>> John Lumpkin:
Okay, anything else under 3.0? I'm detecting some harmony.
>> Kelly Croninn:
I think we just need to be a little bit reality-based with maybe not doing anything new and bold if we can't have a use case advanced in this area, with the exception of the third bullet, perhaps. I mean, certainly AHRQ's work is very important and we can support it in our recommendation, but we'll have to see how we fare in the next ranking process, I imagine.
>> John Loonsk:
Well, if I can, I think there's a role in these recommendations as to try to ensure that some of these priorities are indeed in the next ranking process. So clearly that's, this is not harmful in doing that. I think the other thing that potentially is helpful is if one can tease out the elements of it that have broad use outside of just this particular domain, and if it's just a vocabulary standardization, or harmonization, that's not as easy to do. But sometimes it's, like allergies, identifying the need for allergies to be standardized is helpful in many different domains. But teasing those things out and anything the working group can do to indicate where that would have synergy in other areas would be helpful in the prioritization process as well.
>> John Lumpkin:
Although we do have to be careful that we do have case reporting that we want to be in the queue also.
>>
Yeah.
>>
Right.
>> John Lumpkin:
Okay, any other comments on 3.0? Move on to 3.1, which was formerly 2.1, which is to support AHRQ's efforts to implement the provision of the Patient Safety and Quality Improvement Act. And whoever is working at a computer, thank you for being so fast.
>>
I don't know if you need to add anything else there other than leaving it as a stands.
>> John Lumpkin:
Yeah, I think that Lisa, I think it would be helpful though to flesh that out just a little bit. What are we supporting AHRQ in doing?
>>
I think you know, probably just add something of the key elements, specifying what the key elements are of the Patient Safety Improvement Act.
>> John Lumpkin:
Exactly.
>>
And I can easily pull together several lines.
>> John Lumpkin:
Great.
>>
Just thinking, encouraging voluntary and confidential reporting of events as well as supporting the creation of patient safety organizations to collect, aggregate, and analyze confidential information. So we can flesh that out.
>> John Lumpkin:
Great. Anything else on what used to be 2.1 and is now 3.1? Okay, let's move on to 3.2. And this is looking at those States where medical error reporting is currently mandated by law. And this is, this may be where the ASTHO work will fit in. It wasn't a real firm statement. Is someone from ASTHO still on?
>> Sunanda McGarvey:
John, I know they had said that they had a previous commitment and would have to drop off at around 2:00, 2:30. So I think we may have to get back with them on that.
>> John Lumpkin:
Oh, good, we can say they're going to help us do this.
>>
The recommendation is that AHIC should, AHIC should facilitate reporting in States, is that what the recommendation is?
>>
I think it's not clear.
>> John Lumpkin:
I think what we want to do here is around the issue of encouraging States that are collecting data related to State mandates. So the first step is to try to take a national view on what States are requiring. And then look for opportunities to standardize the, when they ask for certain content, to look at opportunity to standardize the format and the standards of that data collection. So going back to John's comments about drug allergies. If they're requiring to collect data on drug allergies that we begin to work on having an agreement on what that means.
>>
So does this parallel, then, in some sense what we've got in the letter about notifiable disease reporting, that at a much earlier stage, rather than wait until there's a 50 different systems, that at an earlier stage we want to try and for want an of better word, standardize, or find common elements in adverse event reporting in those States where it's mandated? Is that what the recommendation is?
>> John Lumpkin:
I think that there's a strong parallel without there being an organization like CSTE.
>>
Right.
>> Kelly Cronin:
So it could be that JACHO list of “never events”, I mean that's commonly in State laws.
>> John Lumpkin:
It's really encouraging that there be the development of essentially a data consortium, a consortium among the States that are doing it, to do it in a consistent way. And that the federal government would be in an ideal position to do that, probably because of the kinds of reporting it would be AHRQ and ASTHO, I think would be key players, perhaps the State Medicaid. You know, I think we just need to look for ASTHO to give us some ideas on who are the players we need to include so we're fairly inclusive of the agencies, whether it's the State data organization, so the NAHDO, or whatever their new name is.
>>
So John, does this become one of those recommendations where we say HSS in combination with ASTHO should -- ?
>> John Lumpkin:
Yes.
>>
Okay.
>> John Lumpkin:
Any other thoughts on this one? I think we ran out of recommendations. Any recommendations that people want to propose that haven't been included? Have we missed anything? Okay. I think this has been very good session. I think we're starting, we've made some good progress. Just to remind us that we are, we do have a working group. I would suggest that as we get another iteration of that, we'll have our sub-working group taking a look at this so that we can, if we get an opportunity on the 29th, look at a more, a somewhat more refined document.
Okay. Our next item is the hearing that we're going to have on the 29th. We all got a copy of the, this is on response management, unless we have other items. Anything else on adverse events? Going, going gone. Okay.
On response management, we have a draft outline for the testimony that breaks it up into basically into four areas. Outbreak and event management, countermeasure allocation distribution and administration, long-term follow-up, and communications slash alerting.
I'd like to suggest that we try to keep it down to three panels. Just looking at the timeframe, which might have a small potential to give us at least a half an hour at the end to have some discussion. And in my looking at this, the comments that didn't make it before the meeting, is that one area that's not addressed in response management are the non-public health spheres of response. In other words, Department of Homeland Security, Red Cross, State emergency management agencies that may also be looking at development of data systems, for instance I think it's called the CAN network done by the Red Cross and how they would interface with what we're proposing. So that's sort of cross-agency kind of coordination, might be another area to look at.
Other thoughts?
>>
Could I just ask about scope? Countermeasure allocation looks good, and seems to be dealing with large-scale outbreaks of diseases that have a relatively short incubation period, talking about flu and small pox. Number one, outbreak and event management looks broader than that. I'm particularly looking at TB, which you know, Farzad, or Art, or whoever’s on the line may disagree with me but in that context looks to me a little bit like which of these things is not like the other. Is this really about emergency preparedness in the sense of large-scale outbreaks, secondary transmission, broad exposure to a common source, you know, anthrax in an office building or a dirty bomb. Is that the focus that we're on?
>>
Yes.
>>
I think so.
>> John Loonsk:
This is John Loonsk. The thing that has come up in this context relative to sort of, for example, TB, I assume you're wondering about in the context of it being a reportable disease and maybe, we've had in this working group, discussion of reportable diseases. In experience, the information needs, even of an outbreak, even of a reportable disease, when it reaches some threshold, are very different, and at times are not supported by the more traditional reporting structure. Cases in general are, you know, one sort of coarse way of describing that event is an even coarser description because it can include a variety of different activities.
The, much of the notifiable disease reporting is sort of based on confirmed cases, much of the needs of the large outbreak are on possible cases. Information flows at times are, you know, turned around or, you know, adherence to some of the dynamics of what the case definition is, you may not have a case definition. All those things, you know, I think are attractive in terms of putting them on the table because they do relate to the specific emergency responses, certainly that I've seen in, it terms of what the information needs are.
>>
Sure. I think that gets to the point. It’s a question of, TB outbreak is significant. It's an emergency, it's not an emergency in the clinical, it's an emergency in the public health sense but not an emergency in the clinical sense because of the incubation period, the timeframes involved. It's important to identify those folks, but very different than a small pox or an influenza pandemic.
>> John Loonsk:
I think the thrust of that really was not TB. So maybe a different example would be better. But where the needs of contact tracing, specifically, and those needs particularly that are prominent to local and state level in terms of so many responses. But you're right, TB may not be the best example of that because of the timeframes.
>>
Is Farzad or Art still on the phone?
>> Farzad Mostashari:
This is Farzad, but I am going to have to leave momentarily.
>>
Okay. I was just curious whether you had any insight into that, focusing the testimony.
>> Farzad Mostashari:
My two cents is that probably best to focus on the actions that happen rather than the investigation piece. Because that is, there really are two separate parts.
>> John Loonsk:
Yeah, I guess I would consider managing the outbreak to be some of the actions that happen that we really haven't talked about in the context of biosurveillance or in the context of those discussions we did have around notifiable disease reporting.
>> John Lumpkin:
But will we be covering those in our bidirectional?
>>
I'm not sure what you mean, John.
>> John Lumpkin:
Well, for instance, in our bidirectional discussion, at least as I would envision it, we would talk, about okay, we've got an outbreak, now how do we send stuff out to the practicing clinicians on treatment, on how to define case, changing case definitions over time as the characteristic of the outbreak is being resolved. And that kind of takes up some of the stuff in number 1.
>>
Well --
>> John Lumpkin:
As opposed to what Farzad was talking about, what are the actions, how do we facilitate the actions that need to be taken when you have a population at risk.
>> Farzad Mostashari:
The kinds of things -- this is Farzad -- the things I was thinking was resource management, you know, how do we get vaccine out to people, how to we document who got what vaccine versus who needs the supply, those, the resource monitoring.
>> John Loonsk:
So this is John. I would say that what Farzad is alluding to is the broad category and perhaps it's not completely articulated but the broad category in Roman numeral two.
>>
Right.
>> John Loonsk:
What John is alluding to is mostly in the broad category of Roman numeral four, although you also touched on like the dissemination of case definition, which -- but I would return to Roman numeral two as also being important in the context of in some very bread and butter-like activities that information systems are really helpful at, you know, in SARS, in other places where you have a highly communicable disease and what do you do to manage that and how can the systems help and what the needs in that regard.
>>
So am I understanding, then, that 1 is sort of case identification, or either cases or even potential cases, and 2 is then what I think Farzad was talking about, is the response once you've identified and that's an ongoing process, identifying spread, identifying the cases, and patients at risk, then 2 is what you do with that information.
>> John Loonsk:
And how you relate that to your, to the other response, counter measure activities. But certainly one of the response activities is potentially understanding spread. Some of those are fuzzy lines.
>>
Absolutely.
>> Paula Soper:
This is Paula from NACCHO. I think if we wanted to look at the three areas as John suggested, doing three areas that probably would have the biggest impact on not just emergency preparedness, but also for day-to-day public health activities would probably be 1, 2, and 4. I think long-term follow-up is absolutely critical, but maybe not to tackle for this day because I think we could really go a long way in these three Roman numerals, one, two, and four toward not just emergency preparedness but improving the flow of data for day-to-day public health activities as well.
>> John Loonsk:
So we could sort of follow up on that in the longer term, then?
>> Paula Soper:
Yeah. Yeah, exactly.
[laughter]
>> Farzad Mostashari:
This is Farzad. I would totally support including 1, 2, and 4, on 1 I think the focus on things like contact tracing --
>>
Yeah.
>> Farzad Mostashari:
-- as opposed to some of the continuing along the situational awareness kind of general trending, et cetera, stuff, but getting down to brass tacks in terms of the investigation, as a way to draw the line.
>>
I would support that, Farzad.
>> John Lumpkin:
So I'm hearing 1, 2, and 4, as being our focus.
>> John Loonsk:
I think one of the things that I would -- I understand Farzad's desire to be focused in 1. I think one of the reasons that there's some nebulousness to the definition is what the case or event is has varied highly depending upon the type of emergency. So if you look at like Katrina as an emergency, the information needed about people in shelters would not classically be a possible case. But there were definitely, there were data needs there, around numbers of people in shelters and where they were and then eventually about communicable diseases in those shelters, but if you're looking at an environmental event, there may be other data as well. And clearly you wouldn't want it to spin out of control, but I think it is worthwhile to at least have some discussion of the breadth of possibilities of data that are needed so that we make sure that we're not overly focused on cases as the element of data.
>> John Lumpkin:
Well, I think what I'm hearing on number 1, and let me see if this makes sense for three speakers, try not to have too many speakers. One speaker to talk about what was needed in an event like Katrina. Perhaps one speaker who may be from the State level who did Top Off, and what was their experience and what was the kind of data they felt they needed after doing one of the Top Off events. And then the third being a pure, and I think we can perhaps look to CDC, of an example of maybe a food-borne illness?
>> John Loonsk:
I think the trouble with that is that the information needs to vary at the different levels. And I think we may not be able to accomplish that all. It would almost be better if we arrayed a list of issues can then addressed them at the federal, state, and local level, so that we could be inclusive.
>> Farzad Mostashari:
This is Farzad. The other way to structure this would be to take several case examples whether it's a community meningitis outbreak, or Katrina, or 9/11 for that matter, and talk through the different components of the response management, everything from the contact tracing to the countermeasure distribution, the long-term follow-up and communications. But, you know, doing, going through and saying how in each stage the different needs are, instead of segmenting it the other way.
>> John Lumpkin:
Let me just play with that idea. We could have three panels, one on Katrina, one on Top Off and one on a food-borne illness or whatever we decide --
>> John Loonsk:
I think that's too narrow, John.
>>
Yeah.
>> John Loonsk:
It's just not going to be inclusive of the different types of data that are necessary.
>> John Lumpkin:
Okay, well, let me just sort of follow through the idea and then we can, and then the thought would be that each one of the panels would have been from the federal, state and local level. Is there a way to broaden out that idea, or do you, or what would you suggest? Other thoughts on trying to arrange the three panels?
>> John Loonsk:
Well, maybe we could come up with a list of sort of prototype, a list of the scenarios that we could, without getting overly big, we could try to make sure it's sort of inclusive of some needs at different levels and work from there. And maybe it's not far from having a select list and then having it being discussed through at each of the different levels.
>> John Lumpkin:
Would you see the levels, would you see arranging the panels by level or arranging the panels by scenario?
>> Farzad Mostashair:
This is Farzad. I'm, I can't quite visualize it in my mind, but sometimes it's better not to overly emphasize the, particularly here, the levels issue. You know, drawing bright lines between the needs at the different levels may not work in this context as well.
>>
I want to clarify Farzad, I'm still getting a feel for apples and oranges, and maybe we're still in the fruit store, but Katrina or Top Off, meningococcal outbreak or TB, I mean, tend to be of limited scope in terms, I mean certainly more, meningococcal is certainly more acute, but the scope is more limited than a Katrina or Top Off or small pox exercise, I appreciate the fact that the data needs are the same, or, let me say, the data needs may be similar in what you want to find out about an individual patient, in terms of extracting data from a clinical record. I just want to be sure that what we want is in fact the whole fruit stand, or whether we just want the citrus fruits. Are we just looking for the big-scale emergency where, you know, we need to put the pedal to the metal and put people in every, replace the system, Farzad, where we’re putting somebody in every ER?
>> Farzad Mostashari:
I understand completely. You know, this is Farzad. My thinking, it would be helpful for the people who don't have the experience with this on the real level to get a sense of what are the data needs and what is, you know, people, I heard from the state Governor's Association that the legislators couldn't believe that we didn't have a way of polling providers during the flu shortage, of who has how much flu vaccine and how much do you need. So giving some concrete examples of, we had a community meningitis outbreak for example in New York City, we had to do contact tracing, we didn't have the right tools for the job of being able to do contact tracing and managing the relationships there. We then had to, you know, get vaccine out to people, we had no way of reaching the clinical providers to identify their drug users in this case, drug-using patients. So we had to set up shop in methadone clinics and health centers. We couldn't, you know, we had difficulty keeping track of who got administered what. We don't have a way of tracking in the long-term if there are any adverse events from those vaccinations. So that's kinds of taking a story and walking throughout data needs at each step to illustrate what the challenge is.
>>
And quite frankly, a large scale meningococcal outbreak, you can be immunizing or prophylaxing eighty or a hundred thousand people.
>> Farzad Mostashari:
There wasn't enough vaccine in the country to do that --
>>
It can be huge. It may not be nationwide, it can be multi-state and it can be pretty darn big.
>>
Fair enough. Fair enough.
>> John Loonsk:
If I can make a suggestion. I think it would be useful to put together a list of, with a couple of attributes to make that sort of point clear, of the kind of scenario to walk through and then let's just make sure that we have it adequately articulated and then we could have, that would be one helpful step in moving forward.
>>
I'm wondering if it might be, this is just another way of kind of dicing this, but if it might be instructive for people to hear what the needs would be in a smaller-scale something, such as a meningococcal, and then as things escalate.
>> Farzad Mostashari:
This is Farzad. I have to sign off. Now but I think that's a good point. At least a lot of state and local folks have, you have to have systems in use every day with the small- and the medium-size. You can't just pick things up when the big one hits. And I'm not volunteering her for it, but I think Marcie Layton would do a wonderful job talking about the community mening outbreak or West Nile, or any of the other -- West Nile is a good one too -- any of the other experiences we’ve had. I have to sign off now, thanks so much.
>>
I think the other benefit of doing it that way with kind of a escalating incident is that you also do get to the issue of what happens when you need, when local needs to bring in state and state needs to bring in feds. What the different data needs might be.
>>
So it sounds like Farzad volunteered Marcie.
>>
Well, she's fabulous, so, yeah, I would certainly second that.
>> John Lumpkin:
So I think what I'm hearing is that we would do one panel on what do you need to do in order to manage an ongoing outbreak. And then perhaps a second panel on when you’ve determine that you have to do an intervention, how do you manage the resources.
>>
Uh-huh.
>> John Lumpkin:
And then third would be to look at, I'll call it horizontal and vertical communication. So up and down on the public health chain of command, but then also horizontally, with other players, whether that be hospitals or emergency response organizations.
>> Leah Devlin:
This is Leah Devlin. I'm sorry, I've come and gone a little bit on the call. But there were some really neat lessons and important lessons learned from SARS. I don't think I’ve heard you all talk much about that. But it was certainly brought in, John, what you're saying, the hospitals, the doctors, we got involved at universities, a work site. And there's so many lessons we've learned from Canada. That was another scenario you could look to, and did you all talk about the mumps outbreak, too? That was 800 cases, in 19 states and it was flown in? I mean --
>>
That's another one.
>> Leah Devlin:
In SARS, we got into mass quarantine, that why it was sort of interesting for us here. Just a thought.
>> John Loonsk:
Yeah, I mean as we start to do this, each significant event has its own, a little bit of its own flavor and they each exercise different aspects of the system. If you think about anthrax and the lab results that were so prominent there. SARS and isolation and quarantine. Something --
>> Leah Devlin:
A new emerging infection. Something that we had to start from scratch with.
>> John Loonsk:
The other, you know, context for SARS was syndrome complex that was, the initial detection was very sketchy at the outset.
>> John Lumpkin:
It's dawning on me that I think for what we're trying to accomplish, we're not going to be able to do it in just one day of hearing. That we should try to put together the hearing to get us, as many of us on the sort of the same page. I think that we need to have an opportunity, perhaps, by conference call for the Workgroup to have a discussion about where we want to go on this, and then perhaps follow-up with another set of hearings. Because as we begin to focus on where we think the needs, are resource management, are we, I think our, I would find it difficult to think that we would be, after one day of hearing, be able to make recommendations along the lines that we've just done.
>>
Will they give us more time?
>> John Lumpkin:
I'm sorry?
>>
Will they give us more time?
>> John Lumpkin:
You mean AHIC?
>>
Uh-huh.
>> Kelly Cronin:
Sure. There's no real deadlines here. It's really just as fast as we think we can move.
>>
Okay.
>> John Lumpkin:
So I think on this one, if we think about moving that a little bit, that we do adverse event at the next, at the July AHIC. Or the June or July?
>> Kelly Cronin:
We were thinking July would be a better target date if we're going to package both of these recommendations together.
>> John Lumpkin:
Right, but I'm thinking of staggering them. Adverse event at the next AHIC, and then the one after that we would do response management.
>>
Yeah.
>>
I think we could do adverse events then in June. I think we were pushing back to July thinking they would both go together.
>> John Lumpkin:
Right.
>>
We can look at the schedule.
>>
You know, it dawns on me that one area that I don't think we've tapped into, and I think we need to, is the government and sector coordinating councils for health and public health, that are jointly chaired by HHS and DHS, because they have looked at a lot of the resource management issue, especially data going back and forth between public health, healthcare, and the vendors.
>>
Are you talking about things like the National Disaster Medical System?
>>
No, no.
>>
No, that's --
>>
No, this is a government -- there are sector coordinating councils and government coordinating councils for each area of the, each of the HSPDs, I can't remember what that means anymore, out of preparedness for six months and that's what happens. But there are different sector coordinating councils and healthcare has its own coordinating council. There are members of all of the HHS agencies as well as the other federal, state, and local representatives on the government coordinating council, and then representatives from healthcare and industry, clinical care and industry on the sector coordinating councils. And they are tackling issues. They're focusing quite a bit on the infrastructure protection plan, but they have had meetings specifically about how do we, how does information flow for resource management during an outbreak. And I heard some powerful examples of messes that were created because of Katrina, because things were not in place beforehand, information systems were not in place beforehand to allow the information to flow easily.
>> John Lumpkin:
So I think what we can do is perhaps we've gotten some recommendations for discussing some of these events and the kind of information we need, potentially looking at -- excuse me -- looking at the meningitis episode in New York City. I think mumps would be a good one, multi-state response. Looking at perhaps a non-medical episode like Katrina that had significant medical implications. Then -- that would be in relationship to Roman numeral 1. Roman numeral 2 we would want to perhaps invite a speaker from that agency council that was just being discussed.
>> John Loonsk:
John, I just have to add that I do think that number 1 is important enough that we should make sure that we're inclusive in terms of the different types of events.
>> John Lumpkin:
Something else that I left out.
>> John Loonsk:
There are a number, I think there are a number of things we could bring to bear and maybe we can do that offline in terms of developing a list. But obviously we can't be exhaustive, but we have several different aspects that are being considered. One is size was described here. Now, one way of dealing with size is to start small and go big and expect everyone to talk about that full spectrum. I think that's a reasonable thing, but you need to get the full spectrum in.
>> John Lumpkin:
Right.
>> John Loonsk:
And then that obviously would have implications at different levels as well. So if it's inside of a jurisdiction, if it's cross-jurisdictional, if it's, you could take that as far as you wanted. But then the breadth of the different events I think is worth exploring a little as well. There actually were several different Top Off events with several different characteristics. That's not a bad place to start in terms of thinking about them because there had been a fair amount of thought going into what was involved in each of those. And then I'd also do think it's important to consider some of the recent emergencies we've had and to build on them. So SARS was mentioned, anthrax was mentioned.
>> John Lumpkin:
Well, so the question, John, that I would have in response to that, because I mean it's not that I disagree with you, is that what we need are synthesizers. So someone who has had experience with multiple different kinds of events that could give us their sort of here are the lessons I've learned and would tick down maybe five or six different lessons, that maybe one lesson from SARS and one from Top Off and so forth, rather than trying to use a scenario to have someone discuss the scenario and then say what the lessons are I learned from the one individual scenario.
>> Leah Devlin:
You know, I think from one of our conversations we were talking about having, and this may be hard to do, too but having a federal, state, and local level speaker that could talk about their breadth of activities and it's going to be hard to find people that have done that. But we were talking about a state preparedness director that would have been involved in several activities.
>>
Or a --
>>
Or a state (inaudible).
>> Leah Devlin:
Whatever the right level is. That could say, my experience covers these five events and these were the different needs and compare and contrast across them, the data needs at their level.
>> John Loonsk:
We could also ask them to collect information from others, obviously, and not have to have something they personally were involved in.
>>
Right.
>> John Lumpkin:
Between floods along the Mississippi, hurricanes, earthquakes, and various outbreaks, I think we should be able to find someone from a different, a number of different perspectives, federal, state, and local, who had experience on multiple events. Or at least will have, be able to synthesize discussions with colleagues who have been in multiple events. And I suspect with assistance of ASTHO and CSTE, we should be able to find a state person. Certainly there ought to be someone from CDC who could do that. Perhaps, you know, looking at it across, you know, perhaps a representative from NACCHO. And just begin to look at that aspect of number 1, will give us a framework.
>>
I would say definitely NACCHO. All emergencies are local.
>> John Lumpkin:
It's just that --
>>
I know, we have to get that plug in there every time, though, John.
>> John Lumpkin:
Just a question of there are some cities that may be more, have been more challenged than others. New York City was with both September 11th and meningitis and then --
>>
West Nile.
>> John Lumpkin:
So --
>>
Yeah.
>> John Lumpkin:
So we have some examples.
>>
Yeah.
>> John Lumpkin:
Does that sound reasonable for the first panel? A way to flesh that out?
>> John Loonsk:
Was that sort of focused on data needs?
>> John Lumpkin:
Yes.
>> John Loonsk:
And then the things that wouldn't immediately necessarily fit into that would be sort of capabilities like contact tracing and quarantine management and which are important but, and we could address them either by adding them, and also these, or put them somewhere else in this structure.
>> John Lumpkin:
Right. I think then moving on to number 2, because we can think about maybe moving some of those into number 2. In number 2 it seems to me there is a, there are a couple of things that need to be done as part of resource management. One is how do we get biological agents, biologically active agents out to people who need them? And you know, again, I have to believe that there are people, certainly there are people who do that every day with vaccines, as part of the VFC program, with significant experiences in what are the challenges related to getting a vaccine from a central storehouse to a doc who is going to deliver it at a clinic. And all the challenges particularly if you're dealing with something like varicella, which requires a cold chain and so forth.
>>
Let me just ask, in terms of the focus of the working group, are those logistical issues the key for us, or is it the data needs issues such as who needs the vaccine, who has got it, where it is commercially? I appreciate the comment about cold chain but is that on our plate, the logistical issues of delivery?
>> John Lumpkin:
Many states when they deliver their immunization registry created a significant resource management backbone in order to distribute those vaccines.
>>
And a lot of them have built on that for resource management for emergency preparedness as well.
>>
Right, that was my point.
>> John Loonsk:
I think Steve's point is well taken, though, in that that there is a focus here which is a section of what you alluded to, John. And I also think there's some experiences here that we can draw to in a scenario-drawn way as well that relate to this a la the flu shortage and the small pox vaccination program and then obviously more routine programs as well. But that each of those, there were actually information needs and lessons learned from those, like considering how much of the vaccine supply is actually in the retail supply chain in any given time and how you have to get access to information about that supply in the context of considering re-appropriation.
>> John Lumpkin:
I think that there are a number of different areas. My point is that there are systems that are in place, some of which are being used on a daily basis, that need to be thought about, because we're going to need to rely upon those. People tend in a disaster to use what they're familiar with. In addition to the Vaccines for Children program, there's also statewide EMS systems that are also looking at resource allocation. Now, by pointing this out, none of these systems currently work together. And I think that's where we can add some value.
>> John Loonsk:
Yeah, I think that, you know, the approach suggested here is oriented to needs. It is certainly important to also address what pieces are in place and what, where there may be gaps --
>> John Lumpkin:
Right.
>> John Loonsk:
-- thereafter.
>>
It also may get to your point, John, about bidirectional communication. I think that's part of the systems that you're talking about are matching needs with available supply. Isn't that right?
>> John Lumpkin:
That's a major challenge that will be on the plate of the public health system. And related to our charge which is looking at the interfaces, much of the resources which the public health system will be trying to allocate will be the in the domain of the clinical care system. And so we have to look, not only looking at the interchange of who is sick, but who has got a bed, who has got a ventilator, who has whatever.
>>
And who has the Ciprofloxacin and the alternative drugs, right?
>> John Lumpkin:
Exactly. And that gets to John's point about what's in the supply chain. So the thought being there is that we would want to think, we would want to have a combination of people who are doing it, and people who are thinking about how to do it in those kind of events.
>> Paula Soper:
I think Houston has a very good system that was in place before Katrina.
>>
Did you say Houston?
>> Paula Soper:
Houston.
>> John Lumpkin:
System to do what?
>>
Is that Paula speaking?
>> Paula Soper:
Yeah, it's Paula. A system for all of the resource tracking within the healthcare system between public health and the hospitals for beds and medications and ventilators and whatnot.
>> John Loonsk:
Yeah, I don't think we have to go down the bed route because we really did cover that in the --
>>
True, true.
>> John Loonsk:
Actually, you know, having just come back from HIMSS, I can recommend to you that there was really some compelling presentations there that showed the implementation of some of the utilization data and the biosurveillance data in the context of clinical, EHRs in clinical systems. So I just, it's, we have to figure out a way to get people to see some of that stuff because it really took the original biosurveillance use case data accumulation and resource utilization and it's implemented in a number of major clinical systems now. And I think it would be great for the group to see that so that, both as it’s tangible, but also it’s rewarding from the standpoint of how some of this is not as long-term as one might otherwise, you know.
>> John Lumpkin:
Well, playing off of that, could we think about perhaps starting off that second panel on countermeasure allocation distribution and administration, with a basically with a demonstration of that? Or --
>> John Loonsk:
I was really thinking it might be a subsequent activity because it's really more oriented to the previous use cases, around biosurveillance and bed utilization, particularly. I think it would be a good thing for the Workgroup to get exposure to, but it's not exactly on this countermeasure allocation and distribution, which I think has a lot of other implications.
>> John Lumpkin:
Well, again, thinking not about what a system could do, but I think our charge of thinking about how the systems interoperate, if that's already in development then what would we think about in the interoperability of what we might recommend in number 2 with what we've already recommended?
>> John Loonsk:
Yeah, I'm not, I'm sorry if I've confused things a little. I think it would be good from a number of perspectives, including the tangibility, how tangible it is to demonstrate and also as a representation of biosurveillance, the original use case, and bed utilization data. For me, number 2 here is evocative of a number of different activities that are not completely on target with that. So I wouldn't append it with number 2, any more than I would append it with number 1, but I was just suggesting that it could be useful for the Workgroup to eventually see. It will also be a bit of a logistical challenge to do that. What it represented in the context of HIMSS was a variety of different major software, major and minor software vendors coming together and implementing their systems together in a way that will be, I'm not sure we can just replicate it for a working group. It's, you know, this is a once a year event for many of them. But I do think if we can figure out how to do it, a lot of people here would find it rewarding eventually.
>>
John, was that one of the typical HIMSS demos that they do in the convention hall?
>> John Loonsk:
Yeah, it was the interoperability showcase, which was --
>>
Okay, yeah.
>> John Loonsk:
-- bigger this year than it had been previously, and it had a major emphasis on the HITSP adoption and the HITSP standards, and so it actually included a demonstration of the biosurveillance use case, a number of different health information exchange concepts, it included a case report scenario for tuberculosis including augmenting data entry in association with this, so there were two, it was basically triggered by the, and this may be somewhat false, people, but triggered by the prescribing of two TB drugs initiating an electronic form in the EHR for submitting a full case report.
>>
Wouldn't that be fabulous?
>> John Loonsk:
And then the transmission of that. So that was there. And these are in real systems that were implementing this. So I'm sorry, I didn't mean to go overboard with it. But I was just very excited by the progress that was made around the existing use cases, and I think it's something that a lot of people would be interested in.
>>
John, you know, you had suggested lengthening the timeline for this and having a conference call to try to narrow or focus the scope. You know, with Marcie Layton or Leah Devlin or people offering, Laura made the comment about compare and contrast which I think was a good one, to look at various scenarios and which are the scenarios of choice for really highlighting the different things that we want to compare and contrast. Where are the data needs for emergencies, whether it's a Katrina or a meningitis or a small pox, where are those the same and where are those different? How do we create systems that have those common elements, but still have the flexibility for the areas that they're different? And maybe John, you're right in saying that we need that intermediate step, we need some kind of a call, or to bring some, before we called them witnesses, people together to help us think through what are the examples to use and that brings in the question on levels, I think, as well.
>>
Uh-huh.
>> John Lumpkin:
So maybe, I'm going to try a different approach to this. What if we were to sort of have the presentation almost as we've discussed before, with those individuals, and ask them to talk about not just 1 but also 2 and 4? Give them some extended time, and then use the rest of the day to brainstorm with those same individuals as experts. As we try to tease out what, where, what could or should be done in this area related to data needs.
>> John Loonsk:
My experience with 2 and 4 is that they very frequently involve different people as experts than 1.
>>
Okay.
>>
Agreed.
>> John Lumpkin:
Okay. I think we're --
>> John Loonsk:
I'm sorry.
>>
Sorry.
>>
It was a good idea.
>> John Lumpkin:
I'm just tossing ideas out to see if we can sort of crystallize something. I'm not wedded to anything I'm tossing out so far.
>>
Because really you're talking probably your state epi, for the first one, your SNS coordinators for the second one, and your health alert network coordinators for the third one, or the fourth.
>> John Lumpkin:
Okay.
>>
But I like the idea of having them there to brainstorm and I'm not sure how to include the two unless we do two days of testimony.
>> John Lumpkin:
Well, the problem with doing that just the mechanics of it, if you have somebody present in the morning, they're not likely to stay around for the second day of discussion.
>>
Right, no, I understand that.
>> John Lumpkin:
Yeah.
>>
If we, I don't know if we can get two days together anyway, but if wedded another testimony day, could we tackle the first item on the 29th and then add another day and try to get the two, the second and the fourth item on another day, recognizing we're pushing the recommendations out a month or two?
>> John Lumpkin:
I kind of like that idea.
>>
More time for discussion, which we recognized in the last grueling meeting we would have liked more discussion.
>>
Laura, do you mean another day in the following month? Or --
>>
Well, I don't know, we don't have a meeting time in April.
>> Kelly Cronin:
We could add one.
>>
Our next meeting is early in May. And I don't know if we'd want to add or we just want to take that next May meeting day as that second testimony day. And we could work that out in e-mail what preference is.
>>
Question. Is our next meeting during the HIT summit or right after?
>>
The next meeting is the 29th.
>>
I think it's the last day of the HIT summit, so people may already be in town.
>>
The other option is to maybe try and do a day and a half, then? I know that's Friday afterwards is a conflict for some because of spring break, but maybe we could do Wednesday afternoon and Thursday?
It may be too late to try and get those scheduled.
>>
And that overlaps more with the HIT conference, right?
>>
It would.
>>
Yeah, I think so.
>> John Loonsk:
Well, I think taking the scheduling offline, the idea of perhaps breaking it up by 1 and then 2 and 4, as a separate day or a half-day or whatever, would make sense.
>>
Well, maybe we add a meeting in April to do 2 and 4.
>> John Lumpkin:
Yeah, I think that if we do 1, with some good presentations and have, you know, a fairly extensive discussion, we will be well prepared to do 2 and 4.
>>
Right.
>> John Lumpkin:
So I think I like that idea.
>>
Okay.
>> John Lumpkin:
So why don't we play around with that. We can look for schedules for the second day offline. I think we've kind of identified some of the types of people we want to have speak. John Loonsk, I think you're going help with us that. I think I'd like to perhaps, even though we're focusing on 1, it would be nice, I think, to have somebody from the HS, DHS world maybe talk about their view of the same issue. I don't know if that would make sense on this.
>>
Coordinating council or --
>> John Lumpkin:
Or something like that.
>>Brian Carnes:
John, this is Brian. We've got the option of something like their chief medical officer or deputy who are both ER docs and I work with all the time.
>> John Lumpkin:
Jeff or somebody like that, who I think can see the view from both worlds. I think that would be really helpful.
>>
And then the nurse who is the co-chair of the coordinating council might also be a good alternate if one of them is not available.
>> John Lumpkin:
Good. And so I think we've got, we can put together a fairly decent panel with maybe four or five speakers, and then following up with a brainstorming session for three, four hours.
I think we have a plan. Are we comfortable with this point? Can we move on to the rest of the agenda? We've talked about the next meetings. I don't think we need to go over those. I'm assuming there are no other comments on the letter on case reporting. And so I think we're ready for public comment.
>> Judy Sparrow:
Great. Jen, can you bring the public in, please?
>> Jennifer Macellaro:
This is Jennifer. I've put the slide up. It has a phone number if anybody needs to call in. If people are already on the line, they can press star 1 to alert the operator, and there's an e-mail address if anybody would like to write in comments after the meeting. So we'll just wait a few minutes.
>> John Lumpkin:
Great. While we're waiting, I want to thank everyone for their participation. I myself am going on vacation next week.
>> Kelly Cronin:
Good for you, John. That's how you're getting off the AHIC meeting on the 13th.
>> John Lumpkin:
No, no. Unfortunately, I'll be back at work by then.
>>
Okay.
>> John Lumpkin:
Not unfortunately. I kind of like my job.
>> Brian:
This is Brian, I apologize for being in and out. I finally got my floor torn out and my pipes fixed and the leak stopped.
>> John Loonsk:
Congratulations.
>>
Now I can call the wallpaper and painting people.
[laughter]
>> John Lumpkin:
I was just about ready to say that.
>>
I was, too.
>> Brian:
It started off as a small remodeling job that just got more interesting.
>> John Lumpkin:
How old is your house?
>> Brian:
Twenty-five years.
>> John Lumpkin:
That's not supposed to happen.
>> Brian:
When you put a drill through a pipe, it's never good.
[laughter]
>> John Lumpkin:
Who was doing it?
>> Brain:
Fortunately, an insured and bonded contractor.
>> John Lumpkin:
I always hate it when I do it myself.
>>
You get a remodel for free.
>> Brian:
Yep.
>> Jennifer Macellaro:
This is Jennifer. It doesn't look like we have anybody calling in today.
>> John Lumpkin:
Great. Well, it gave us a great opportunity to catch up. So thank you all, we'll be in touch. Please send suggestions for speakers in so we can start putting the panel together. And with that, have a great weekend, everyone.
>>
Bye, now.
>>
Thank you.