American Health Information Community
Personalized Healthcare Workgroup Meeting #7
Friday, August 17, 2007
Disclaimer
The views expressed in written conference materials or publications and by speakers and moderators at HHS-sponsored conferences do not necessarily reflect the official policies of HHS; nor does mention of trade names, commercial practices, or organizations imply endorsement by the U.S. Government.
>> Judy Sparrow:
Great, thank you and welcome, everybody to the seventh meeting of the Personalized Healthcare Workgroup. Just a reminder that we're operating under the auspices of the Federal Advisory Committee Act, which means the meeting is being conducted in public and there will be an opportunity at the close of the meeting for the public to make comments. Let me just remind the members of the Workgroup to please speak clearly and distinctly, and identify yourself before you speak. And when you're not speaking, if you could please mute your telephones, it will reduce any background noise. Why don’t we begin here in the room and go around and introduce ourselves, and then Jennifer, you can introduce members on the telephone.
>> Alan Zuckerman:
I'm Alan Zuckerman, I'm a contractor with ONC assisting Workgroups.
>> Becky Fisher:
I'm Becky Fisher, and I'm a patient advocate.
>> Alan Guttmacher:
Alan Guttmacher, from the National Human Genome Research Institute at the NIH.
>> Mary Beth Bigley:
Mary Beth Bigley, with the Office of the Surgeon General.
>>Paul Cusenza:
Paul Cusenza, [inaudible] with 21 23andMe [inaudible].
>> Greg Downing:
Greg Downing, Office of the Secretary.
>> Kristin Brinner:
Kristin Brinner, Office of the Secretary.
>>
[inaudible]
>> Judy Sparrow:
Jennifer, if you could introduce those members on the telephone, please.
>> Jennifer Macellaro:
Sure. On the phone today we’ve got your co-chairs, John Glaser and Doug Henley. Brad Therrell from the Newborn Screening and Genetics Resource Center. Betsy Humphreys from the NIH. Joel Kupersmith from the VA. Beryl Crossley from Quest Diagnostics. Emory Fry from the DOD. Janet Warrington from Affymetrix. Steve Matteson from Pfizer. Steve Teutsch from Merck. Marc Williams from Intermountain Healthcare. Andrea Ferreira-Gonzalez from Virginia Commonwealth University. Anthony Achampong for Cheryl Austein from HRSA. Mark Rothstein from the University of Louisville. Did I miss anyone on the phone who is an alternate for a member today?
>> Marie Mann:
This is Marie Mann.
>> Jennifer Macellaro:
I'm sorry, I did miss you. Anyone else?
>> Catherine Kohler:
Hi, this is Katie Kohler from CDC.
>> Jennifer Macellaro:
Okay, thank you.
>> Philip Pochon:
This is Phil Pochon from CDISC.
>> Cathy Fomous:
And Cathy Fomous from OBA.
>> Chris Lamer:
Chris Lamer from Indian Health Service.
>> Dina Paltoo:
Dina Paltoo from NIH.
>> David Parker:
David Parker, Indian Health Service.
>> Judy Sparrow:
Is that it?
>> Jennifer Macellaro:
Okay.
>> Judy Sparrow:
Quite a crowd for a Friday afternoon. Let's turn it over now to doctors Glaser and Henley.
>> Doug Henley:
Thanks, Judy. And on behalf of myself and everybody else. Thanks for being here on a Friday afternoon. We'll try to be succinct today and stay certainly within the confines of our allotted times so folks can get off for other events this evening and/or for the weekend at the appropriate time. I would just like to reiterate Judy's earlier comments about please mute your speakerphone or your cell phone if you aren't speaking. And again please state your name as appropriate as we have our conversations during this call.
Our first item of business today is to look at and review the Thursday June 7th minutes of our previous telephone conference call. So I present those to the Workgroup for any additions, edits, deletions, corrections or comments. All right? Hearing none, then without objection, we will consider those approved. Thank you for that.
As you can see on our next item for agenda, we want to have a quick recap of the July 31st AHIC meeting, which you have as one of your attachments with the set of recommendations that were presented to the AHIC. I'm happy to report that Dr. Glaser did a wonderful job, along with Greg Downing and all these recommendations were approved by the Community. I must also add that they were received, I think, and I'll certainly invite John and Greg to make comment, as well, that they were received, I think, with a great deal of excitement. The Community felt that the recommendations were aggressive, they were on target, they were important. And the basic feedback was this is an important area of work relative to the issues of interoperability and health information technology that can move the country forward in an important way, to improve patient care and health outcomes. So let's get on with it, to the extent that later that afternoon they approved the use case that we had also presented to them without objection. I would point out that recommendations 1.0, 1.1, and 2.0 go to the AHIC explicitly; whereas the other recommendations 2.1, 2.2, and all the 3-point-whatever recommendations go to HHS for implementation. And we'll hear more about that from John and Greg here in just a moment. But again my sincere congratulations and thanks to all the members of the Workgroup for excellent work done, particularly the workgroups on genetic and genomic testing and family history that led to these recommendations. Let me pause there and see what other comments John and Greg or Kristin might have. John?
>> John Glaser:
I don't have much to add, Doug, other than you forget yourself as one of the great presenters there. I do think it was very well received and at reflection of some terrific work on the part of you all.
We'll move on the next item, although Greg may have additional comments about having gotten those approved. Let's go off and make those happen.
>> Greg Downing:
Thanks, John and Doug. I think the only thing from the staff's perspective is that the interchange and the depth of dialogue coming from the AHIC members and the leadership that both of you portrayed was really quite rewarding. I think there was acknowledgment by a number of the AHIC members of the extensive dialogue and work that the subgroups and the Workgroup itself did for coming forward, and in particular, from the government side, anyway, the interagency dialogues that have taken place in terms of reviewing documents in advance of coming forward. So I think there was full recognition, including by the Secretary, of all of the work that this group has done to getting these recommendations brought forward. So thank you all.
>> John Glaser:
Any questions or comments from the folks on the phone? Terrific. Well, again nice work. Now we'll go into the item regarding the implementation plan. So having had obviously these recommendations approved, the discussion shifts to steps we need to take to make these recommendations, turn them into reality. For that, Greg and Kristin will lead us through the process.
>> Kristin Brinner:
Okay. This is Kristin. You'll see the first slide, which is the AHIC priorities and use case road map? It’s even less readable on the computer screen than on paper. But I think that was part of the point, to show that they are working on a lot of different topics and have a lot planned for the future. Just as a recap, in 2006, they had three use cases. 2007 they had four. And then 2008, which is a little confusing because those are the actual use cases that are being developed right now, so they're being completed in 2007 for further implementation in 2008. And those, some of those were approved at the July 31st meeting, including the one for personalized healthcare. And you can see for 2009 and beyond, and we can continue to add to this list, there's just a listing the various topics that Workgroups have been discussing that could potentially be grouped into future use cases.
The next slide is a rough development timeline for the 2008 use cases. And I will describe what is exactly a prototype and a detailed use case is in further detail once I've gone through this timeline. The prototype use case is being developed right now and will be released for public comment on the 27th of August. And John Loonsk has volunteered to discuss this use case with the Workgroup by conference call and Webcast. We're working on scheduling that, but it looks like it’s going to be September 5th at 4:30 p.m. We'll send out an email notification about this once the details are finalized. And that comments on this use case will be due sometime at the end of September, given that there's a four-week public comment period. And this prototype use case really just describes the flow of the use case at a very high level and is used to start engaging the stakeholders in this process.
Following that public feed pack period, the Office of the National Coordinator will take the public comments and process them and begin to develop the detailed use case over a period of five or six weeks. And by my calculation, that brings us to about the beginning of November, when that detailed use case will be published. And that will be followed by a period of four weeks for public comments, and that would bring us to the end of November. Taking those public comments, ONC will then take them and combine them into the detailed use case and revise as necessary. And the plan is for them to release the final use cases by the end of December. So these use cases will be done by the end of this year. The detailed use case takes the prototype use case and then documents all the events and actions at a much higher level of detail.
So the use case concept I think is best explained if you look at past use cases, and I was just going to just quickly run through the medication management use case as an example. So the prototype use case, for all of the use cases, and the medication management is just one example, describes the stakeholders, so that might include the clinicians, the consumers, health information exchanges; lists obstacles and issues, that might include the CPS issues we've been dealing with, auditing, patient identification; and finally the prototype use case lists the perspectives that will be identified and addressed in the use case. And for the case of medication management, this includes the clinician, the pharmacist, and the consumer. It then takes those perspectives and describes several scenarios where the interactions across these, interaction across the perspectives is then described. So for the medication management use case again, the inpatient medication reconciliation, access to current medication, allergy information, prescribing process, and ambulatory care. So for each one of these scenarios, the use case basically generates a diagram that shows how the different perspectives will interact. So in one example of the scenario, for the inpatient medication reconciliation, this may include the clinician, how the clinician gathers medication allergy information, how the clinician orders the right discharge perspective and it also includes the different health information exchange aspects, so whether you're drawing from EHRs, PHRs, pharmacy systems, et cetera. And it basically generates a diagram of that. And what the detailed use case does is it then breaks down that diagram into a really granular level of information. So it describes all the different aspects of these different interactions. And so it's a detailed use case basically just blows out the prototype use case into a much higher level of detail.
So in terms of the followup from recommendations that we had accepted at the AHIC, that kind of covers recommendation 1.0, that the Community should advance the area of personalized health care as a priority of use case development, and recommendation 2.0 that an extension to an existing use case, that's the harmonized use case for EHRs and lab results reporting should be developed to address specific needs for genetic testing. In terms of further Workgroup engagement on the recommendations, I'll just go through the recommendations at this point. Recommendation 2.1 is that a multi-stakeholder workgroup should be formed to identify what types of data information are gathered when performing genetic and genomic tests and to identify the appropriate standards, et cetera.
So as we better understand the shape that the use case is going to take, and we can work from the very detailed level of information the genetic and genomic tests subgroup already generated, and that’s the matrix where they laid out the various different types of genetics testings, we can work with ONC as it develops its detailed use case; and then I think, importantly, ONC has also offered to help facilitate our interactions with HITSP as they then take the use case and begin to identify what standards will be used for those different types of information exchange. We will start with the genetic and genomic test subgroup in terms of evaluating the use case and then we can convene the stakeholder group as necessary as this process continues.
Recommendation 2.2 remains a coordination activity led by the NIH, so it's possible in the future we may want further update on these activities.
Recommendation 3.0 is that a multi-stakeholder workgroup should be formed to develop the core minimum data set and common data definitions available for primary care collection of family health histories. And that activity at this point is being led by Mary Beth Bigley of the Office of the Surgeon General and Greg Feero at NHGRI. And they have worked with us to convene a group of stakeholders. We have been reaching out to these stakeholders over the last couple months for recommendations. And basically what they asked for the Workgroup were any documents the group, that any of these stakeholders had already generated in terms of what they believe a core family history should contain and what processes they use to generate that document, whether it was through stakeholder engagement, internal processes. We just kind of want an understanding of what people had already done in this area, since it's been extensively covered already, and we didn't really want to reinvent what other people had done. And the first call for the stakeholder group is scheduled on Wednesday the 22nd. And then this will be followed up by another meeting in September to discuss the documents that they have been generating. And through these documents, we'll continue to work with the Office of the National Coordinator as they develop their use case and further along the line in terms of the standards identification.
And then lastly recommendations 3.1 and 3.2, we continue to work with HHS and NIH as well as Indian Health Services, Department of Defense and the VA, among others, to engage them in this process and find out what they already are doing and just to try and coordinate the efforts. So that covers the recommendations and the use case development process, and I'd be happy to try and answer your questions if you have any.
>> John Glaser:
Yeah, I think any questions for Kristin?
>> Marc Williams:
Marc Williams, Intermountain. Kristin, I hate to rain on your parade a little bit, but the September 5th date may be problematic for a number of us that are participating given that the Secretary's Advisory Committee on Genetics, Health, and Society oversight taskforce has an all-day meeting scheduled that day. I know there are at least four or five participants in this group that are going to be involved in that meeting.
>> Greg Downing:
Thank you, Marc. We've been talking with Sarah about that. And we're waiting for John to get back on Monday and we'll coordinate that. So we're aware of that potential conflict. And we'll work to see if we have a second session or if there's some way to address that in the agenda of the meeting that you're attending.
>> John Glaser:
Any other questions, comments? I think it was nicely done, Kristin, of laying it out here, especially for people like me. Blue is ONC and yellow is us. I think what would help me anyway, and perhaps help all of us, is if you all put together a schedule throughout the fall of the various calls, their topics. And if we lay out as you did verbally the dates that sort of bound the comment periods, if you could also lay that out. And particularly if there's opportunities during the comment period for us either entirely or interested members to get together and see what the feedback might be, and, as you mentioned, engage John in perhaps some of those conversations. So a layout of the dates and calls throughout the next couple months I think would be helpful.
>> Kristin Brinner:
Sure. We will be happy to do that. We were thinking that once they are released, it would also be good to schedule just one more subgroup call after people have had a chance, both the genetic test and the family history subgroups, so that the subgroups would have a chance to discuss amongst themselves as well.
>> John Glaser:
I think that would be terrific. And the other is sort of the approach to working with HITSP in an iterative fashion towards the end of this timeline. It sounds like we still have some sorting through about how best to do that.
>> Kristin Brinner:
Yep.
>> John Glaser:
Okay. any other questions or comments? Perfect. Thank you, Kristin. Why don't we, Becky and Amy, see if we can get an update on the CPS activity that you all have been doing?
>> Greg Downing:
John, before we start, I just wanted to note that Steve Posnack from the ONC has been coordinating from the CPS Workgroup is here also as a resource for the community. I'm not sure if Deven is on the call today at all or not. But as you know we have been coordinating this subgroup's activity potentially in parallel with the CPS Workgroup. So sorry to interrupt.
>> John Glaser:
No, that's fine, Greg, thank you.
>> Amy McGuire:
This is Amy. But I'm not sure, I think Becky's on the phone.
>> Becky Fisher:
I'm here.
>> Amy McGuire:
I'll go ahead and get started. And, Becky, if you're there and want to jump in, please feel free to.
So what we've been working on is we spent some time as a subgroup, and particularly Becky and myself and some of the staff, talking about what our charge is and how we can be most useful and helpful to the larger working group. And so the way that we've decided to go, and I would love, I think we would all love some input from members of the larger working group if this doesn't seem like the appropriate way for us, but we thought that we would focus our attention and our energy on providing the larger Workgroup with some conceptual frameworks for ways to think about confidentiality, privacy, and security as it pertains to genetic information in the electronic health record. And we're currently working on putting together a document, which I don't think the larger group has seen because it's still very much in draft form, that we're hoping to use to really address three major issues that we identified as kind of the three major issues that would be most useful for us to address in this context.
The first is kind of a larger debate about genetic exceptionalism and whether additional or special protections ought to be afforded for genetic information as it's stored and used in the electronic health record. And we spent most of our time discussing this and trying to really lay out the pros and cons of treating genetic information as special and providing additional or special protections for genetic information. And I guess when we talk about protections, I think we're talking about both legislation that might protect genetic information but also special protections informatically within the electronic health record, whether it should be somehow separated out or stored in a special way or separate way. And I think where we are right now is we are still trying to decide, and if there's suggestions on this, I think we'd be open to them as well, about whether we should be proscriptive on this or more descriptive of sort of the pros and cons of treating genetic information as special or having special protections. This is obviously a very big debate, and it's something that people have very strong feelings about, so I'm not sure if we can actually reach any sort of consensus, but we haven't yet tried to do that. But right now kind of what we've been doing is thinking about the ways in which genetic information is different from other types of health information, if any. and what some of the arguments in favor of having additional or special protections for genetic information might be. And then thinking about the counter-arguments for what some of the risks of doing that might be. And if nothing else, I think sort of laying out those different arguments might be useful to people as they're thinking of the different ways in which they might want to try to protect genetic information differently than they would protect any other health information. When I talk about genetic information here, I guess one big question is how broad in scope are we discussing. That's been a discussion for the larger Workgroup, and I think we would include family history information, as well, in most of these discussions. So we've been working on that primarily, and as soon as we have sort of a draft of that that the subgroup has agreed on, then I think we can send it around to the larger group for comment and for additional suggestions.
The other thing that we're thinking --
>> Joel Kupersmith:
Could I say something about that? This is Joel Kupersmith from the VA. If you're going on to another topic. We are doing focus groups with veterans probably between September and November, which will touch on those issues. And I know that the Genomics Institute, we're using the same group, actually, as a consultant, as they are. And I know their effort has been publicized, also. But certainly I think that would inform what you're doing. And I would certainly think that before you're proscriptive, you ought to see what descriptive shows, whether there are major differences among groups of people and that sort of thing.
>> Amy McGuire:
You're looking at how they view information, whether it's special or not to them?
>> Joel Kupersmith:
Absolutely, yes. It's only one of the things we're looking at, but it's part of a whole effort to try to get veterans' input into what we're doing.
>> Amy McGuire:
Yeah. I think that's a really important point. We would love to kind of follow up with you on what you're finding and things like that. I think that it is important to get individuals' perspectives and all the sort of different stakeholder inputs on this issue of genetics exceptionalism. You can address it from many different levels, from sort of the conceptual level, from the policy level. And we're trying to articulate the many different perspectives on why or why not you should be treating it as exceptional or not exceptional. And I think that that's one perspective that we don't have, so that would be great to follow up with you if you don't mind.
>> Joel Kupersmith:
Sure.
>>
Just a comment, when you use genetic information, there's two dimensions of the, two axes you should be explicit about and maybe distinguish, because when you say genetic first you think about just the codon, but when you look in your discussion of the biochemistry tests in the prenatal screening, you’re calling that genetic, and some people would say cholesterol level is high is genetic, so you have -- you don't have a clear border in the kinds of information. So you should probably add some definitions or else you're really talking about all information already.
The second part is that when you're talking about that patient's own information that they're volunteering to some context that they're agreeing to something, there is a difference between them volunteering their own and their uncle's or their brother's in the family history. So there is a little distinction there you ought to at least kind of highlight.
>> Amy McGuire:
Yeah, and I think I'm not remembering. I know at one point we were sort of laying out what our definition is. But I think that that's actually one of the biggest issues that we discussed sort of in the arguments against treating genetic information or whatever you want to call that as special, which is how do you even define what is genetic information and there are all of these blurred distinctions. And so that makes it very difficult to have special protections for special types of information. So that's a very good point. But I agree that we should have some definition of what we're actually talking about.
>>
We are working or have worked our way through some definitional issues as far as presenting it to the public.
>> Andrea Ferreira-Gonzalez:
This is Andrea Gonzalez from VCU. I have two points to make. I have just come across an article about Genetic Information: Special or Not?, published by the group with Wiley Burke, that actually has engaged in some focus groups with individuals of the public to ask that specific question. And I think it's very revealing to read that article. It is from the American Journal of Medical Genetics.
And secondly I want to point it to the attention that the Secretary's Advisory Committee on Genetics, Health, and Society overtaskforce is currently wrestling with that particular issue. On our charge from the Secretary, we were asked to look at genetic testing to see if it's different or not for any other testing. So I would strongly encourage you to either start a dialogue or some collaboration or exchange of information with our group to see where we come out.
>> Amy McGuire:
What's that group again?
>> Andrea Ferreira-Gonzalez:
The Secretary's Advisory Committee on Genetics, Health, and Society oversight taskforce. We have just been charged by the Secretary to look at the state of oversight of genetic testing in this country. And one of the questions that the Secretary posed to us was that particular issue of exceptionalism. So I would strongly encourage you to contact Sarah Carr in the office to maybe see how we can start exchanging information as we prepare to discuss this issue.
>> Amy McGuire:
I think this is a great plan. I think we’ve recognized as a subgroup that this is a very large issue that is being debated on at all levels. And so we’ve spent a lot of time deciding how we could best draw on the existing literature and present it and package it in a way that might be practically useful for individuals and groups that are trying to make decisions about what kind of special protections they want to do. So I appreciate that input.
>> Andrea Ferreira-Gonzalez:
Do you want me to send you the cite for Dr. Wiley Burke's article, too?
>> Amy McGuire:
I think that would be great. I think I’ve come across it. But it would be great if you could send it.
>> Andrea Ferreira-Gonzalez:
Great.
>> Becky Fisher:
We’ve done a literature search on this quite extensively. And a lot of us have a degree of crossover with Sarah and some other people. So we're definitely open to your input, but we are already looking at much of what you suggested.
>> John Glaser:
Other questions, comments? I had the opportunity, although I know the full group hasn't, to review an early version of the document that you mentioned earlier in your comments here. I think it's terrific. Actually really, really neat and very, very thoughtful. So I first of all want to say I'm sure the end product will be as good if not better than what you have. But you have already done terrific work. I congratulate you. It's really quite nicely done. My encouragement to you, and I realize that there's lots of people looking at this issue in lots of different forms and that there are diverse opinions on the topic, but to the degree the Workgroup can come to convergence or consensus on some principles for things along those lines, I would encourage you to do that. I don't expect you to solve the problem but I think that would be helpful to the degree that’s possible in your conversations.
>> Amy McGuire:
Okay. Just to kind of finish up the other two components of what we're working on, and we've done much less and I should also acknowledge, I know Becky kind of jumped in there, but Becky really kind of created the draft of this. She's really done a lot of work on this even though I'm speaking. But the other two components that we’ve kind of been working on, we haven't gotten to these quite with the amount of time and energy that we've addressed the genetic exceptionalism issue with, but we plan to, is the question of the integration of genetic information in the electronic health record. And we've kind of laid out sort of a conceptual framework for how we might want to think about this by breaking it down into the ways in which genetic or genomic information is generated and the context in which it's generated. So, for example, it's generated in the research context, in the clinical context, in the private industry commercial context, and the type of information that's generated in each of those contexts. So trying to come up with some sort of matrix or at least framework for thinking about the ways in which and the types of information that are generated. And then to think about whether or not and how each of those categories of information might want to -- we might want to include in the electronic health record. So there may be certain types of genetic information that's generated outside of the healthcare context, like in the private commercial context that we may decide should not be included in the electronic health record. So if somebody goes to a company once they get started up like 23andMe and gets their genome sequenced and goes into the doctor's office, and says here, I want you to keep this, you can tell me what it means, that may not be an appropriate role for the doctor to have, and we may not want to include this information, at least at this time, in the electronic health record. We haven't discussed that. That's just an example. So trying to think about and there might be other scenarios like in the research context where it might be appropriate to have it link somehow to the health record but may not want to include it in the health record for various reasons. So those are the types of things that we want to think about with regard to that. Becky, anything else about that one that you have to add?
>> Becky Fisher:
No, I think you covered it.
>> Amy McGuire:
Okay. And then the last component is the component of public education. So we're going to try to, to the best of our ability, try to think about and identify the educational needs of various sectors and stakeholders from the consumers to patients to physicians to researchers related to each of the categories of genetic or genomic information that we identify. And then I don't think we can really develop the educational content, but at least to try to do sort of a needs assessment of what some of the educational issues might be and the educational needs might be of those different groups.
>> Doug Henley:
Sounds great. Becky and Amy, thank you for your great work on leading the workgroup. We appreciate the work of everybody in moving forward. I concur with John. It's an important topic. And we look forward to your future recommendations on that.
As you all know, on September 17th, we get together in D.C. face-to-face for another PHC Workgroup meeting. And this one will be focused on clinical decision support tools and we've got some excellent presentations scheduled for that meeting. But as you may also be aware, John Glaser has been working extra duty with a CDS separate workgroup. And he's going to give us an update now on how that is proceeding and how that relates to our meeting on September the 17th. John?
>> John Glaser:
Yeah, you guys may not know this but if you work on a lot of AHIC stuff, you accumulate points that can be redeemed for valuable prizes. So I'm trying to get a toaster out of all of this.
You probably have the agenda in front of you. And I think the sort of core intent of the day is to really bring the group, to educate the group and bring us all up to a common base understanding of decision support, both some of the tools that are available and some of the experiences, et cetera. To the degree we can, we will try to outline how we might tackle it. But really that's the secondary intent. The core intent is again a broad exposure to a wide range of things that are going on across the country, both in the government sector and also in the private sector. So as you can see on the agenda in front of you, or at least one of the attachments --
>> Kristin Brinner:
John, this is Kristin. I’m actually going to jump in because we didn't send out the agenda.
>> John Glaser:
I'm sorry. Then let me give you an overview of it. We're going to obviously have some opening comments and turn it over to Jerry Osheroff and Jonathan Teich, who did work on the roadmap for clinical decision support, which I presume we should send out, I know we should send out to you, so you all see that before the meeting. This was work that was commissioned by ONC to arrive at a national roadmap for advancing the use and the effectiveness of decision support. They did a very nice piece of work. This was also in conjunction with AMIA, the American Medical Informatics Association. So we're going to ask Jerry and John to both facilitate the conversation, but also to present an overview of the roadmap that they and their group developed. And this was presented to AHIC last summer.
There is then a panel in which we'll have five different organizations slash individuals presenting some of the commercial clinical decision support tool work that is being done, so you'll get a feel for the various aspects of this and different ways that people approach this.
We have after a lunch a second panel, composed of folks from both the government sector and the private sector, talking about how they have used decision support and how they've gone through the process of developing the evidence that is then instantiated in the decision support logic.
So that's really the core is both of those two, exposures to the tools and also people trying to use the tools. And then if we have time, we'll begin to see, well, given all of this, how might we think about addressing this topic of clinical decision support given the focus that we have broadly on personalized healthcare? So I suspect that most of the plan formulation or things that we will do will occur in calls that follow the gathering on the 17th. And I think the 17th should be a very interesting day. So let me just stop there and apologize for thinking you all had this and see if there are any questions or comments.
>> Kristin Brinner:
John, this is Kristin. I just wanted to jump in and supplement your comments. The rationale behind the two panel discussions was really for the first meeting to serve as a good information gathering session. We felt it was important on the first panel to bring in some people that have either been developing clinical decision support tools that then get inserted into electronic health systems or people that are working on larger systems where within those systems they are actually developing the tools themselves. And that we could then follow-up with the discussions about the various aspects of evidence development. And this has all been kind of in light of, or you can reflect on the Incidentaloma paper that we sent out that highlights the importance that clinical decision support will play in terms of supporting the aspects of personalized healthcare, interpretation of genetic test results, and family history information. So that paper really serves as a prologue for the discussions that we are going to be starting at that meeting. And you're absolutely correct, this is just the beginning of the discussion, and we think that from this meeting we will better hone in on who we want to speak to further about different aspects of things that we learn and areas that are potentially right for us to make recommendations.
>> Andy Wiesenthal:
This is Andy Wiesenthal from Kaiser Permanente in Oakland. Can you all hear me?
>>
Yes.
>> Andy Wiesenthal:
Because the telephones blew up in Oakland about 45 minutes ago. I was just able to call in. They just fixed it. I wanted to chime in on this particular issue just to ask if anybody saw the Wall Street Journal yesterday.
>> Andrea Ferreira-Gonzalez:
We did.
>> Andy Wiesenthal:
You did. So that might be an interesting case to take up because for those who didn't see it, the FDA in the Wall Street Journal featured this particular intervention. The FDA has just decided that in the labeling of Warfarin, all Warfarin products, clinicians will be advised that there are some genetics tests that they ought to do. And if you can just think of how many people in the United States are on Coumadin for more than a few days, the implications of that decision alone are staggering. And the kinds of decision support that clinicians might require in order to help decide whether somebody needs the test and if they get the test, what kinds of discussions to have with the patient. I mean, there are millions of people on Warfarin every day. So this is just, it's not exactly the tip of the iceberg, it's a rather large iceberg itself. But it's the beginning of a trail of initiatives that the FDA is likely to take.
>> Greg Downing:
Andy, this is Greg Downing. Thank you for that.
>> Andy Wiesenthal:
If you mentioned that before I called in, I'm sorry.
>> Greg Downing:
Thank you for that prompt. This is indeed sort of exactly the kind of scenario if you read the article, we'd be happy to distribute it, that demonstrates potential for a healthcare provider, potential confusion over this issue. And I think that's been fairly well characterized in that paper. Indeed for this session, we have one of the laboratory companies that has been working with the clinical decision support entity who has a tool and has been working with a number of the health plans. And one of our current agenda issues is to feature the handoffs, if you will, in parallel with technology development of the knowledge facilitation tools to address this issue. So I think that case study, as you suggested, is certainly one of the feature aspects of this September 17th meeting. And we've been doing a little bit of handoffs and coordination with these entities that are working together and that will be present at the meeting. So we think this is an important common case that could provide a good learning structure for this Workgroup and beyond as potential areas that the Workgroup's activities could focus on.
>> Andy Wiesenthal:
Just so people have an understanding of the volume, we, Kaiser Permanente probably has, on any one particular day, about 75 to 100,000 members taking Coumadin.
>> John Glaser:
Goodness gracious.
>> Anthony Achampong:
This is Anthony with HRSA. Does anyone have the electronic version of this Wall Street Journal article that can go around?
>> Greg Downing:
Yes. This is Greg. We'll send it out to the membership today.
>> Doug Henley:
It was in the New York Times, as well.
>> Anthony Achampong:
Okay, thanks.
>> John Glaser:
Any other comments, questions? Thanks, Andy.
>> Andy Wiesenthal:
You're welcome.
>> Marc Williams:
Marc Williams. Something a little more prosaic. When will the logistical information for the 17th go out?
>> Kristin Brinner:
It should be going out early next week.
>> John Glaser:
Anything else? Look forward to seeing you all there. I think it will be a remarkably interesting day and look forward to it.
The next stop we have is on information exchange for newborn genetic tests. So if we're ready, we can go ahead and start.
>> Greg Downing:
John, Doug, I don't know if we want to do a bit of a prologue for this, but I think we've had some background prior sessions reflect this Workgroup's efforts and the visioning session back on March 23rd, that in the priority aspects, we identified areas that could be developed further for potential consideration as actionable by the Workgroup and then ultimately potentially for areas that AHIC itself would be provided from the Secretary. I think as mentioned in the last meeting, staff has been working with a number of the members of the Workgroup in facilitating the development of potential Workgroup areas. And two of those are being presented today. These are new for discussion but parallel the activities on CDS and confidential and privacy and security subgroup activities. So the chairs today are being presented, will be presenting these two background areas on the newborn screening test health record information interchange as well as the interoperability of the clinical research and for clinical aspects of pharmacogenomics. So these next two presentations have been worked on as development areas to be presented here today, and the chairs I think have the discretion of deciding whether we, how we take this forward after the presentations and the discussions on this call. I'll stop there.
And we have Marie Mann is representing HRSA. Michelle Puryear was not able to join us today, has worked with the staff in coordinating this across several of the Federal agencies and the public sector organizations that are involved in this area. And she will be making some comments, as well. And Brad Therrell from the National Newborn Screening and Genetics Resource Center will also be participating in the discussion. And as Marie will provide the background descriptions of that organization as well as Dr. Mike Watson's organization in the overall aspects of newborn screening coordination at the Federal Government level. So I'll stop there. Marie, do you want to make any opening remarks or comments on this session?
>> Marie Mann:
Yeah, I thank you for the opportunity. And I'm just going to say a few words, just provide some more of the framework for what we're talking about, the subject of newborn screening, before we move on to Mike Watson's presentation. But when we're talking about newborn screening, I think many, probably most of you are familiar with we're talking about newborn screening and heel sticks. But we also -- heel stick screening -- but we're also talking about newborn hearing screening. And this screening is at present, is a State mandated activity. And that in fact all States mandate newborn screening as a public health activity.
While all the States do mandate it, they vary in the number of conditions screened by each State. And the decision on what constitutes the screening panels for each State are made at the State level rather than at the local or individual level. Besides the disparate screening panels, the States also vary in how the various components in the newborn screening system are implemented. And when we are talking about newborn screening system, we're really talking about about six components. And they include education, the process of screening itself, the follow-up of the screening results, the confirmation and diagnosis process, the treatment and management, and then the monitoring and evaluation of the program.
The successful follow-up of the out of range lab results with the appropriate management of the results, requires the processing and sharing of essential information among multiple partners. And they include both the screening laboratory, the program, the newborn screening program follow-up staff, the child's primary care provider, the diagnostic laboratory, subspecialists that may be involved, and then of course the child's family. And for many of the conditions that are currently screened, short turn-around time from the screening to the diagnosis and to the intervention is extremely critical. So that we really, it's necessary to have effective information exchange.
What that information exchange allows is that just-in-time information to the health professional and the families with linkage of the decision support tools. It also allows for it to avoid any missed opportunity for a screening. So to identify those children who might be missed in an initial screen. It allows for the coordination and the continuity of care as well as the long term tracking of the children with the identified conditions, both for the program performance monitoring and continuous quality improvement, as well as for understanding the natural history of many of those conditions which are very rare.
So to address the different aspects of the information exchange, the issues of developing metrics for reporting of the results, developing the appropriate standards for the raw results, standardizing the diagnostic codes, adoption of these standards by the vendors who supply the diagnostic equipment, and developing standard guidelines for clinical decision supports needed.
And then finally just a couple points about how newborn, in thinking about newborn screening tests, how they may be different from other genetic or genomic tests. The testing, as I'm saying, is generally done, one is that you obtain the specimen generally in the birthing facility rather than the doctor's office. And the ordering of, the person ordering the test is often, actually most often not the one who is going to receive the results. So it's critical to make sure the right person receives the results. The testing itself is generally done in public health laboratories; however, there are some States that contract out the screening itself to for-profit clinical labs. The patient name may change between the time of ordering the test to the time the results may be reported. This is not always, but it can happen. And so that needs to be taken into consideration. The followup and the confirmatory tests are common, but there may be a long lag time of many months before the actual testing may be completed. In that interim, the provider may have changed. And then some of the conditions that are identified through newborn screening, as I said, are, may require emergency intervention. And that may require the child being taken to emergency room before even being seen by the primary care provider. And then some tests, such as hearing screening, may be done using point of care devices at the bedside rather than actual specimens going to the laboratory. And then the results will be still reported back to the primary care provider in the ambulatory setting.
And then the results of the tests needs to be maintained over a long term period. And so it needs to follow the child and transition into adult care. And then it's important to know, too, that when we're talking about the screening itself, it is not necessarily just one test. In fact, in about 25 percent of the newborns in this country have a mandated second screen. So for many of the children they have two screens. And then an additional number of babies will have a second screen due to whether the specimen was inadequate or there was a result that was out of range. And so the results from these multiple specimens are essential, need to be linked, and often this is a difficult process. And then finally, to insure that all newborns are screened. And that is not necessarily being done at this time. But to do that, it really requires linking the newborn screening results to the birth certificates.
So I just want to make sure some of those points are conveyed to you. And certainly after Dr. Watson's presentation, we'll be glad, Dr. Watson, Dr. Therrell, and I will be glad to provide any answers to your questions that you may have.
Before I introduce Dr. Watson, I want to, also on the phone with us today is Dr. Brad Therrell, who is the director of the National Newborn Screening and Genetic Resource center, which is funded as a cooperative agreement with the Maternal and Child Health Bureau and HRSA. Dr. Mike Watson is the executive director of the American College of Medical Genetics, and he was the director of the HRSA-commissioned project to outline a process for a uniform screening panel and system. And he's currently director of the National Coordinating Center for Regional Genetics and Newborn Screening Collaboratives, a program that is also funded by Maternal and Child Health Bureau and HRSA. So, given that, I will pass on to Dr. Watson.
>> Mike Watson:
All right, thank you, Marie. If it looks like I'm sweating, it's not the pressure. I've seen most of Fort McNair today. They were certain it was the Mary Surratt building, they’d written Switzer wrong, so they sent me deep into the bowels of Fort McNair.
>> John Glaser:
Glad you survived.
>> Mike Watson:
Really. Well, I wouldn't say I looked like I survived but I made it here.
So in addition to the things they said about me, I'm an adjunct professor of pediatrics and genetics at Washington University where I ran genetic testing for about 15 or 16 years before I came to this job. So that's really my background going into this. First off, who is controlling the slides? I'll go through a couple of these quickly because I think we've already covered some of the things on them and I want to leave enough time for some conversation at the end of this. So I'll give you sort of the party line as to why newborn screening is important.
Ninety-nine percent plus of babies born in the United States are screened in newborn screening programs. That's 4.1 million infants every year, so it is the single largest area of genetic testing in the United States. It's one of those things, as Marie said, and I'll show you a diagram that tries to show this a little more clearly. But it requires some significant interoperability among programs. State public health programs have their own information systems. The private sector has as many as you can imagine information systems. There are a wide range of provider groups that interact with newborn screening, as well as diagnostic laboratories, physicians and it's one of the largest sorts of programs imaginable. As Marie mentioned, there are, 28 of the 29 conditions are metabolic diseases. The other is hearing screening and that occurs in the nursery. Some other screens that are occurring within the nursery setting that will then have to add a new dimension to this of interaction with the State programs who will control the quality assurance and back out to the private sector where management, confirmation and those sorts of things will occur.
It's a ton of genetics, frankly, in newborn screening. People think about things like PKU as being PKU. But if you get hyperphenylalaninemia on a newborn screen, it could be any one of five genetic conditions and other causes. And when you go through those 28 conditions that are listed in our newborn screening panel, including hearing loss, we're talking about over 140 gene loci being involved in the potential diagnostics that come from those screen positive infants. And it's a wonderful opportunity because it defines a birth cohort that would greatly benefit by starting life with an electronic health record. And there are certain States that have gone further in linking birth records, maternal serum screening records,and newborn screening records. Most States have not made that much progress, but it's certainly a direction in which the programs are moving. Next slide, please?
This I'll skip because Marie told you the six critical parts of the newborn screening program. Educate, screen, figure out what they’ve got, and make sure the programs are working right, and treat the kids.
This begins to try to give you a perspective of how the pieces of program and the information flow within a newborn screening program operate. At the top there, on the left side, boy, I can barely see that. Hospital nurseries is obviously where most of the, is where the functional tests begin, and that would be a screening test like hearing loss. That information is reported on the newborn screening card on which the blood is, has been deposited on a filter. And that goes back to the newborn screening program. And then that blood spot is the material that is analyzed by the newborn screening laboratory.
Out of that newborn screening laboratory comes the information that the infant is of course screened negative. And that is the only very first step. And I understand that LOINC has made some efforts towards defining some of the tandem mass spectrometry analytes in LOINC, but my sense is they haven't thought much about the fact that it's a two way street. All you send out of the newborn screening laboratory is the screen positivity of the infant. That can range from very low levels of positive predicted value to relatively high levels. So you could have varying numbers that turn out to be positives at the end of this whole process. And it varies from State to State as to how that occurs. The primary care provider, in some States it goes out to a specialist as well as a primary care provider. And they then establish the diagnosis for that child, either that it was -- they hate it when I say false positive because it's not really a false positive in a screening program, but it turns out to be a not true positive, and that goes back to the program, as well as any diagnostics that might be identified that define the true diagnosis. That could be anywhere from enzymology to mutation analysis to things that also may include some clinical information in an infant that might be found to be positive in a diagnostic setting. That has to go back to the newborn screening laboratory because that now is the point on which they determine what are their positive rates and who do they follow for the long term or the short term? At a bare minimum, those are the kind of major pieces of information that define newborn screening at the very front end step.
But after that, you obviously have now a child in chronic care for the rest of their lives, basically. And that's a very significant number of infants. We estimate that somewhere in the neighborhood of 8 to 10,000 infants each year enter a chronic disease management program and have on going treatment and management of whatever the condition they had. Some of them are very rare conditions and there may be only a couple of laboratories in the country or world that do the diagnostic testing for that particular condition. Others, like sickle cell, are much, they're much more readily available. Hypothyroidism, readily available. So there is a wide range, both of complexity and the numbers of the people who might be involved, depending upon what you screen positive for.
So then you move out into that whole clinical laboratory area where information is -- really, I separated that from diagnostics because historically the newborn screening programs have really not done a whole lot after finding out what the diagnosis of the infant was. But it's clear that we want to move towards long term follow-up, tracking the outcomes of these infants to make sure that the newborn screening programs are functioning as best they could. And there's a world of interest in having well established registries of these patients to facilitate second generation drug clinical trials and other things. We really are looking at moving towards a national activity around data collection, long-term follow-up, developing really repositories of things for research and for really guiding us in management, all of which we envision ultimately ending up in national data collection activities. We’ll have to see how that goes and we'll be having some meetings with state policy people with regard to issues they face in sharing things at that level. That will occur sometime this fall. So next slide, please.
So the minimum information as I said is screening test results. Out of the newborn screening will come analytes, and analyte levels. Unfortunately the case definitions of what is positive vary considerably across the States. So that requires that for the very base information that's quantifiable as to the level of analytes, one then has to have reference ranges for that laboratory’s testing in order to compare that particular quantitative result to their reference ranges to say yes, it's normal or abnormal. As it turns out, most programs actually go back to physicians with something less quantitative, which is probably the second level of the LOINC codes, which is not all that quantitative analytical information, but it’s screen positive for this or that and it's associated with these conditions or that condition. And that's a common scenario. And there's some result interpretation that allows that to occur in the physician's hands. And it may actually include providers who access newborn screening results before they’ve even received notification from the program. There's a fairly wide delay on screen negative children in getting the result out to the primary care physician or the specialist. The screen positives tend to get called out, or rapidly made aware to whoever the provider is. In fact, there are times when these medical emergency situations where babies can crash with these diseases before they even get to their first visit. I'll show you some materials we’ve developed to help physicians through that kind of an emergency medical situation. But what would be valuable to the physician population is when they see that baby the first time, if they don't have a result they have an easy mechanism to go in and see, in an electronic system, that screening occurred and that it was negative, in the absence of that being present in the chart for that infant.
Then we have the diagnostic provider area that is reporting back to the newborn screening program or to the primary care provider or to both in a collaborative management setting where the laboratory diagnostic information goes back and that can range oh, all across of number of different analytes and enzymes, mutations and things. And then clinical information that might also be supportive of the fact that an infant had that condition, in the event that they didn’t have the critical mutation early on. Next slide, please.
So there's some communication and information needs of the various participants in this sort of a program, clearly interoperability in electronic communications systems across this wide range of provider groups would facilitate the program operating efficiently and well, it does include States, it includes local institutions. Most of the genetics community is located at academic medical centers in the United States, so all of those institutions get linked into these systems and then it would also require some standardized laboratory and clinical languages to allow at the bare minimum, even in the absence of interoperability and the ability to easily exchange, if we're at least collecting information around standardized laboratory and clinical languages, it is compatible and we can begin on a national level to look at really how these programs are functioning, and things of that kind. Next slide, please.
So this was the report that we did that identified a number of issues, it related to newborn screening programs and their efficiency and recommended a uniform panel of conditions that ought to be included in screening programs. It's been endorsed by the Advisory Committee on Heritable Disorders and Genetic Diseases in Newborns and Children. It's advisory to the Secretary of HHS. Next slide, please.
As you begin to look at individual conditions, and I only pulled out a couple because I wanted to give you a general sense of the range of clinical and laboratory information that becomes part of that first stage of back and forth between the clinical laboratories, the physicians, and the newborn screening program. So this is the example I alluded to with hyperphenylalaninemia in newborn screening. It could be done by older methods of bacterial inhibition assays, it could be done by semi-quantitative -- I’m sorry, by fluorometric assays. And the most, the assay with probably the best positive predictive value and really cost effectiveness, because it can be tied to the broad utility of tandem mass spectrometry used in newborn screening is tandem mass spec, where we need the languages that are specific to those analytes. Obviously these different kinds of assays would have different reference ranges attached to them. So that's the first level of information out of the newborn screening labs that there is too much phenylalanine present. Next slide, please.
At this next level now we're out into the private sector, the diagnostic setting. And there's two sets of information for phenylalaninemia that are commonly developed. At the laboratory end, the quantitative amino acids, plasma amino acids that are looking for a phenylalanine to tyrosine ratio is one of the critical measures. Urine pterins, red blood cell pteridines, and then mutations in genes for phenylalanine hydroxylase, which is associated with PKU. Unfortunately there, I think last I looked there's over 450 mutations in that gene that can be associated with PKU. There are also two biopterin cofactor biosynthesis genes involved in that pathway. And those are slightly different treatments and conditions. And then you can also have two other genes that caused biopterin cofactor regeneration defects. So there's a total of five genes right there that can be ultimately, or start at least in the differential diagnosis of hyperphenylalaninemia, could lead you through the diagnostic pathway to decide which one of those it is or if it was nothing at all.
On the clinical side, in the short term, the minimal, there's a minimal dataset that is sort of due to emergency medical management or endocrinology. And then you have some long-term things that are diet parameters, how effective the baby has been treated and is behaving with the treatment. There is -- boy, I can't see anything. Prealbumins, ferritins, transferrins, I won't read them all. They're on your slides. But this is just to give you a sense of all the things that are done at the diagnostic side. They can obviously come back to the program and minimize the amount of information they get. But to maximize some of the data for long-term follow-up, those become fairly important parameters for monitoring the baby and how well they're doing.
There are a lot of other conditions obviously in the newborn screening programs. And a wide range of clinical parameters that might be involved and organs that might be involved in those diseases. So it's a fairly vast array of things. In the short term, I'll be putting together about a half a dozen of these to send to Betsy Humphreys at the Library of Medicine just to ask to what extent do we already have these languages available for these particular sets of conditions just to get a sense of what's not there. I think that we've been told it's there for the front-end screening but not as much for the back-end of the process where we're reporting back to the programs. Next slide, please.
So as I said, the long-term outcomes are going to be important and follow-up and management are important, and all lead into understanding the natural histories of these treated diseases because that is now the new measure against which new drugs will be measured in clinical trials. Because there hasn't been a natural history study of PKU itself since the late 1970s, since these babies have been in treatment that long. So we're looking at now the second generation of the treatment clinical trials. Next slide, please.
I’ll bore you silly with this. Well, you probably can't see it. It's on our Website at ACMG dot net if you're interested. These are all the conditions that are screened in the programs. The top section is the core panel. It defines analytes from organic acidemias, amino acid disorders, hemoglobinopathies. And for many of those conditions, there are that secondary list of conditions below which are really those other, often very rare conditions that may be part of the differential in working through that particular analyte. Next slide, please. So right there you can see that we're talking about 54 conditions that could be identified out of the newborn screening program.
One of the things we began to do in order to facilitate the ability of primary care providers to respond appropriately, given that any one primary care provider won't see more than a couple of these patients in a lifetime, or some in a year if they're interested in children with special healthcare needs, they may follow a few. But for the most part, they couldn't pronounce 3-methylcrotonyl carboxylase deficiency, they have never heard of it or seen it before until it appeared in these panels. And as we appreciated the fact that these panels were expanded, we wanted to develop materials that would help them in their management. We put together a set of materials that at the time we called act sheets, because they were based on the actions the provider should take rather than the fact sheets that had been historically part of the programs. And really in electronic terms they can be thought of as pop-up education materials. We make them available and we're moving rapidly towards trying to increase their availability by, as point of care educational tools, by manually attaching them to the report that comes out of the newborn screening laboratory so that they get that pop-up education thing in their hands at the time they are dealing with the result itself. We did that with, keeping electronic medical records in mind, with the real end point being their integration into an electronic medical record as a pop up education component and, for instance, Marc Williams I heard earlier, Intermountain Healthcare has already integrated these into their EMR systems as pop-up educational materials, as has the Regenstrief and a few others that have EMR programs in the country. Next slide, please.
I just want to give you a sense obviously we're not the only ones playing in this game. You have already heard from HRSA, so I won't tell you anymore about them. The Regional Collaborative system is one that's very actively involved with working with the primary care providers and trying to begin to coordinate collaborative management among specialists and primary care providers around these conditions. There are seven regions across the country. The National Coordinating Center is at the American College of Medical Genetics, and the fundamental goal of that whole system is to bridge public health, primary care, and specialists, with the fundamental goal being to move system -- I’m sorry, to move care closer to local communities, to improve access at the local level. Obviously it's that electronic step that allows more of the primary care providers at the local level to become more directly involved in working with their patients whereas historically they would have just shipped them straight out to the specialist for the most part. And then the National Newborn Screening and Genetics Resource Center, it has a strong relationship with the individual States and the newborn screening programs themselves. That's probably the thing that distinguishes Brad's activities from mine is his strong relationship with the State programs and ours with the provider groups and specialists and primary care providers. And then the other interested group is NICHD, who has this interest in developing a national newborn screening translational research network founded around data collection and using resources available from those States that can make them available. They dry blood spots for research and things of that kind. And is very much interested in this whole longer-term outcome aspect of newborn screening being built up. Next slide, please.
So, just to give you a quick sense of these ACT sheets, I'll walk you through one. I think the reason people are supportive of them and certainly special organizations are, is that our board of directors signs off on them as practice guidelines, and by doing that, we buy into them for the long term. Once we do one, we are obligated every three years to insure they're current, they're revised, they're retired, or they're reaffirmed. And in genetics they're mostly revised, I'd say, given the pace at which the field moves. All of them are in a public area on our Website. But go out to the newborn screening laboratories and through them with screen positive results. And our regional collaborative groups are beginning to integrate with their State programs, and they're making them available. Next slide, please.
One of them you can't tell much from it but this is just so you know that it fits on one page, which was one of our fundamental goals was to make it simple,straightforward. Next slide, please.
This is one for MCAD. It's defined by having too much acylcarnitine C8 with lesser elevations of C6 and C10 acylcarnitines. When you see those, MCAD is probably the only thing in the differential that you're thinking about. What we've built into these was the differential for that set of analytes that define the condition, a very brief condition description. Next slide, please.
And that gives the primary care provider, really what I think it gives them most is the ability to interact as if they're very well informed when they talk to the family. They can tell them what the course of events is going to be for working that child up. And when it's a condition like MCAD that can be a medical emergency where the babies can crash prior to getting to their primary care provider after having been identified as screen positive, it tells them what to think about, what to talk about when they’re telling the family that the baby screened positive. They can ask whether there's lethargy, vomiting, that’s an indication that they’re beginning to crash, and to head to the ER. So it has that kind of level of action information. It tells them something about what is going to occur in the diagnostic evaluation step. Next slide, please.
And then some clinical considerations around the condition itself. And then we provide links to some various national resources that are at various levels of depth and content that allow people to learn to the extent they wish to learn more about that condition before or after they speak to the family and are dealing with the child. And with the other specialty providers who. they are interacting with as this child is being worked up. Next slide, please.
On the back page of it, we make them State specific. And we allow the States to begin to integrate some of the links back to their own newborn screening programs. We build in both clinical and testing referral networks. That's being expanded right now into a complete geospatial information of all the genetic service providers in the United States. It will replace these links that are to less detailed kinds of information about the providers. Next slide, please.
I'll skip that one. That will kill you.
[laughter]
But that shows you this is just biochemistry if nothing else. So what we did as part of this exercise, in addition to the ACT sheets was we laid out algorithms. And in fact in that packet of materials you've got, this thick one at the start of the session, every one of our ACT sheets and algorithms, diagnostic algorithms is in that packet. What they do is walk somebody through at a very fundamental level what are the diagnostic steps that are going to take place? It may be plasma amino acids. Fundamentally, these translate out into clinical decision support tools at a very intermediate level. And that's one of the things we appreciated very early on. You know, people talk about clinical decision support tools and pop-up education as they might be one thing. But I think in reality they’re going to be very different things for people coming into this at various levels of expertise and knowledge. A tool that allows someone to speak in an informed manner with patients and other providers is not necessarily the same clinical decision support tool that the expert specialist will use to go through that diagnostic workup of the infant. Next slide, please.
That's one you can see. They can be quite complex. This is C5-hydroxy acylcarnitine. There must be eight or nine different conditions that are ultimately part of that differential diagnosis. Each one of those steps down through the algorithm is based on some different analyte being positive at various stages of the workup. However for this set of conditions, as with some others like PKU, you can also screen positive because your mother is affected. Or I shouldn't say affected because sometimes the mothers aren't clinically affected, but they are carriers who have not expressed. Carriers is probably a bad word, too. They are non-expressing individuals who have the appropriate genotypes and are non-penetrant at least at that point in time. Next slide, please.
You've got all these in great detail in that thick packet so I won't go through too many of them. Next slide, please. Keep going.
This is where I'm picking up speed. These are just showing you in detail different steps along those algorithms. Next slide, please.
Okay. I'm going to go through these very quickly. This is just a set of surveys done by the American Academy of Pediatrics, I’m sorry, of primary care providers. And the degree to which they found these things useful and beneficial to them in practice. We now have gotten responses from over 100. The results are quite consistent with what I'm going to show you. Can I have the next slide, please?
The bottom line here is when you look at the pie chart, purple, which is three quarters of that, is very definitely true. And then the light blue, which is out sort of one o’clock to three o’clock, is somewhat true. So those are all the positive ends of statements made by those people who evaluated these materials. And now what I'm going to do is walk through the basic questions that are at the top of them because the pie chart doesn't change much as it goes through it.
Do they find, are they clearly stated, is the content germane? And the vast majority thought they were useful and germane.
Would they recommend them to their colleagues, if you spell colleagues right, yes they would.
Is the content organized in a manner that would facilitate easy action on the part of the pediatrician? For the most part, yes.
Is there inclusion in the, are the algorithms also useful, clear, concise and such? Again, very high positivity around those aspects of them. Next, please.
Where would they like to see them? They want to see them everywhere. They want to see them out of their newborn screening program’s Website, they want to see them through their professional organizations. They’d like to see them in government sites, HRSA and other places where these materials are housed and some of the resource centers we've been talking about. Next slide, please.
All right and now this is just a quick lead in to another way that we’re, another activity that relates directly to this. As you know Katrina was a really bad day down in New Orleans and it had significant impacts on the newborn screening programs. And as part of our activities, we've begun to look at what we can do to help patients when they're in these disaster areas because they've been identified because they need to be on a treatment and it has to be maintained for them to be healthy and well. In many of the conditions that can be critical, in others less critical over these short-term intervals. So we've begun to look very hard at disaster preparedness for newborn screening. Next slide, please.
It's clear that the electronic health record for the patient is going to be an important part of this. We really are looking at three different tools that we need. We need something for the patients themselves to have to carry to take to others' locations wherever they might get dispersed during the disaster. We're looking at materials that help that specialty group reconnect. The genetics group in Tulane and LSU had huge problems reconnecting with themselves. It was only by luck that somebody that knew how to get through satellite communications, more luck that somebody actually had a laptop with all the patient's medical records in it, that they could actually interact with them by telephone, make sure they were getting their prescriptions, talk to manufacturers to insure that they were able to direct deliver if they had to a therapeutic to a patient. So it begins to involve a much greater number of patients when you’re trying to make sure that that management and treatment piece is maintained.
So I’ll go through the next slides at about 2 seconds each because you know what a disaster looks like. That's the lobby of the place where the biochemical genetics group in Tulane was housed. Next slide.
So when you think about these patients, they're actually quite different than many patients that are routinely seen out there. Usual patients in a disaster need security and the basics. Reliable information and appropriate shelter. This set of patients [inaudible] need all those things but they also need those special foods, the special medications that they're on, often orphan drugs. Special laboratory testing needs to be available because this disaster could strike anywhere along the pathway. The infant could have just screened positive. They could be where their sample is lost in the mail and now it's a manner of knowing that they had a sample drawn, it wasn’t received, and they still need their screening done. It could be out in management where they need to be maintained on their treatment, and have access to those specialist providers who care for them. And then special information is also necessary for many of these patient populations, because most physicians, and certainly ERs, have limited experience with this patient group. And ERs do what they do, you walk in looking like you’ve got something, they'll do whatever the most likely thing is, and that's often not the best thing for some of these patient groups. Next slide, please.
So the people that had to get interconnected were patients to their providers; patients to their treatments, which often involved pharmacies and manufacturers; patients to emergency services; providers with their other team members; providers with patients’ medical records; providers with diagnostic laboratories, if they were in the course of finding out some of them screened positive and the laboratory they typically used goes down you need access to others; and then you need access back to the treatments as well. Next slide, please.
We're almost done. Our next steps are, really we are in the process of a vast expansion of these action sheet materials. We've already generated them for a range of non-S allele-associated hemoglobinopathies. We're developing them around the concept of family history. And we're looking at them around transitioning health from pediatrics to adult care. Most adult physicians, you know, a pediatrician hasn't seen some of these rare genetic diseases in newborn screening, you can be sure that the adult institutions have not, because most of the patients before screening did not make it into the adult medicine world. Next slide, please.
And we're doing them around genetic testing. We're developing a range of these and we're targeting those things right now that are most commonly referred out of the primary care community. Things like cystic fibrosis carrier testing, screening. Fragile X testing. There are a number of genetic diseases that are much, are more likely to come from primary care than they are from a specialty genetics group. So we’re targeting those as materials that can help the primary care physicians understand, really, what the appropriate response are. Even things like Fragile X are complex now, with Fragile X tremor associated syndrome appearing in adulthood, female carriers having premature ovarian failure, there's a whole lot of new information that requires us, that benefits by us keeping these things all up to date. And we're actually planning right now a meeting with EMR companies and clinical decision support tool companies to start making them aware of what they're doing instead of going into the literature, pulling guidelines, translating them into these things and selling them. We think we could work more efficiently with them at the front end by developing them in a way that is similarly compatible to their needs, and perhaps by [inaudible] some of the proprietariness of them all. Next slide, please.
We're organizing our regional collaboratives to be much more active in bridging between all the various communities. We're growing out these long term follow-up programs. And all of this data collection activity is coming into evidence bases now to help us really understand whether everything we're doing is right and whether we're doing it as best we can. Next slide, or maybe no slide? Thank you.
>> John Glaser:
Thank you, Mike. I'll give you a couple moments hire to catch your breath. Any questions or comments for Mike or Marie?
>> Andy Wiesenthal:
This is Andy Wiesenthal. First of all I'm a pediatrician so this is just stunning. I would have loved it when I had my patient with Maple syrup urine disease. I’m wondering if you are aware of the clinical document architecture standard?
>> Mike Watson:
I would not describe myself as --
>> Andy Wiesenthal:
Well, and that's good, that’s fine. There is a standard for, even if you do nothing else, and others can help you with this, I'm at way too long distance for you, but there's a standard way that's approved by the HL7 committees for putting documents into a form that can be imported into electronic health records and moved between electronic health systems. And that's an available standard now. And so many of these documents are very suitable. They're brief and they have clear, definable categories. You've done a beautiful job with that. And it's just ideal for building them as CDA to be documents. Clinical data architecture. If you can do that, I think your conversations with the vendors and others like them will be very smooth.
>> Mike Watson:
I appreciate that. I am familiar with the language. I just try not to speak it for fear that somebody thinks I know what I'm talking about.
>> Andy Wiesenthal:
I know exactly what you mean.
>> Mike Watson:
No, but that was our intent. And we did these as much out of guilt as anything, I'd say. Moving towards an expansion of newborn screening. That if it was dumped out on the primary care community, it was going to be a disaster. So on a shoestring we have been putting these resources together. I think it's getting up to this level of visibility that we're able to hopefully better resource and identify partners in this whole process to make these things much more readily usable.
>> Clem:
This is Clem, Mike. Very good presentation. Now I'd like to reinforce the CDA thing. The FDA is doing exactly that with their, what would be the package inserts, so they can be pushed around very easily. Secondly, I just want to emphasize one thing you said in your talk that but maybe everyone didn't hear quite. The current phase of newborn screening is that you’re measuring biochemical measurements, except in hearing, not in genes. Just so everybody understands that. The second thing is, nobody reports out the numbers, the value, the levels of the results, or almost nobody except for maybe TSH. So the challenge here is getting this report out, one challenge, getting the report out the real numbers instead of just saying it's abnormal. Maybe you could elaborate on that, Mike.
>> Mike Watson:
I agree that that's critical information and should be part of the system. Whether it's always useful to the primary care provider could be arguable, I suppose. I think they could learn from it, I think they clearly need that second level of information that interprets what that quantitative information means. But I think there is a lot of work that is greatly benefited by having that bottom level of detailed quantitative information that can be attached back to reference ranges. I’ve, actually, in the last week, getting ready for this, I talked with a number of these, of these software companies that are developing the materials that the newborn screening programs are using. NeoMetrics is a company probably does this for about 25 of the States, roughly 20 to 25 of the States. PerkinElmer also makes this available through their newborn screening systems, I guess I’d call it, however they're much more deficient in long-term follow-up and any follow-up information being part of their systems. And then many States build their own. California went to Deloitte Touche and built theirs. Deloitte Touche pondered going to NeoMetrics as a subcontractor to do it for them, but ultimately they did their own. And a number of States have done their own. So I think you have this mix, of all those I talked to, they all agree that if we can come to some really a national standardized language, international standardized languages, that they can built that into the systems. They can build a crosswalk between the languages that the State has decided is the way they're going to describe a certain thing that hadn't been standardized previously and crosswalk it back to those national languages. So that at least for those States that are willing to make the rapid transition, at least we have that linkage in place. So I think there's a lot of willingness to go down this, and a lot of interest, I think, in there being standardization in the programs. If nothing else, it would greatly help some of the case definition, the standardization of a case definition of what is a positive and allow us to really get to compare programs, because there are some that operate the same testing with a two percent positive predictive value as compared to another State that might run that same test at a 40 to 50 percent positive predictive value, which has huge implications for when you begin to spend money on different steps along that diagnostic algorithm.
>> Clem:
So you're saying the labs could fairly easily send out the numbers that the machines make in a standardized way, especially with the fact, and also with the interpreter thing which is now all that's reported.
>> Mike Watson:
There sense is that they can. PerkinElmer started with a pretty firm yes and then they wanted to have a meeting. So they're meeting about it right now. But I think they can be inspired.
>> Betsy Humphreys:
This is Betsy Humphreys. I know many of the people on this call or we've met in other rooms have heard me say that if these big lab companies need encouragement to help with this by actually setting up systems to report out the values, then that would seem to me that that's a perfect job for the Secretary and his overarching AHIC to be the one to encourage them to step up to the plate and do this.
>> Joyce Mitchell:
This is Joyce Mitchell from the University of Utah. And we've done some studies where we were looking at the way that a fair number of the genetics tests were reported and stored in the electronic medical record today. They're just not stored in any way that is computable or that has the complete results in there. Most of the tests, including the newborn test, are reported back in just a format which is a text statement, positive, negative, with some implication of what it means. But if anybody ever wanted to go back and figure out the State lab, and where it is that they had set their cutoffs, and what the specific data were for newborn screening or for any of the other genetic tests, they can't get it. They would have to go back to their laboratories themselves that did the testing and lots of times that's very hard to track down, certainly for the primary care docs, it would be hard, if they want to, but even for the geneticists, who probably need that data more, it's difficult. So I would put in a big, something that would come right after what Betsy said, and to have a recommendation that the tests actually be put in an elemental form into the laboratory data, probably a special laboratory subsystem, so that they're stored from that point on. They don't have to be repeated in the future. And you don't have information lost along the way.
>> John Glaser:
Let me check here, Greg, shows my lack of knowledge of the protocol. I think the types of work that Mike and Marie have described obviously don't quite fit into the use case that we are fronting here for a wide variety of reasons that have been covered. Are we in a position to return subsequently to AHIC with further recommendations along these lines?
>> Greg Downing:
Yes. That’s a good point. I think now turning a little bit to the work of this group potentially, should you decide to engage it, is what is it that's sort of within scope here, and more specifically addressing, as we've just completed a recommendation for a use case along the lines of genetic test in electronic health records. Again emphasizing, the focus here is facilitating information exchange and interoperability of health records overall. First of all we've had a number of discussions with some of you, and also internally with ONC, about how the use case prototypes of the current recommendations that are being taken forward to develop the prototypes and the use cases overall could potentially address some of these issues for looking more longitudinally if there is a need because of the different pathways that this information flows, there could be some areas here that would potentially look at alternative use cases.
I should mention we've had discussions with the Population Health Workgroup, formally known as the Biosurveillance Workgroup of AHIC. And in prior discussions on a population health area, a newborn screen has also shown up on their areas as well. So if the action coming forward from this today is to further develop this through a subgroup of this Workgroup to further develop the needs, potential areas that are actionable for the Workgroup itself and to potentially take forward to AHIC, we would recommend to you that several of the representatives from the Population Health group that are associated with public health laboratories and State and local government bodies that are represented on that group also be engaged in this subgroup.
So I think what we would ask for from staff is some discussion here about how the Workgroup would want to approach this, the subgroup section of the genetic test subgroup that has been working on the prior recommendations may also want to engage in this. But I think the issue coming forward for the chairs here and the rest of you is to engage this and define specific actionable areas that may not know whether those are coming specifically as AHIC recommendations or back to HHS. But that's probably the option to expend.
>> John Glaser:
Okay.
>> Janet Warrington:
Could I ask a question? This is Janet Warrington. And this is a question for Mike, but I think it's relevant to this discussion with respect to the future that we are envisioning. Mike, do you see a future where there will be greater alignment between the State labs and consensus on what constitutes an actionable positive screen and potentially more testing, genetic testing incorporated in the newborn screening?
>> Mike Watson:
Yeah, I think both of those are likely to occur. I think the thing that will drive the States towards more uniform case definition and performance is going to be collecting the information that actually allows other people to look at how they're performing. And I think that might underlie some of why we don't get the kind of detailed information that's useful. I think it translates out into those huge differences I mentioned about positive predictive value. So, yes, I think there are things that would take them down that road ultimately. I think genetics is moving into the programs. Cystic fibrosis screening is commonly an immunoreactive trypsinogen followed by a DNA test or by another IRT assessment. In the hemoglobinopathies it's not inconceivable at all that there's a lot of molecular diagnostics integrating into the programs. I think it's really, as the technologies become easier and easier analytically, the ability of the State labs that may not have had much experience in these technologies increases. And I think the more they can do to minimize the number of kids they send out into the world who are going to come back as negative, then the better the programs will function and the more we are going to reduce the back end cost, which is really about following all the kids who are screening positive. So I think there are various places where genetics can fold right into the program. So as a front-end assay I have never been all that sanguine, because there’s the variable expressivitiy. I have always favored some of the functional assays like the tandem mass spec that tell you, based on a metabolite, that a biochemical pathway is not working right as being a more direct way of asking the question than the mutation itself. But as the cost of mutation testing goes down, I can clearly see where there will be, and where penetrance is high and variation and expression of the condition is fairly tight, that they could be very useful markers.
>> Marie Mann:
Janet, this is Marie Mann from HRSA, and just specific to that. There is a venue by which some of this discussion is taking place and that's the advisory committee on heritable disorders and genetic diseases in newborns and children. In fact, there is a subcommittee as part of that advisory committee that's beginning to look at some of the practices in the laboratory and moving toward some uniformity in how the different labs may approach screening.
>> John Glaser:
So I guess the question on the table is whether there would be folks who are on this workgroup who would be interested in working with Mike, Marie, and others, [inaudible] from Greg and Kristin, to explore this topic further. If there are, which would be terrific, we would look forward at one of our upcoming phone calls to hear back from you about what we ought to do, whether it's an AHIC use case recommendation or some other set of things that might be done to facilitate the work that Mike and Marie have described.
>> Betsy Humphreys:
John, this is Betsy. And maybe Clem wants to weigh in on this, as well. I mean there really seems to be some very specific things that need to be done to fix this situation like moving ahead with the national agenda to have these actual numbers reported out of the thing. And I don't, I mean everything is more complex than you think, but it just seems like if we could take advantage of whatever we have in front of us here as this process to move that forward, that would be a very good thing.
>> John Glaser:
I would agree, Betsy. And I think part of what would be interesting and shows my lack of knowledge about how the sort of Federal Government and other organizations work is what's the best way to do that? Whether it's AHIC recommendation or some other set of things that we might, speak to your point, some quote quick wins here.
>> Greg Downing:
I just want to note, John and Doug, that we have made a preliminary presentation to the Workgroup that Marie mentioned last fall just as an overview of what the personalized healthcare initiative and this Workgroup is about. And one of the discussions that we think we can facilitate is more close interaction of that advisory body with the efforts of this one, should this issue move forward.
>> John Glaser:
Doug and Greg, and Kristin, how about if we rather than try to take a straw poll here, if you all would, Greg and Kristin, follow up with folks who are on the committee to see who would like to spend some time looking at this. And then we can work to find a time at a meeting or call up ahead to lay out some possible ideas about how to facilitate conversations, advance certain quick and targeted areas. Given the conversation that surrounded the presentations, it sounds like there's a fair amount of interest in both the topic more broadly but also helping some of the specific areas that would help the cause here.
>> Doug Henley:
I agree with that, John. I think that that process probably will work very well for our next conversation, as well.
>> John Glaser:
Okay. Does that make sense to folks on the phone?
>>
Yes.
>> John Glaser:
Terrific.
>> Doug Henley:
Okay.
>> John Glaser:
Let's move on with our next presentation. Thanks to Mike and Marie and others for a great presentation about the value of information exchange as it relates to newborn genetic testing and to the members of the Workgroup. And hopefully by our, no later than our meeting of September 17th we'll have some follow-up on that, as well.
On our next topic, which deals with the area of pharmacogenetics, again, we want to think as a Workgroup about how we enable interoperability of the research component of this area of medicine, linking it to electronic health records, perhaps to things such as e prescribing and clinical decision support that again ultimately will improve the quality and safety of healthcare. And Federico Goodsaid is here today to give us the presentation on this topic. Federico?
>> Federico Goodsaid:
Hi, thank you very much. Thank you for inviting us to talk about this topic today. First of all, I need to thank the revised Warfarin label, the Wall Street Journal, the Washington Post, and everyone else who has been talking about what information for sure we would like to see in a patient's record. And today I would like to have a chance to discuss something about a big chunk of work that we do at the FDA regarding exploratory that we do not expect would end up in a patient's record. And you're going to be hearing this over and over again today. And I hope that by the end of the presentation today, you will be able to go back to the document that my boss, Felix Frueh, has been writing with you, and just press select all and delete and start from scratch and we'll be all right.
[laughter]
Today I'd like to -- Excuse me?
>> Doug Henley:
I think we're getting some background conversations Federico. So, again, if those of you on speakerphones or on cell phones, please mute those so that we don't get the background noise.
>> Federico Goodsaid:
So the agenda today will cover two sides of the story. The first one on pharmacogenomic biomarkers, what they are, and I understand that I cannot project the video, so we will hear some of the audio, 90-second video on irinotecan drug safety that was issued by the FDA a month and a half ago. Secondly, how their identification can be shared with the FDA through the voluntary genomic data submissions. Third, the scientific standards in the measurement of pharmacogenomic biomarkers that we have been able to develop as the result of our work with the voluntary genomic data submissions and that are synthesized in the companion guidance for the pharmacogenomic guidance. And fourth the biomarker results we would want to include in EHR. And this will be a very nice model to look at. This is a table of valid genomic biomarkers in the context of approved drug labels that is on the FDA genomics Website, and that, as you will see, can serve as a [inaudible] point of reference of how we can start looking at lists of biomarkers that we could anticipate might be measured in the near future.
We will also discuss today a little bit about biomarkers in which this information should be collected, the content of individualized reports, the annotation and linking of biomarker information with clinical data, eHR versus eHR plus genomics and perhaps we can talk after the presentation of the next steps.
Pharmacogenomic biomarkers that we work with include both mutations, the next slide, please.
Pharmacogenomic biomarkers that we work with include both mutations that are often known as single nucleotide polymorphisms, as well as differentially expressed genes. And one of the main questions that I think comes up often is how do you translate a definition such as this one, differentially expressed genes, into something that a patient or a physician might be able to understand? The FDA has taken it upon itself to generate a safety video that is able to show the, that is able to show --
>> FDA Video:
Everyone responds differently to medicines. The dose of a drug that cures one person can be ineffective-or even toxic-in someone else. Although many non-genetic factors play a role in how an individual patient responds to a particular drug treatment, many differences in drug response come from genetic variation. Individualizing a drug's dose based on pharmacogenomic information can help a patient get the most benefit from a drug while minimizing side effects.
Take the example of treating colorectal cancer with a chemotherapy regimen that includes Camptosar, irinotecan. A gene, called UGT1A1, produces an enzyme that metabolizes Camptosar. Variations in this gene can influence a patient's ability to break down the drug.
About 10 percent of the North American population has a variation of the UGT1A1 gene that reduces their ability to metabolize Camptosar, leading to high blood levels of the drug and a higher risk of toxic side effects. If these patients are given standard doses of Camptosar, about half of them will develop severe neutropenia, which can be fatal.
A test, called the Invader UGT1A1 Molecular Assay, can identify whether a patient has the genetic variant affecting the metabolism of Camptosar and thus would be at higher risk for developing severe neutropenia. The drug label says to consider lowering the starting dose of the Camptosar for those patients found to be at high risk.
You can get more information on adjusting dosage levels for Camptosar by going to our Website.
>> Federico Goodsaid:
Okay. So the media which is accessible at that link that shows at the bottom of the slide allows anyone to get a very quick look, 90 seconds, that's all it took. A very quick look at what is it, what the problem is, what the biomarker is, and what the test. And to hopefully be able to get everyone on the same page as to why it's important to do this testing. We can go beyond that, beyond a video like that, which is part of the educational component we’ll be seeing here, by looking in more detail at how we can come up with lists of pharmacogenomic biomarkers to be used for different kinds of therapeutic treatment.
The next slide shows, the next slide, please, shows the experience we've had with the process that has been described earlier here called the voluntary genomic data submissions. This process we have had a chance to look at what different pharmaceutical companies and different clinical groups are working with in terms of novel biomarkers related to different therapeutic agents. In VGDS we have seen a world of biomarkers which we may or may not have known before but which we urgently need in drug development and regulatory review, and we've also identified the need for consensus in this field where the technology does not stand still. I emphasize this because when we use the term interoperability, I think in every other sense it has been discussed here, it makes a lot of sense. But when you go from a research or exploratory application to a diagnostic application, let's not get carried away. Let's make sure that we understand that platforms used for research or exploratory purposes have constraints, technical use, et cetera, et cetera, that may not be related at all to platforms that will be suitable for diagnostic applications. Therefore I am not anticipating, and I'll try to convey that to you today, that it's high on the list of things to do to find a way to have these platforms necessarily always talk to each other. These things are apples and oranges most of the time. And we have the knowledge of 40 examples that we can talk about regarding at least the part about exploratory biomarkers and the platforms that support them. Next slide, please.
This slide shows what the VGDS process is like. You start with data coming out of a company that is submitted to the Agency through the voluntary genomic data submission. The FDA receives, tracks, and archives that data. The data goes into the Interdisciplinary Pharmacogenomics Review Group. The IPRG sets up a review of that data and a meeting with the sponsor to provide some feedback. And at that point, a report is issued. And if the sponsor is interested, that data, which is normally confidential, can be shared outside the agency. We can synthesize what we have learned from those experiences independently of the proprietary nature of that data, and we can certainly share that synthesis of what we've learned.
We have three years of voluntary -- next slide, please. We have three years of voluntary genomic data submission. Now, this toddler has taught us quite a bit. And it's a very robust toddler.
It has, as I said earlier, 40 submissions -- next slide, please -- 25 sponsor meetings. It has already impacted new policy development and our notion about best practices in the generation of submission of genomic data. It has also helped us with the education in both directions, of our reviewers on how to do this and of our sponsors on the notion that we may even know how to do this. It has meant a new pathway for communication between sponsors and the Agency. It has really created a precedent for that. Next slide, please.
We have covered in the VGDSs several therapeutic areas, cancer, Alzheimer's disease, hypertension, hypoglycemia, depression, obesity, rheumatoid arthritis. And we have also covered several technological areas related to those therapeutic areas, the biomarkers, the devices, the microarrays, the analysis software, the biological pathways that can be used to identify how those different therapies are working.
The next slide shows that we have learned quite a bit in terms of clinical science. What is the clinical impact of pharmacogenomics marker? Clinical studies design in drug-test codevelopment: how many groups, how many patients should be per group? Next slide, please.
We have a number of general observations about these exploratory biomarkers. And I keep using the word exploratory here. The time-dependence of gene expression changes versus histopathology, for example. Sometimes these two map with each other. Sometimes the gene expression changes precede the observation of the pathology. The lack of temporal association does not necessarily mean that these changes are useless. It may mean that you have a way to in fact predict what's going to happen. You see here the genesis of the development of new biomarkers. It's the start of these biomarkers. Not necessarily their qualifications or their applications. Diagnostic applications of biomarkers require the correspondence. Even without that association there can be quite a bit of knowledge gained from those changes.
The next slide shows that for a certain type of VGDS submissions, the primary interest of the sponsor would be to discuss a particularly circumscribed or strictly methodological aspect of that data. Sometimes you will get more, sometimes you will get less ancillary data that allows you to understand what that microarray data was about. The experience reveals that even when the data is not complete, we get to learn quite a bit about how microarrays, about what microarrays are telling us. VGDS is a forum for scientists from different Review Divisions, they can get together, non-clinical [inaudible] and perhaps share information that otherwise they would not even get to see.
The next slide one additional reminder, these are exploratory biomarkers. They are not by any means ready to be put in a patient's record. And the platforms that support exploratory biomarkers support exploratory biomarkers in that context. Not as diagnostic platforms that would be taken on to a patient's record.
The review, the next slide summarizes some of the review experience that has led to understanding about some of the technical issues that need to be resolved with these exploratory biomarker platforms so that we get results that make some more sense. And these include the fact that we have been able to understand how we will be able to take exploratory biomarkers to be valid biomarkers. There are several steps in the generation of gene expression signatures associated with therapeutic outcomes. And that each of these steps can be completed in more than one way. Since you're eventually in this case generating some sort of signature that tells you that something's happened, you have to be very careful to understand what is the best way to do that. There is no consensus on the protocols, algorithms or biological interpretations of microarray gene expression data. And our job has been over the past few years to get information that allows you to at least capture what consensus there is. And that has been captured in the companion guidance to the pharmacogenomic guidance that is in the process of being issued over this coming week or so.
>> Greg Downing:
Just an anecdote. We haven't had that before last night. So we haven't had a chance to synthesize that or develop any kind of discussion around that.
>> Federico Goodsaid:
The companion guidance. We'll bring that over, yeah. The only reason why I'm bringing it up here, the next slide, please, was to show how we came about to come up with that guidance. Illustrating, again, the usefulness of the VGDS experience to do that. And that usefulness has meant, has been carried over into being able to bring together regulatory people, platform providers, academic experts, into what is called the MicroArray Quality Control Consortium. And this Consortium has been very useful to try to look at all those issues that have to do with the consensus on how you run microarray platforms, how you run exploratory microarray platforms. And to come up with recommendations from colleagues, from public forums. We had both a concept paper last year as well as a workshop. And those two were brought together into that companion guidance that we’ll have a chance to go over. Again, that guidance is intended to cover the generation and submission of genomic data to the Agency. And it is intended to clear the air with respect to all the steps that have to do with microarray data generation. So that's the part about the technical issues associated with coming up with reproducible data out of a microarray. That's what that guidance covers. For purposes, which are primarily, as I said earlier, to identify exploratory biomarkers.
Now, you get exploratory biomarkers, now the question in the next slide is how do you take those exploratory biomarkers to be qualified? And I am not going to discuss about the process, but I am going to say that it is important to get qualified biomarkers because then you can establish an independent cause and effect. You will know exactly what that biomarker measures. They may or may not be surrogates for clinical endpoints. The conclusions from these measurements are acceptable in the context, and I'm talking about the FDA, in which these biomarkers were qualified. The specific claims having to do with those biomarkers would be having to do with the safety issue or a regulatory decision.
The next slide we can get to the core issue of what we would put in a patient record. Why and when should a genomic biomarker result become part of a patient record? First of all, when we know what we're measuring. Second of all, when we know why we're measuring it. And we think that as a first step, it's worthwhile.
And I have a little bit of that in the next slide. But you can also go directly to our Website, the genomics Website of the FDA, and see the whole table. That table shows you, it's a table that has the valid genomic biomarkers in the context of approved drug labels. We are going through a very profound process by which you would qualify exploratory biomarker without touching on that. But you can do, and what we did in this case was to take a look at the information that is already present in our own drug labels. It's already there. It's already been hashed out between clinician, sponsors, and the Agency. And we collected all that information on this table. The table includes the biomarker, a representative label, meaning one drug that happens to have the context in which that biomarker is supposed to be used. I'll go back to that column that talks about the test and the number. The specific drug related to that label as well as other drugs that may be associated with that same biomarker. We define a biomarker in context. I go back to the word context here. You can have the same biomarker on this table being defined over and over and over again in different contexts because the biomarker doesn't mean much without the context in which it is applied, the package, the whole package together that we need to work with.
And on this table, we go through all the, as I said earlier, the biomarkers that have been identified in drug labels. And the context in which they're supposed to be used, the context that is supported by the label itself. The number on that column goes from one to three. In most instances, most of the biomarkers that we have from those labels have a three attached to them, meaning that the information from those measurements is important to know in considering the final decisions having to do with therapies that are on this table. But that the measurement itself is not required for the therapy that you're working with. When the number is one, what we say on that table is that those biomarkers are required. There's just a few, less than a handful at this point from this table that does that. We have Herceptin, we have a couple of more that where the label clearly says if you're going to prescribe this drug, you're going to know this result before you do that. Everything else is at this point with a two or a three indicating that that test is not required for that therapy. Information is useful, the information is valuable, but it's not required. And there are many reasons why you will get to those different levels. It depends on how much information you have at the point at which that label is issued.
So what kinds of biomarkers would we expect to get in the -- next slide -- first?
>> Lisa Rovin:
Excuse me for interrupting. This is Lisa Rovin at FDA. Does everybody understand what a drug label actually is? What goes into it and why this is the moment at which a biomarker would obtain its diagnostic or therapeutic relevance with respect to an EHR? If everyone knows, we can move on. I'm just not sure. The label isn't that piece of paper that tells you how many times to take it. It's what we use to call the patient package insert. And it has the information that's been submitted to the FDA that we think is relevant for clinical decision-making. So the ones that Federico referred to as number one, as being essential, are places where the actual drug will not have efficacy unless you've done the genetic testing. What goes into a label as telling the clinician for whom and under what circumstances the drug will be efficacious. And so you get the number one when you can't know if the drug will actually be efficacious without the genetic test. There are other biomarkers in which some percentage of the population may react badly or may not respond. It's different than a number one. Based on the information that we have, you may be a slow metabolizer, but a lot of other clinical bits of information go into the decision-making. The physician is going to have a whole clinical picture for the twos and the threes. So this is some information that the clinician may or may not find useful based on the whole clinical picture, which of course the FDA can’t have. But I want to stop here in case people want to ask Federico any questions about what that label is and why it he considers it so important.
>> Doug Henley:
Questions from the group?
>> Lisa Rovins:
Good, everyone knows about drug labels.
>> Federico Goodsaid:
That's very helpful. Once again, I need to thank the fact that the Warfarin revised labels just came out yesterday. Because that's exactly what we're talking about here. That label is on that table. And it tells you exactly what we are referring to here. The biomarkers from which the EHR information should be collected should include therapeutic and diagnostic labels; those that have been, have gone through some sort of regulatory qualification, some notion that they work the way we think they work; that have a consensus that has been captured in a table similar to the ones that I just, that we just showed today, because we feel that a table like that would be very helpful to establish a point of reference for any kind of therapeutic program.
And then on the next slide, when you get to the individualized report, we would anticipate that an individualized report would include the genotypic information, which could be the genomic sequence data, which could mean actual sequencing for genotyping. The phenotypic information, which is telling you a little bit more than the hardwired parts of the genomics of that individual, transcriptomic, proteomic, metabolomic, serum markers, histology, imaging, all those biomarkers would be included in that phenotypic information. And the clinical information with which those two other need to be linked. And I'll talk about that in a short while. One thing we do not want to get in those individualized reports is information about exploratory biomarkers. That is the, an invitation to a really spectacular Pandora's box. You need to be able to keep things separate.
In the next slide, we talk about annotation and linking of these data. So the proposal here is that there would be a lookup table with the definitions of what the biomarkers are, and the contexts in which they're supposed to be used and interpreted. So that table would serve as a guide for a physician, for example, on what is the consensus of how these biomarkers are supposed to be used and interpreted. And what I'm referring to here as context guard, which hasn't been registered by the FDA yet, the idea being that a biomarker will be applied in the clinical context in which it was qualified and no other way, that the biomarker application will default to clinical research if you start doing this outside the context from which it was qualified. If you don't go through the level of qualification required to reach a consensus on what this biomarker actually measures, then it would be really not appropriate [music]. And in terms the next slide -- next slide, please.
>> Doug Henley:
Yeah, you've got some background music. Somebody put you on hold, I think, Federico. We're off now.
>> Federico Goodsaid:
Next slide, please. The difference between eHR and eHR plus genomics would have to do with being able to sort out issues having to do with the data format and standards, you've already touched on much of that, there's no big mystery there; infrastructure needs; the cost/benefit of doing all of this; and knowing who is going to pay for it. And those four items I'm sure will be addressed by this group. And that's it.
>> Doug Henley:
Thanks, Federico, very much. Interesting presentation. Before we talk about next steps, any questions from the Workgroup? Okay. Hearing none, John and Kristin and Greg, it seems to me that if there are those, much like we did with the previous presentation, if those on the Workgroup that are interested in pursuing any aspects of this area of pharmacogenetics as it links to clinical research and HIT and clinical decision support that we might want to query the Workgroup and see if there is interest. And if so, then how do we capsulize that in terms of some potential action items as we move forward? Thoughts from the group? Greg or Kristin in particular?
>> Greg Downing:
I think perhaps what we need to do is to go back and sort of crystallize the aspects of this that address in interoperability. I'm not sure we kind of got those identified specifically enough yet for this group's discussions. And we'll certainly work with the principals on that. And I guess my suggestion would be at this point if there are a core group of people that have expressed interest in this and we work together I think to harmonize this a little bit with the overall goals of the Workgroup, I'm not sure we got those elements clarified yet today. But I think the overall concepts discussed here and presented I think are consistent with the general approaches that the Workgroup's been working on.
>> Steve Teutsch:
This is Steve Teutsch. I think this is another area where we probably need to coordinate with the Secretary's Advisory Committee, because they are in the middle of finalizing a report on pharmacogenomics, as well, and some of the issues probably could benefit from dovetailing with that report.
>> Doug Henley:
I think your point's well made, Steve. I think both on this presentation and the previous one, there are obviously a lot of folks doing a lot of work on the same topic in a lot of areas. So that coordination is very important for both of these, which I know the staff's been working on.
Okay. Well thanks again Federico for the presentation. And we'll look forward to some follow-up as staff has had a chance to work with other stakeholders to look into this topic a bit more as it relates to interoperability and our particular Workgroup.
John, to you next for next steps and actions?
>> John Glaser:
Let me tell you what I have here, and folks can obviously point out omissions or commissions along the way. Sort of running in order along the sequence of the agenda that we had, whatKristin and Greg will do is get out a calendar of conversations and phone calls, et cetera, that are related to taking the recommendations through the use case development process. And so we look forward to members of our two subgroups who contributed mightily to the recommendations contributing to the use case development. I think a little later this fall, we'll engage with HITSP about how to take the detailed use cases and iterate our way through, recognizing the immaturity of a variety of aspects associated with the standards here. So I think that's the next thing, is just to make people aware of the various conversations that are coming up and to invite them to contribute as much as they would like to.
I think, expecting that at some point we'll hear back from Becky and Amy after they conclude the drafting of the work that they're doing, and sort of bounce that off of a variety of other groups and constituencies going on. But I look forward to having your all’s work put in front of us and talk about what we ought to do relative to that.
We have the meeting coming on the 17th. We'll make sure we'll get the logistics out here. But look forward to seeing you all there. It should be an interesting and certainly an informative day.
And then last but not least to follow up on the last two presentations, and again thanks to our presenters, what Greg and Kristin will do is reach out to or invite some, whoever would like to contribute to some conversations, largely I think at this point to sort of formulate approaches and things we might do that would help advance these two areas, realizing again that there are a variety of groups that are looking at this so we don't want to obviously be redundant or at cross purposes necessarily, but let's take a look from the particular lens of interoperability and standards of what we might do here. So whenever that deliberations or set of deliberations are finalized, we look forward to hearing back from folks.
So that's what I've got. Greg, Kristin, or Doug, let me know if I’ve missed anything?
>> Doug Henley:
I think you nailed it.
>> John Glaser:
I had to nail something before the week was up, otherwise I can't go home.
>> Doug Henley:
It is Friday.
>> John Glaser:
Comments from the Workgroup committee? Terrific.
>> Doug Henley:
Okay. I think we're ready, then, for public comment, if any.
>> Jennifer Macellaro:
Sure, this is Jennifer. I just put a slide up with the phone number for people to call who have been listening over the Web. If anyone’s been listening over the phone, they just need to press star 1 to ask a question. There is also an e-mail address there if people would like to write in a comment after the meeting. I'll check back with you in just a minute or twoo.
>> Doug Henley:
We'll see if there are any public comments. Pause for a moment for that. We'll hear back from Jennifer.
Are we getting the signal, Kristin, from anybody about public comment?
>> Greg Downing:
I think we need to wait just a second.
Nothing here, Doug.
>> Jennifer Macellaro:
It doesn't look like we do have anybody on the phone, no.
>> Doug Henley:
Okay. Thanks very much. Thanks to everybody again for first our presenters for your time in presenting to us today. Becky and Amy, thank you for your time that you put in and your discussions today about the CPS subgroup. Thanks to staff for a continued job well done. And thanks to everybody for your continued great work on this Workgroup. We look forward to seeing you on September the 17th. I hope everybody has a wonderful Friday and wonderful weekend. John, any final comments from you?
>> John Glaser:
No, Doug, other than echoing what you just said. Thanks, everybody, we'll look forward to our next conversation.
>> Doug Henley:
All right.
>>
Bye everyone.
>>
Thank you.
1