 Greg, are we connected now? Yes, it's Greg Fierro. Can you hear me? Yes, we can. Excellent, I will go on. Okay, we're joined by Greg Fierro, who many of you know until recently was at NHRI in various capacities, and was instrumental in some of the issues and program planning that has led to this concept clearance. And so, and it's the only concept clearance we're talking about this council meeting, and Anastasia Wise will present it to you now. Great. So I'll be presenting to you today a concept clearance on family history implementation in the challenging setting of routine clinical care. As Eric mentioned, Greg Fierro is on the phone with us today and was really instrumental in developing this concept, which has been in development for a couple of years now at NHGRI. It's also been worked on in conjunction with a trans-NIH family history working group, which consists of 18 different NIH institutes as well as the clinical center, and we recently got representation from the Health Resources and Services Administration HRSA as well. This family history initiative is proposed to develop and evaluate methods for the collection and utilization of accurate family history data in the time-pressured setting of practicing clinicians. So we're really looking to take the family history tools that have been developed mostly in more research settings and try and tailor them more for clinical settings where there's more of a time constraint and potentially different information that would be more useful to collect. This would fund two RFAs, both cooperative agreements, one for study investigators and another one for a single coordinating center. And this initiative falls within the strategic plan domain number five, which is improving the effectiveness of health care because we're really looking at this in a clinical setting. Family history is currently a standard part of medical history collection and is frequently utilized within clinical genetic settings, especially for rare diseases. Family history of disease is also the single strongest measure that we currently have for diseases such as corny heart disease that's genomically mediated risk of disease. However, if we look at data from the Emerge Study, and these are the five phase one sites, you can see that whereas all five sites have over 80% collection for diagnoses, only two of the sites collected family history data, and even those two sites, more than 70% of the data was either missing or incomplete. So though we know that family history data is important, electronic health records may help us to collect this data in a more integrated and standard method if we develop better tools for this type of collection. As well as family history may be useful for informing our future interpretation of data within electronic health records such as genomic sequence information. NHGRI has been involved in family history activities for a number of years. We're involved in the My Family Health Portrait Tool that was developed with the U.S. Surgeon Generals Group as well as family history demonstration projects related to the use of the My Family Health Portrait Tool and have supported initiatives and incentives for including family history within electronic health records. NHGRI was also a key player within the State of the Science Family History and Improving Health Conference that was held in 2009, and Greg Farrell was an instrumental part of this. This conference led to 25 different recommendations for research areas in family history and where this needed to go in order to be able to better implement it in clinical settings. And what we've done is trying to distill those down into four key areas for this initiative. The first of these focuses on being able to collect family history information within a clinical setting where you often have time constraints on the clinician's time to be able to collect this data. Therefore, the collection methods need to be practical, efficient, and cost-effective in order to be useful in the clinical setting. We also would like to focus on strategies for integrating patient-entered data from personal health records into the electronic health records. So not just necessarily collecting this information within a clinical setting, but perhaps there are ways for this information to come from the patients themselves in other settings and then get integrated into the health record. Our third point that we pulled out from the recommendations from the State of the Science Conference was related to how and when family history should be collected and utilized in order to optimize health outcomes. This may also relate to which conditions is most conducive to collect family history information on and what questions are the most useful to ask. And then finally, we want to make sure that any family history data that is collected as a part of this initiative is going to be accurate. Overall, family history data has been found to be pretty accurate within a number of studies. In this paper from 2012, we can see the congruence of family history data with a number of separate accuracy measures. They looked at three different common diseases that are related diabetes, hypertension, and overweight and asked individuals both their own personal self-report of whether or not they had any of these three conditions as well as asking about their family history of these three different conditions. They then compared the report from the individual on their family history to self-reports from the relatives of these individuals, their first-degree relatives, and found that in general the report from the individual who was giving the family history and the report from their relatives tended to be pretty congruent with rates around 80 to 90%. They then did physician assessments of both the individual who provided the family history as well as their first-degree relatives and you can see that the congruence between an individual's own self-diagnosis and their physician diagnosis is very similar to that between of their relatives and what they reported as being their relative's risk of disease status. When you're looking at family history, you tend to get an increase in false negatives as you move further out from the individual that you're asking about the family history. So first-degree relatives you tend to end up with less false negatives than when you start working out to second or third-degree relatives. And there's also been some studies that have shown that maternal relatives may be under-reported compared to maternal relatives when you ask about family history. Family history can also capture information both about genomics and the environment. Looking at this study of over 80,000 Swedish adoptees, though, you can see that when you looked at the adoptees and whether or not their adopted parents had coronary heart disease, there wasn't much of an associated risk with the environment of their parents having coronary heart disease. However, when you looked at adoptees who had at least one biological parent who had coronary heart disease, there was an increase in risk associated with having a biological parent who had coronary heart disease in these adoptees. Family history data can also be used to potentially increase the accuracy of risk classification categories. In this study, they looked at individuals and whether or not they were being classified into high cardiovascular disease risk groups. Initially, they did a routine assessment of these individuals for cardiovascular disease risk and then looked at the standard family history information that was already available within their medical records. And as you can see, there's not much of a difference with the routine assessment between the risk assessment using the family history data or without the family history data. However, when they then added in a structured assessment, an additional questionnaire that was asking specific questions about family history information, there was an increase of 5% of the individuals being classified into the high cardiovascular disease risk group. And this is overall 40% more of the patients being classified into that group than were previously classified using the routine assessment. And this is important because individuals that are classified as being at high risk of cardiovascular disease are then recommended for interventions that could prevent disease in the future. There have also been individuals who have been pretty skeptical about whether or not using this sort of genomic technologies in primary care practice or in family history could potentially be useful. This is a quote from Dr. Berg, who was the chair of the State of the Science Conference, who actually upon seeing this study actually thought that it was a great example of how family history could be used and clinically useful in acknowledging risk classifications and potentially shifting patients into more accurate classifications. Family history information can also be used in conjunction with genetic information. As you can see, the population risk of colorectal cancer increases as individuals age. And if you look at family history, that increases individuals risk about double. When you start looking at genetic information, you can see that having greater than 13 risk alleles puts individuals at even greater risk than what was captured using family history information. However, when you combine those two types of data, you find that there is a subgroup of individuals that are at even higher risk. This can be seen as you start looking at individuals with more and more risk alleles, that the pattern continues, and it seems like both the genetic information and the family history information are both adding unique information to this risk calculation. Overall, we've seen that family history information can be collected accurately. It can be used to find information on genomics, and it can be used in conjunction with genomic information to provide additional information to risk calculations. So the overall goal of this initiative is to take current family history approaches that were typically developed in research settings and try and develop them into more streamlined methods that could be used in routine clinical settings such as primary care. We have three specific goals. They'd be to develop these streamlined methods that would all be electronic methods that can be put into electronic health record systems to then develop and evaluate risk algorithms using these streamlined methods. And finally, to work on increasing the efficiency and interoperability of the family history data, both its collection and its use within the electronic health record system. We would expect that these studies would compare their risk algorithms to either more extensive family history approaches or their current standard clinical practice, that they would evaluate the acceptability and the added value that the family history information brought to their risk assessment and do this both within the context of their standard algorithm and their new algorithm that they're developing. That they would provide information on their data sharing plans and informed consent procedures, meet current standards for electronic health records, and make sure that they are going to be both inter- and intraoperable so that the electronic health record systems can speak to multiple systems both within that electronic health record system and between electronic health record systems. And finally, that they would work to harmonize with other related research standards for data collections such as the Phoenix measures that are developed by NHGRI. Examples of applications and some of these studies might include incorporating family history data with other clinical information or genomic information that's currently to be found in their electronic health record systems. They could work on developing strategies for the entry of the family history data into the electronic health record system as this best to be done by the patient themselves, by relatives, by the clinician. Look at the economics and efficiency of family history collection as well as developing complementary personal health record interfaces that would allow the patient themselves to be able to see this data and interact with it. They could also look at LC issues such as patient perceptions of sharing family history data within the family and with clinicians, as well as more research-oriented questions such as the association of genomic variants with specific phenotypes that you might find in the family history data. As we mentioned previously, there is a trans-NIH family history working group that was involved in the development of this concept. This is an NHGRI-led working group that has been recommending ways to take the recommendations from the State of the Science Conference and recommended the development of this initiative. They provided feedback on this concept and two of the key points that they wanted to have highlighted was that they really felt it was important not just to focus on developing the tools for family history collection but to have easy ways for clinicians to be able to use and interpret the family history data that in order for this to become useful in clinical practice, it needed to go a step beyond just collecting the data to providing back information that could be used for clinical care. And also, they highlighted that interoperability of electronic health records is really key in order to be able to make tools useful in multiple different clinical settings. There are multiple ICs that are enthusiastic about the development of this initiative, but they can't commit funds until they see a draft of the RFA. Overall, we would anticipate funding four to six investigator awards and one coordinating center for this initiative, which would have NHGRA committing roughly four million dollars per year for four years for a total of 16 million dollars. And we would encourage the participation of other NIH ICs and especially those that have been enthusiastic about this from the Trans-NIH Family History Working Group. And with that, I will take any questions. Ms. Stasia, all the things where you said that things that might be able to be done with these did not include figuring out how to collect that family history more accurately, did it? Or maybe that was embedded in there. It seems like that's a problem still, right? It's not obvious that the way that most people do it is the best way? Yeah, we would expect that any study that came into this would be doing some sort of measurement to be able to determine that they were collecting the data accurately and could focus on that. That's definitely part of one of the main goals is to make sure that it's accurate collection. Okay, and can I ask another question? Maybe this is broader than just this one, but when you come up with these ideas and you make these concept clearances, how do you think about a budget? You want four to six or four to five or whatever. I have no idea how much it would cost to do this. I mean, is it five million dollars a piece or fifty, you know, five hundred thousand dollars a piece? What do you do to do those estimates? I guess it's not really calculating. It might estimate costs based off of the personnel and supplies that would be necessary in order to conduct an average project that would be responsive to the initiative that we're developing. So in this case, we considered the fact that you would need bioinformatics support to be able to develop the system for being able to collect the family history data and integrate it into your electronic health record system, as well as nursing and clinician support to be able to potentially collect the information and determine how accurate it is. And then based off of that calculation for how many support personnel and supplies that you would need, we get a rough estimate as to what each center would cost. We add in, you know, approximate amounts for, you know, indirect costs and that sort of thing, and that's how we get an approximate number. So for this being four million dollars per year, that with four to six sites, that's, you know, approximately a million dollars, each a little less. Counting on other, the ICs are the other intramural programs. I mean, not intramural, the other NIH institutes. Is that what ICs mean? I'll just forget. So that, if that, if there was a lot of interest and that came in, would that cut the amount that you would need to spend down, or you would just add to it? It could potentially cut the amount down. Okay. The other thing, I think this is something that I think will, maybe it's for a closed session, I don't know. But in general, how do you deal with, what if it was half that amount? Could you still do the initiative? And not this particular one, but I think in general, that's one of the things to be thinking about, because not that we hope that that happens, but it could, I guess, that there could be cuts. So would you lose the program? Part of this is prioritization overall. But the other question is whether you can do an initiative with, you know, different amounts and, you know, with a grantee the answer is always no, of course not, but at least we have to think about it. And I think we would need to make sure that we do still have multiple sites and the coordinating site or component to this initiative, because it is important to make sure that it's being done in multiple different types of clinical settings, along with multiple different patient populations, in order to make sure that you're developing tools that are going to be applicable across a wide variety of clinical settings. So you were going just where my question was going to go. So is the assumption here that there exists or will exist a single set of methods or approaches that work from oncology to cardiovascular to rare diseases all well? We are not assuming that there is, you know, a priority going to be one set of particular methods that are going to work. An investigator who's coming in with an application could choose to look at multiple different diseases and see if a similar set of questions work for all of them. He will choose to focus on a particular disease or phenotype and work on developing a set of questions that work particularly well for that particular condition. So you'd become comfortable with that one. I would chime in. It's Greg Furrow just to say that, hopefully there's not much phone delay here. I apologize if there is, but I would just anticipate that there would not be a one-size-fits-all tool that there would probably be several types of tools that would need to be developed depending on the clinical setting. Primary care obviously has considerably different needs for an oncology clinic in terms of what level of granularity they might need to get, but it certainly would be nice to have a suite of tools that was developed that had some degree of interoperability so that data that was collected in a primary care setting could be foundational for the oncology environment. John, did that answer your question fully? That's exactly what I would have assumed as well, and I'm pleased to see that that's the direction you'd take it to try and get something interoperable but not assume that there's a one-size-fits-all. Yeah, so I believe that this is something that should be done, but I'm curious why it would be done at NHGRI instead of something like a common phone because it seems to me it's something that applies NIH-wide. So that's one question. The other one is what's proposed here is say maybe four projects or something. I feel like that's just the beginning and how do you see this truly being implemented on a nation-wide basis so it really has the impact and the effectiveness that it should? Okay, so I can start with the why NHGRI question. So NHGRI has been involved in family history initiatives for a number of years now, so we actually do have a group of expertise here at the institute that has been involved in the area for a while, as well as we're disease agnostics, so if we're trying to develop tools that are going to be able to be applied across genomic medicine or integrated with genomic tools in the future, it does fit within our goals to be able to work within this area to develop family history, electronic medical record tools, and we do have experience with programs such as eMERGE with working with electronic health records. In part... In part that it would be different for different diseases. So we are trying to partner with other NIHICs, potentially to get additional funding for this initiative. If they came in with particular conditions that they were interested in funding, that might mean that their particular dollars would go towards a particular disease phenotype, but we're really interested in trying to get something that's going to be a little bit more agnostic and making sure that this is something that's going to be applicable across a wide variety of different phenotypes. And then your other question was related to where do we see this going in the future? Lamentation, so it truly has the impact and effectiveness that it could. So this is really designed more as a small pilot study to try and start developing these types of tools, which is something that NHGRI does have a lot of experience and is developing tools for genomic type applications. And in the future, this might be something that could then partner up with some of our other initiatives, potentially looking at applying some of this information with genomic sequence data, or moving in that direction and integrating with some of our future genomic medicine initiatives. But right now we're really trying to get the information on how do we collect this data and what additional information does it add? How do we do this effectively? I would just add too that I think that, I think it's a great fit for an NHGRI because it complements well the highly technical genomic information. It really integrates that very nicely and I think there's a lot of expertise here and the community of people who traditionally are funded by NHGRI will be exactly the community I think that can really tackle this. The other thing that I would just add is that it's also highly dependent to make this work, I think, on synthesizing it with the EMR, with the electronic medical record. And so I think highlighting that is really good and it's the kind of thing that a lot of that NHGRI has expertise in as well. Thank you. Jill? Maybe I missed this in the presentation but can you just say a word about what the state of the art is here? I mean, presumably, there are a lot of commercial efforts in electronic medical records also and presumably some of them have to deal with family history and I'm just wondering what the state of the art is. So there are some family history tools that have been developed. A number of them were developed specifically in more of a research type context so a lot of them take a long period of time to be able to fill out which might work if you have a patient who's able to be able to fill things out ahead of time and bring information in but you still need to be able to have that be integrated into the electronic health records and have information that can be fed back to your clinician in a format that's going to be useful for clinical care. So these tools are targeted at the patient and the patient entry? We're not specifying where the data entry has to be. It could be patient entry, it could be clinician entry. It could be that you're actually having multiple relatives enter data into the same system. Let's add a little bit again on the state of the art. There are as a result of the stage two meaningful use criteria, menu criteria that just got put forward regarding family history collection a variety of vendors I think that are looking at that and beginning to sort of reassess their capabilities and their commercial EHR platforms. There's a couple of features that I think will be difficult unless there's sort of a governmental effort in this area to ensure or happen. One is interoperability, which of course has been a conundrum for all aspects of EHR work. But the other is the interpretation component. Though the vendors may improve on the ability to gather structured or allow gathering of structured family history information they're unlikely to do much in the way of investment and the validation side of the way the data is collected, the user interface for example in terms of the accuracy information and they're also unlikely, at least the larger vendors in the near future I think are unlikely to work on developing and validating interpretive tools that would allow the clinician to make some clinical decisions and certainly I suspect they're a long way off from thinking about well how would you know if data integrate with this? And so this effort I think might stimulate some work along those lines with a little initial public type investment. I was going to ask whether it's possible to expand this to include a collaboration with the VA, a collaboration with the Department of Defense or whether SBIR as a funding mechanism might be thought of. I'm using an electronic medical record right now and it's okay for collecting the information it's actually quite somewhat clunky to get the information in. It will not generate a pedigree despite the fact that there are perfectly good programs that could do that. I think something that would encourage the vendors to get their act together and provide what people need would be an important thing that you guys could try to do. And we can definitely try to reach out to those groups. That's actually how we ended up with HRSA being involved is they were working on some work looking at pediatric settings and integrating some family history information and so they're interested in what we're doing with this initiative and we can definitely reach out to other groups as well. So I think one of the things that's potentially likely to distinguish this approach from a lot of the others that have been taken is a lot of them have been really one-off operations where somebody in their own silo develops a family history tool that they think works well for them. They put it out there and then the history is that there's not much adoption. I mean I think that's the history of family history tools. And I think what I would hope comes out of this is something a little different. Because it's a consortium and because there's going to be lots of at least four or five groups sitting around the table thinking about this hopefully ultimately doing their own thing but converging a little bit in the course of this that there might actually be a product that now gets a little bit more adoption. I think that's the real thing that I see is different here is to actually have a consortium that works together on this. I'm hopeful given the history that NHGRI has with consortiums actually being successful and producing good stuff that would be a key distinction that this project would bring that all of the other family history tool projects that have been out there have not been as successful because they lack that consortial approach. If you don't have Epic and Cerner at the table then I think it's going to follow the same fate that Rex just described. These guys are very widespread and the electronic medical record I use because of my medical center is Epic. And getting them to change anything is rather difficult without encouragement. I think it's an opportunity to have them at the table but just like we heard in the Genomic Medicine meetings and I'm sorry the Emerge meetings with them they're looking I think to us to try to tell them what that needs to look like and then once this group has come together and put some constraints around what that looks like they can find a way to actually build a tool or integrate it. Is this group their customer base? Because that's who they have to hear it from. We would hope that the applicants to this group would be clinicians who are going to be using the family history data which should be the groups who will be using those Epic and electronic health record systems that would be their clients. Groups like Kaiser Northern California which uses Epic. I mean you need big groups. And we have engaged those groups with some of our other initiatives. They won't pay attention unless a big group bangs on their door. Any further questions? So I understand the intent and I think I understand the intent and the plan to have this be clinically very useful. Is it also the case that it seems to me that it could be the case that from an NHGRI research perspective this could also be extremely useful because it seems to me the way the technology is moving the way everything is moving in a small number of years looking out say you know three to five years I can imagine NHGRI wanting to do pretty large studies and as some of the discussion earlier without this type of information collected efficiently you wouldn't be able to do that. Is it also... Yes and we would hope that this would feed into future initiatives and research here. We need a vote. Oh normally you do that. You want me to do that. So we need a vote. Which means we need a motion. I knew that. Is there a motion about this concept clearance? Second. All in favor? Any opposed? A couple of three. Okay. Abstain. Okay. Passes. Alright. Thank you. Very much. Yes sir. And we have a few minutes before David Page is going to join. So Greg you're still on. Do you have anything else you want to add? No thank you very much. Thanks for joining us. Bye bye. So I can... We can start. As soon as the computer is ready. Go ahead and disconnect. What? Disconnect from the phone or leave that on. No leave it on because I assume that's how David's coming in. Yeah we have somebody else who's dialing in in about 10 minutes. Let's confirm that. Oh by the way I can answer while we're waiting for the computer to reboot because I have to give a few slides to set up our next discussion. I did get some information. Our communications team got a little bit more information to answer Bob Nussbaum's question before relating to Lucia Hindorf's issue about what can we access in terms of IP information. But we can get city-state information about what's coming in and we can get some analyses about whether it's a government or private entities and I bet you educational entities are maybe not. But so I think we can get those kinds of demographics. We don't get to any sort of more specific information than that. So maybe what Bob you heard the answer from NCBI was that they are limited in what they can get can't get to specific people or a whole lot of demographic information. But that's consistent with what we can at least know. I know I've seen information about what countries and at least then the breakdown of government or otherwise.