 Okay, thank you and so we'll, for the summary we'll go to Dick Winshill-Vaughn. I hope everyone can hear me. I want to explain first how this group operated because both Heidi and Laura supplied their slides early violating most rules of academic slide making and that gave the group an opportunity to interact with them and they have changed the slides significantly from the beginning to address those topics which were basically on our list that is genomic testing, actionability, validation, and lab reports. And I want to thank both of them and the entire group because I see one of my functions is being certain that the members of our group, that is Murray, David, John, and Kim have an opportunity to make their comments with regard to these specific issues. What has come through from both of these presentations is that the EmergePGX project has certainly forced us to deal with practical issues different at each site and I'm sure that Ithacarcuil and Chris Huit would say that would agree that at the Mayo Clinic this has been a major catalyst and I want to be sure I say that very clearly to addressing very practical issues of implementation, not all of which some of us had thought through carefully so that this has proved to be very important. I want to also go out of my way to say that that PGRN Seekery agent was developed by the investigators of the PGRN, but that's the royal we, really it was Debbie Nickerson whom we heard from just before lunch and Steve Sharer at Baylor who I think with their colleagues we need to thank for the fact that Laura could say that that reagent has performed so very well. What I would like to do then at this point because we have individuals who have been working through these issues in a variety of different ways is turn to the members of our group who have already put a good deal of effort into helping to shape this particular session and it seems only appropriate Kim has gotten back to me and actually emailed me with some thoughts so Kim this is your chance to highlight any of the issues which both Laura and Heidi have done such a nice job of outlining in the context of how CIDR is interacted with these projects or anything else that you think addresses the topic we were given. Thanks Dick, can everyone hear me? Yes. There you are fine, you might speak up a bit. Okay, I just had a couple of comments. I think my personal opinion is that as Dick pointed out the PGRN-Seq projects and the clinical validation may outforce some issues and really spurred a lot of work in the area of implementation and that's an area of heavy work going on at Hopkins as well and I do believe projects like this, there's a huge amount of work to be done in the area of implementation which has been discussed a lot this morning, implementation is a very complicated process and my personal opinion although I obviously generate discovery data as an outside advisor here and not promoting my own work would be that an area for Emerge 3 to continue to contribute to really would be more in the area of implementation of these types of clinical results than in discovery but that's my personal opinion but that's what I'm here for. I think that's pretty much all I had to say. And we wanted to give you that opportunity. Murray, you participated in these calls and comments from your perspective on the topics that we've been assigned. So one of the things that concerns me is that after it's specifically for PGRN but it could be for any kind of testing that we do but one of the things that concerns me is, you know, we have developed this test. Information goes into the electronic health record along with clinical decision support and as we know, as we learn, the raw data or the data is sort of c-attached or something in some kind of other file. So with large numbers of people going from, you know, having medical care at one institution and going to another institution, the issue of portability and standards and things like that I think are also very important and that with, you know, since Emerge we have different electronic health systems which I think we're in a position to actually see what happens when a patient moves from one institution and health care organization to another. How do those reports travel with the patient? Let alone, you think about new information that pops up and how to generate that. What do you do just in kind of a lateral move with this information? David, any comments that you want to make and we'll have John next before we open this up to everyone who might have something to chip in? I don't know whether David, are you there? Maybe we'll move on to John. John, well, both of them participated in the calls and maybe we can go back then and ask anyone in the group to comment on either their experience because there are a large number of the Emerge sites that have been vigorously participating in all of these projects and are facing the same problems and we'll go back to our chair to guide that discussion. Okay, the floor is open for comments or questions from any of the presenters, reactors, or summarizers. I think a concern that a lot of people have with returning results is the CLIA sample issue and we face with the PGRN results that we want to return to are CLIA certified and many of our bio-repository samples are not CLIA certified. Some sites solve this by recollecting people. We solved it by doing the testing in our bio-repository samples and recollecting the people who we want to return to and I think if we're going to look at rare variants, it's not necessary to have 50,000 CLIA appropriate samples because we're not going to return results to that many people but we need to think about when we return results, how will we be returning CLIA results to those people and we're using bio-repository samples that aren't CLIA certified for anything else? So it sounds to me like that with something just empirically sprang up that some people made the decision to do it one way and some recollected and some just focused genotyping panels and some made their pipeline, their recent pipeline and this seems to me to be a wonderful focus of a paper on the alternative operational and cost implications of these various ways of taking a research result and making it clinically palatable. This reminds me of the old conundrum where a sample was collected for a rheumatological study and I couldn't use it for a cancer study or a derm study and you couldn't use it for something else and now we're getting broad biobank approval for a lot of things. So when a study goes into such a biobank, it may be for a study of a rare variant and you can go back and collect that one but the great mass of those may eventually be used for pharmacogenetics and then those are results which are clinically applicable to a substantial number of people. So I think genetics is not unique but has a very high density, a high probability of running tests in research labs which could also be run in clinical labs and we're all used to running things in non-clear approved labs and collecting our samples that way whereas in a cancer chemotherapy trial you'd never think of grabbing a serum sodium on your own or running it in your own lab, you just use the clinical lab. I think this is David Carey. Can you hear me? My audio wasn't working so I came down the hall to Mark's office. So as a group if we can move in that direction in the future of collecting things that are clear approved when they will have potential broad applicability a year from now or two years from now or five years from now that will probably be a very good thing. That was one of the points I was going to make because we have a lot of research samples that are not suitable for clinical testing and that was a bottleneck for this project so one of the things that forced us to do was think about how to get around that so one of the changes we're making is to convert our research biobank to CLIA so we don't have that issue in the future and the patients are consented for broad use so I think it will give us a lot of flexibility in terms of how we use it. The other challenge we had was we had to reevaluate our research consent because it did not allow us to put data into patients medical records so we've now changed our broad consent that gives us that ability. The other, I think the beauty of this project was that it really has many moving parts that have to connect and these are things that aren't usually connected in our health systems and I think Dick used the word catalyst. I think we use this as a catalyst to drive some change here. I mentioned a few, also some changes to the way we work with our molecular diagnostics lab and others. David, I'm glad you made that point. I used the word catalyst but in some ways this is a disruptive technology. It's not the business as usual and certainly in a big place like ours or yours we found that we have to adapt and I think I'm just saying what you're saying. Absolutely and I think it's been an evolving process for our molecular diagnostics lab. They've been very supportive but I think it is a change in how they do business. This is Julie Johnson. I think after this conversation I'm a little confused and I really beckon back to the original topic of balancing discovery versus implementation because to me there was some implication that implementation meant that clinical results were being reported into the medical record and yet it sounds like the very, very vast majority of genetic information generated and emerged is not able to be placed into the medical record. This might be a sort of distraction area but I guess I'm curious then where the implementation piece fits if the results aren't clinical. Julie, just to clarify, those of us that are returning results in the electronic health record those are all being returned under a clinical research protocol. They're not being transitioned into a peer clinical result. So while they're in the EHR they are still a research result that is where the patients are consented for return within the clinical setting and also consent to have that result placed in the EHR and to persist in the EHR. Yes. And then clinical actions or decision changes and approaches or whatever can be made based on that even though they weren't generated in a clinical, in a CLEA environment? Well the results are generated in a CLEA environment. I mean that's the point that we're making. You can do a research study and clinically return results and use a CLEA approved laboratory so you can actually use the results clinically so that's how we're actually doing it and while we are building decision support around certain use cases particularly in pharmacogenomics it does not restrict clinicians from using the results more broadly because they will be in the electronic health record and they have been tested using clinical grade studies. So maybe I misunderstood Laura's presentation because it sounded like most of the data were not being generated in a CLEA environment. Is that not the case for like the PGRN-Seq? The PGRN-Seq data is generally not being generated in a CLEA environment but then they're validating certain genotypes in a CLEA environment for return in the EHR so every single one of the participants is having some genotypes validated for return but yes the PGRN-Seq sequence data is generally not being generated in a CLEA environment. Is that helpful? But it could be. But it could be. Yeah. And really at the Mayo Clinic it's in our Department of Laboratory Medicine in a CLEA environment that the sequencing is being done. Could I make a comment? I don't know whether people can hear me. It's Susan Wolff at Minnesota. Yes. A lot of these issues are also being encountered in the CSER studies as probably some of you know where some of the consent forms are actually asking permission to incorporate even CLEA-generated research results into the EHR. So I think there's a consent issue that's emerging because of the consequences, the potential consequences of this information going into the EHR. So I think also looking at what's the consent format being used for moving that information would be very helpful. That's the next panel. For the discussion and especially talking about gene insight and integration of gene insight into systems into the EHR systems. One of the questions that comes up is sort of who has responsibility for the decision support when the variant status changes. One approach is that the lab has responsibility. They figure out that there's this new variant and then they're trying to track all patients and tell patients. The other way to do it is to say there's a way of sharing knowledge about the new variant and the EHR system uses its own infrastructure and decision support to do that. And we see a conflict in that. And actually we'd prefer not to have the lab do it because what happens then is you start exposing different ways of doing decision, you know, decision logic is dependent on each individual system that's generating that knowledge in it. It's kind of just a system architecture question, but I don't know what experience people have had and whether you have the same feeling or it's fine with you if the lab does it and then you figure out some unique way to, you know, do genetic results different than you do all kinds of other decision support in the system. This is Heidi. I can expand a little bit on how we've implemented that. So the alerts that we send go directly to the physician and then it is the physician that is responsible for relaying that information to the patient if they feel it's appropriate. In addition that because all the genetic data is structured, the EHR environment itself with, you know, approval from the makers of it can implement additional clinical support rules on top of the structured data. So, you know, it is integrated and it's the physician's responsibility to return that data. We do allow them to either sign up for proactive alerts or they can simply only get the information when they go into a patient's medical record because they're seeing the patient the next time. Is that consistent with other environments in the EHR for that type of information? How does the notification to the physician happen? Is it happening through the common mechanism of the EHR or are you just sending an email or how does that actually happen? It is being generated by the gene insight software system which is fully integrated into the EHR. Now it isn't a part of the main EHR system and actually when we implement EPIC there'll be just an interface between the two. It'll be a single sign-on but it is a separate software system that's generating those alerts that then get emailed to the physician. That point was not that initial analysis but what happened three years down the road when you discovered that somebody is the star. We've never seen that before but now we know what it means and so our clinical interpretation of the genotype changes three years after you did the original analysis. Who owns that re-notification? Yeah, that's a great question and one of the ways we've set it up is that the clinic sites have a system for the whole clinic and that way if care changes from physician to physician from the original ordering resident, let's say, to a physician later that clinic is getting the information and they often designate a genetic counselor who manages it. We have talked about trying to manage who's responsible for what within the EHR environment, although it gets a little complex. We also have a new study that we're starting to look at to potentially allow patients to know when alerts are sent on their genetic information so that they, if there's no physician any longer, the original ordering physician is no longer caring for them that they can find another physician to manage that result change because these things do happen years later and they already have. So that is something that we continue to try to explore how best to support. This is Mark. Justin, there's no established standard of care more broadly about who owns that process. That would be a potential thing that could be included as a study question for an RFA, which is to study both the process of how re-annuation would take place but also to study the also related issues relating to re-annuation and update. It seems to me there's a precedent here for the interpretation of microbiology results, which on a shorter time frame often change. They'll be preliminary and then they'll be revised. Sometimes the revisions replace the initial interpretation in such a way that the clinical implications are quite different. So it seems to me territory that genomics has to make up or envision as never having occurred before in clinical environments. Susan speaking, I think though to support the prior comment there remains an enormous ethical and legal controversy about responsibility for re-contact and updating. So I would support the idea that it really needs further work. Dan, I think the other difference is that in the microbiologic example, and this is also true in pathology, that you can really fairly cleanly define an episode of care associated around those things that does have a relatively delimited time frame where those things are less clearly defined at the present time in the genomics. So I think, again, whether this is a place where genetic exceptionalism is really relevant or not, I don't know. And part of the analysis would be to say, how is this the same or different from those examples? But it is not something that there's clear consensus on. The other thing with microbiology is, I know that we've had to implement special decision report rules for TV cultures because by the time they come back, their page is off the screen and people miss them. And you know with micro, it grows in a certain amount of time or you've done. You don't have the case of the bottle grew three years late. So you have a rotating function there. Sure. Okay. Well, that's been an excellent discussion. And so what we'll do... Is there anything else? Are there any other comments or questions? Okay. So we'll segue from testing to informed consent to education governance and our presenter is in the room here in Sunny Bethesda. The blue on the top and white on the bottom.