 We're going to start the clinical informatics for varied electronic health record systems session now Where are we're going to have our first talk for the state of science and gaps sandy arison from partners health care? Thanks, I believe that Clinical informatics is going to be increasingly critical to the point of potentially becoming a roadblock to the adoption the proper use and ultimately the efficacy and overall impact of genetic medicine and It Compounding this making it more complex is the fact that clinical informatics is such a broad Field with many different components that have to evolve in order to properly support the use of genetics So what I'd like to do today Let's cover parts of clinical informatics that I think are most relevant to genetic medicine and share my views on Where those parts are in their evolution in terms of moving towards providing our community the support that we desire so the foundation of Clinical informatics is the movement and acquisition of structured data There are amazing things happening in the world of NLP But I think in order for us to build a durable foundation for the kinds of support that we're after We have to get more structured data moving through the clinical ecosystem And that's really hard in the field of genetics It's hard because we have providers often ordering from laboratories that are in different organizations Using systems from different vendors than the provider uses so we have to push increasingly structured information across these inter organizational boundaries and that in turn is hard because developing interfaces that are generalizable broadly Implementable and reproducible is in effect a multi-dimensional Problem we need interfaces that are capable of transmitting Data for many different types of genetic tests many different types of data about each of those tests Many different scopes of data Potentially that are relevant to transmit and that's just the content that needs to be transmitted mechanistically we need to establish technical Legal and regulatory security context need to establish Actual methods of data transfer and formats for data transfer and for fields that are particularly important to harmonize and structure Well, we need to define the ontologies and the codes that we will use so effectively We need to cover all of these different areas in order to develop a fully robust generalized interface between a laboratory and provider some in some cases for Situations where the lab and the provider are using it systems from the same vendor It is possible to cover many of those to the best of my knowledge No example where you can cover all of these bases for situations where the lab and the provider are using different Vendors as far as I'm aware there's no situation where even a majority of these different bases are covered Which means that what we have is a situation where those labs when they Establish these interfaces to the providers There's a cost associated with that which becomes a barrier Which means that not as much data moves through the ecosystem as we need to move in structured form Which in turn means that the whole rest of clinical informatics is less powerful than we need it to be that doesn't mean That there aren't examples where clinical informatics is currently protecting patients in great ways There are examples doesn't mean that as we move forward we won't have Increasing a number of use cases where that kind of support won't be provided it will But it does mean that the scope of the impact in terms of the number of people that we can assist the number of Communities that we can assist the number of clinical scenarios that we can assist the number of places that we can assist is severely limited By this lack of movement of structured data So progress in this area is potentially Valuable in a very generalizable way So once data moves from the laboratory to the EHR there really two different scenarios For how it can be managed and processed it can be placed directly into the EHR Probably in the most straightforward way to handle it the disadvantage though is you're limited to the structures and the EHR's ability to receive and understand And manage that genetic data alternative is to interject a Ancillary genetic system that actually receives the data from the laboratory Usually models it in a much more detailed way event just sends to the EHR the data that the EHR is capable of Managing well both of these in use today And you know can be implemented robustly It's also worth thinking about where Clinical interpretation of variants occurs at present it occurs in many different places Which leads to a scenario where it is difficult to ensure that The varying interpretations are maintained consistently across the ecosystem in Solving that problem requires the establishment of interfaces between the different sites where varying interpretations are maintained Multiple barriers to that technical barriers barriers in in terms of places wanting to keep data proprietary But where these interfaces have been stood up they prove and helpful So once data is moving it has to be displayed in the clinical environment again Multiple different ways to do that as Kelly mentioned You know baseline scenario you have a PDF or scan facts coming in getting placed into the EHR It's probably going to be placed in a generalized section of the EHR Maybe a pathology section, but not managed as you know as an independent genetic result if Structured payload is sent into the EHR then you can use whatever native Capabilities the EHR has for managing genetic data If an ancillary genetic system is involved then you have the ability to establish a single sign-on transfer patient context Interface which essentially allows you to project displays into the electronic health record That seamlessly integrate with that environment, and then you can version and improve those displays independent of having to version the EHR itself This point I did just want to mention that user interface design is an extremely difficult challenge throughout all of IT But in particular in the clinical realm it is very important It's important for ensuring that Screens that are presented are intuitive and that when data when people look at data in those screens They understand it properly, so they apply it properly, but it also is a multi-disciplinary Area and therefore it can be expensive to pull together the different disciplines that you need to in order to develop good user interface Designs, but I think it's something that in our area. We really have to focus on more Okay, so we're displaying the data now in order to do more complete forms of clinical decision support Ideally we federate data together across institutions through networks and repositories There are multiple different ways that that can be done There's the model that ClinBar uses that eMERGE is going to use to set up its Deidentified repository Multiple different entities sending data into a single repository. Maybe a public repository. Maybe a controlled repository Other alternative is peer-to-peer networking Both of these have advantages and disadvantages I think that both are necessary in order to build out the ideal ecosystem But in addition to the networking process, there's the question of the content that is shared I think we as a community are doing an increasingly good job of sharing genetic knowledge There are some examples where aggregate data independent of cases are being shared to assist clinicians But I think that the real holy grail here is to find ways to share Individual case data that can drive more complete forms of analysis and clinical decision support with an eMERGE Looking at ways of sharing genetic results phenotypes indications, but to truly get to what we want here We really want to get to outcomes and that becomes incredibly difficult for multiple different reasons one of which is the Deidentification and security context that you need to Set up in order to make sure that you're appropriately Securing and protecting the patients within that these these repositories This is an area where it will be incredibly exciting to see what comes out of the Precision medicine initiative and how that Helps the clinical informatics field in general So all of this is focused on leading to clinical decision support and really assisting The way that clinicians make decisions two different kinds of clinical decision support There's passive decision support where you're essentially augmenting displays that clinicians navigate to with more functionality a number of people here KC mark others have been real drivers of the creation of e resources info buttons that enable clinicians to drill down on from information in the EHR Often in patient context specific ways to get more information about the data that they're seeing it's possible to add annotations To screens in the EHR, which can be helpful different calculators can be hooked up to help with risk assessment dosing And an area that I think many folks here including myself are extremely excited about is population based analytics Potentially actually real-time population-based analytics that could be integrated into clinical displays Give information like if I was to run this genetic test on this patient What's the likelihood that I would find a variant that would be significant, you know How is creatinine likely to trend in this patient with or without intervention if I release this patient today? What's the likelihood that they'll be readmitted lots of amazing things that people are beginning to start to look at? doing in the EHR environment and then you have active Clinical decision support where you actually are going to proactively interrupt the clinicians workflow With information that you think is particularly important perhaps the lowest-hanging fruit in this area has been pharmacogenomic interventions These are being so digitizes working actively in this area a number of different places have these stood up and working today where Clinician orders a drug if there's no evidence that the clinician has ordered a needed Pharmacogenomic test before ordering that drug the system alerts them if there's evidence if the result from a pharmacogenomic test Contra indicates that drug order system also alerts them specific areas where where those are being stood up We've found knowledge update alerts to be extremely Useful so situations where a variant has previously been identified in a patient new Information is learned about that variant that could potentially be impactful to that patient's clinical care sending the clinician a proactive alert and Caesar and digitize the Caesar EHR Working group and digitize are working together to create an example of a risk-driven reminder So in this case working on identifying which patients have a genetic predisposition to Lynch syndrome and therefore You know should have reminders for colonoscopies go every two years as opposed to as opposed to every 10 So I think overall in this space We really are just at the beginning of the beginning and yet already material ways that clinical informatics is Helping patients. I'm helping to protect patients. So I think it's really exciting to think about where we can go In this space going forward. Thanks a lot Thank you, Sandy. Next we have Casey Overby from Johns Hopkins