 So, I'm going to start out just by first defining clinical informatics. They realize that there's some variation in flavors and informatics, and also talk about some of the roles of EHR and clinical decision support plays with precision medicine. So I also talk about technology and information gaps impacting precision medicine and some of the clinical informatics approaches to address gaps. And so much of what Sandy said will probably be described here in some detail as well, which kind of validates some of these gaps that have been identified. So just briefly, bioinformatics can be thought of as a science to some extent where the domains are really driving the areas of research and then biomedical informatics methods and techniques are common across those domains. And so, since this is related to clinical informatics, we're really thinking about the actual clinical care activities like medicine, pharmacy, nursing, dentistry, et cetera, and they're all patient-oriented. This is just a brief definition from AMIA. And to further define clinical care activities related to precision medicine, I like to use the term P4 medicine that's coined by Lee Roy Hood at the Institute of Systems Biology, which is predictive, preventive, personalized, and participatory. And these areas are very well represented and ignite, for example, predictive in terms of family history of risk and susceptibility, being able to prevent adverse drug events with pharmacogenomics knowledge, complex disease, risk advice, which integrates both clinical data and genomic data, and then it's participatory, they're participatory examples where patients want to self-manage complex diseases. And so decision support is the bridge between, is the bridge to overcome many of the barriers to realizing precision medicine. And this is kind of the starting point for ignite. And so this was from a paper in 2012, which just outlined some of the barriers to realizing precision medicine as limited genetic proficiency of clinicians, limited availability of genetic experts and the growing genetic knowledge that need to be managed. And so we're able to, so we're able to use clinical decision support to really make the information that's been discovered available at the point of care. And so I like this figure to really show how clinical decision support can facilitate precision medicine. First it outlines what is the typical workflow when a patient comes into the office and then they need to wait in the waiting room until a nurse or a medical assistant sees them, a physician then sees the patient and then a final treatment decision is made based upon all of the data that's available to them. And what's shown in this figure also is that there are several paths that could be supported by clinical decision support. So importantly there's a lot of data and information sources that could be accessed for clinical decision support. So we see the dashed lines going from the first two boxes are data collected from a patient that are being stored within the electronic health record. And then also once a physician is seeing a patient they may order a laboratory test which is then interacting with another clinical system. So some of the areas where we're beginning to see our other data repositories for clinical trials data, molecular databases, biomarker repositories, we talked about passive decision support earlier where that might be accessed to journal articles, also information from other patients that are at other EHRs. As within the IGNITE network there are several institutions who are working together and we may be able to use population data from multiple EHRs at some point for risk algorithms, for example. And so on the other side of this figure it just shows that there are several points of delivery within the workflow of a patient visit in this case. So we see that screening or if the genetic test results are already available for a patient they may be filtering test results for different scenarios and those could be presented at the time when a patient comes to the office, for example, also assisting in visit acuity, also suggesting possible differential diagnoses and suggesting up-to-date treatment protocols. So these are all areas that could be facilitated by clinical decision support. And so now thinking about what are some of the technology and information gaps for implementing precision medicine, I broke it down into those three main areas. So the healthcare delivery process, thinking about the workflow. For each of the projects within IGNITE there are different types of workflows that are relevant and may not be. And so the different transactions for interacting with the EHR may inform how decision support is made available and what the timing is and what data resources are available. Also there are multiple stakeholders. We've talked about patients, healthcare providers, lab professionals, bioinformaticists and health IT professionals and also pharmacists. So there are several stakeholders who may be involved depending on the domain and also at what point the patient is being cared for. In terms of the data and information sources for clinical decision support, there are various sources that are relevant and so we've already talked about lab data versus EHR data and so we need to be able to assess those data together. Their data storage and access and exchange requirements to be able to integrate and access them in a private and secure way. Also ensuring high quality and being able to identify actionable data. We talked about or we heard earlier about variants of unknown significance and Sandy discussed the option of reinterpreting variants over time and so we need to have processes for identifying when that actionability changes. And so that brings me to the last area of delivery of decision support. So how we deliver those data may vary depending on the environment and also the actionability of what's being delivered. And there's dependence on vendor-specified capabilities. So across the network there are several EHRs that are represented and even if you have the same vendor like Epic, there's different flavors of Epic and so being able to take that into consideration when we're trying to come up with generalizable implementation approaches is important. Also current clinical decision support is inadequate and part of that is because the kinds of data that we're dealing with where it needs to be reinterpreted over time and again Sandy brought up some unique aspects of the projects that he's involved in where there is infrastructure for that kind of reinterpretation but in general a lot of the approaches have been implemented for different types of scenarios. And so when I talk with physicians sometimes I hear that they've gone through their whole process and then they get an alert message that says what they've ordered for their patient is incorrect and so then they have to start all over again. So being able to intervene earlier in that process for example would be useful for some types of data. And so now I'm going to talk a little bit about how there could be some approaches in clinical informatics to help address some of these issues and it's kind of a cross between informatics and implementation science within this group and so some of the lines are blurred there but a key process in informatics is doing a needs assessment. So first understanding how the workflow and the context is shaped to inform how your decision support will be implemented is important. So we want to know what are the pre-EHR, EHR and post-EHR tasks? Who are the stakeholders and how are that data used? In addition we could think about the workflow of implementing decision support rules as well where we're starting with most institutions have a rule authoring environment and they're developing a rule that generates an alert based upon a clinical event and the patient data that's available and once that fires then there's offered choices that then need to be responded to. And before you go through this process there are approaches to monitor that implementation and so that's one of the things that could probably be pursued within the Ignite Network in terms of before actually implementing decision support, seeing how many times it's going to fire and whether it's firing more often than you really need it to fire. Also after it's been implemented if there are ways to monitor the screens of folks who are using the decision support to see if they're responding how you would expect them to you can also capture different types of process measures such as whether they're ignoring it, whether they spend a second to two seconds on the alert and so you can learn from those kinds of processes. Next thinking about the data information sources we know that again needs assessment is important and so are there needs for data that aren't currently captured, are they being captured as free text versus as something more structured that can trigger an alert message so we need to understand what our data looks like and whether we have what we need. Also using standardized terminology and data exchange standards are important when we're talking about integrating those data, integrated knowledge environments that can bring those data together could be one area and then being able to share interpretations that are authoritative, concise and informative are important and some of those things we're seeing more of. So the last point is on delivery of decision support, being able to characterize CDS capabilities we can draw from a lot of the work that's already been done with clinical decision support and being able to characterize what those capabilities are. Can they support alert messages? Can they support passive and active decision support? A recent review article showed that most EHRs lack some CDS capabilities that are required but even though they lack those things there's a potential to use other modalities for decision support. Also understanding readiness to adopt CDS, user experience and design considerations are important and being able to measure implementation over time which can represent adoption as well as downstream outcomes and this is just some of the outcomes that can potentially be considered. And so this is just summarizing those three main areas and some of the potential ways to address those challenges. Thank you. Thank you Casey. Next we have Josh Peterson presenting on behalf of the Ignite Network.