SHARPfest

PIs and senior leadership convene in July to discuss opportunities for synergy.

Ideas for the agenda - what are the burning issues and use cases that we can raise at this meeting?  

- Coordination of data representation between SHARP3 and SHARPn
- Data security coordination with SHARPs
- What are the potential show-stopper topics that are not currently covered by the current SHARP efforts - any additional gaps that need to be addressed?
	- De-identification is an example
	- Authentication of users
- Data sets and their availability - make these available to improve the output of the SHARP efforts
- Within SHARPn we are developing a large set of free text files and make this available - how do we continue to make this available beyond the life of SHARPn
- Data segmentation - the ability of patients or providers to restrict what part of a record can be shared
	
Beyond SHARPfest, what does SHARPn need to be doing 
- How can each project group support the other groups withing SHARPn (NLP supporting the phenotyping group, e.g.)
	- we know we want to develop generalized tools, but how do we expose these more broadly within our workstreams?

Pipeline diagram
- Big part of the pipeline was about getting data from one entity to another - how much effort goes into making that work (we are going to redesign our uber-schematic to more appropriately illustrate this)

How do the various SHARP projects measure their value - cost/time savings, benefits in the real world?

Cross-team collaboration - the demonstration project exposed many of the detailed issues that present roadblocks/challenges.  Tactical agenda for the next couple of months is to complete the pipeline project to do just that.  The tracer shot project forces us to examine those issues.  We will face many of the problems that the SMARTapps folks have run into.  Task #1 should be for the phenotyping group to start looking at the APIs for SMARTapps; we would learn a great deal from this by adding these other SHARP components into the pipeline.  

Also raises the questions around what kinds of security/provenance issues will we reveal. (Mark Frisse is "all in.")

- How do we franchise this? If someone wants to set up a UIMA pipeline, how would we make this easy for them?  Configuration documentation, source code, etc.  How do we make these open and available?
	- If we look at what we've done now, we should start putting those out so people can play with them and give feedback.  We need to talk about how to do this faster.  (This should be a SHARPfest discussion.)
	- Would be very useful to make it easier technically for these pieces to be made available.  Right now, it's all very hard to simply install elements like cTAKEs.
	- We are starting to deploy these tools and we do need to spend time now around how do we package these for widespread distribution?
	- Mirth Corp has done a lot of this.  Appliance packages and other methods are ways to make this easier.  The pipeline would be an example of an appliance package.

The laying out of the next year is something where we need some formality.  We want to keep that strategic view about the end of year four, but need to get tactical.

Need for use cases to drive the vision of this effort.  Need to refresh the use cases that focuses on collaboration issues.  Would be good to launch on some use cases that are relevant to the individual institutions, the SHARPs and the Beacons.  
.
SHARPfest - plan right after this a meeting to summarize what happened there for the larger SHARPn community.