Additional Meeting Notes
Ross's Meeting Notes
- Make sure you update your bio on the wiki
- Our goal is to assist in connecting folks and help disseminate backgrounds
- Follow this meeting on Twitter: SHARPnORG
- Main goals for the meeting
- Get to know each other
- Get to know the different workstreams within this meeting and how we can make the whole project provide more value than the sum of its parts
Connection Points between SHARP Areas
Privacy and Security
- Both the SHARP Research (Area 4) team and the SHARP Security (Area 1) team assumed the other team would deal with deidentification issues
- We need to discuss how these issues will be addressed through one area of research or another.
Open API - SMArt
- How does the money flow in an open app development environment?
- Still working out how vendors will view this. Right now, we're focusing on open source applications, but this is an issue that needs to be addressed.
- Stan and Chris are very interested in collaborating on this.
- Critical that the data be well represented in this and standards have a critical role
- Often with RESTful methodologies, the secruity protocols sit outside - how are you going to build this in?
- Looking at using OAuth - Open Authentication (authorization?) - allows this to happen in a well-structured manner.
Wil Yu - Overview of ONC Programs and HITECH Act Efforts
Note to Self: Recorded end Q&A and Discussion - File STE-000.mp3 Q&A
- Chris C: There was a discussion about being able to ask questions of the national infrastructure - what is the status of that?
- Chuck Friedman calls this Element 3 (Meaningful Use of EHRs being the first, HIE being the second)
- We are working with IoM to scope out these efforts
- ONC was about 30 folks last year; now at 85 plus many, many contractors
Propellant / Repellant Forces
Close scrutiny of the project
On the inside – would be satisfied if we made broad progress;
on the outside, they will be looking for focused success – so
they may be looking at insights created around a single medical domain
Repellant Forces Repellant Forces
Close scrutiny of the project
If people can’t see the use of individual accomplishments to move toward a resolution. Analogy: NLP – you can come up with the best tokenizer and who cares – you need the other pieces on both sides of that
Marshall you NLP forces to solve _______.
Middleware: Gets at the notion of how we consider success
Identifying terms that are of interest and assign them to codes (Entity Recognition) is an end to itself for NLP, but we could support secondary use as defined by others – that would be an important piece that those trying to solve a particular problem (like curing lung cancer) as valuable. So the end game is not necessarily public noted impact, but that those who would see value in the middleware we create would acknowledge the value.
Can we choose a couple of clinical domain areas to focus on? In my experience, it takes more time to do data prep than anything else.
From an NLP perspective – the time investment is data prep, training data sets, etc. Get the sense that we are off working on our own and we’re not connecting these groups. If we don’t start nailing down specific domains and work on those connection points, we won’t show value.
Group sees value in creating a story or use case to follow through the various workstreams.
Would also want to support Peter in being more proactive in just doing our work by reaching out to some of the other SHARP programs to make sure we are providing something that would be of value to them. Maybe we don’t actually do the “secondary use” but provide the middleware to the other SHARP research area teams that would be of value to them.
First 4-year grant I’ve ever been on – this is much longer horizon than I’m used to. Implies on the propellant side – we have to do more than set up data, etc. we have to do work that is worthy of four years of work. On the negative side, we can come up with ideas that are timely today and outmoded in four years. (like putting this on the “cloud”
The Methodologies we will be applying in NLP, then how this will enable phenotyping, this will help provide feedback and extraction of use cases. When I think about what this middleware is going to do for me, I see NLP and phenotyping both as being middleware – so it makes a lot of sense to reach out to the SHARP areas and also look to the Beacon programs – all of the work they are doing—measuring meaningful use, quality, cost efficiencies, etc. will be valuable to them.
Chris – I’m interested in epidemiology, but I don’t want to do epi here – I want the epidemiologists to say that what we are doing is saving them a lot of time in getting their work done.
Steve – I’ve been involved in longer projects where it was always helpful to describe the arc of the project and develop milestones along the way. At the end of the project, what is the objective.
This is a very American view of the process – the invisible hand will take care of that. Need to show the value at year 2 because you are going to be worried at that time because you will be thinking bout the next round of funding. Link up with an epi group or some other group – you are sitting on a treasure – I would like to see you revise ICDs – if you can improve the speed of this cumbersome process, that would be a great success.
If you can link your capabilities to the iPhone project at Harvard, that would be very helpful.
Lacey – ONC is looking at what we are doing now… not our 4-year objectives, but what are you delivering now and in the next six months. Also, this is NOT a renewable grant (though the ONC has hedged on this a bit). They have also intimated that, if they are as successful as they hope it will be, there will be monies made available. So they will not be renewable in an NIH sense, but they will lead to new opportunities.