I have mentoring from 1 PM to 2:30 PM CDT tomorrow (Tuesday)
OK … we can do it at 9am PDT 12pm EDT if that’s OK or Wed/Thurs I’m available 9:30 PDT to 1pm PDT any of those days
Tomorrow is out for me because I also have a meeting prior to the mentoring. Wednesday and Thursday look good. What about 10:00 PDT On Wednesday which would be 12:00 for me (CDT) and 1:00 PM (EDT) for Kevin?
Kevin was kind enough to call me today so we did go over some short-term ideas we can discuss then
Sounds fine.
Conference Call Info:
- 713-535-9011 ID# 88961
Keep everyone in the loop 
– By everyone I mean I wanna know what happens
Great meeting with @yehster @rhoyt @tony
Now: Enhancements (Kevin’s Notes)
- Loading the laboratory data as separate panels instead of one monolithic report for all NHANES data
- Improved searching/reporting based on lab values
- Reports export as .CSV
Grants: What are the target projects that would attract a grant.
- Interoperability Open API for FHIR (use the spec MI2 wrote for Reddy) $75/K put
- NLP: Progress Notes → CTakes → FHIR → EHR
- Need to get progress notes to de-identify and process
Hosting:
- Pay model for hosting needs to be determined, University sponsored, or paid for individual students.
- Cost of hosting EHR instances (with and without HIPAA) - MI2 provide some costing from our experience
- Scalability / Containerization is important - to move to container model/installation - Proposal from MI2
User Support:
- Training Videos (QA side effect) - Proposal from MI2
Other possible:
- Data set management and sharing
- In Patient Care version
- Sophisticated Med Management
- Enhancements to Clinic Rules tools
- Interface / Port the lh-toolkit Concepts Dictionary model to LH-EHR
Great summary of our discussions. One other item to include in “short term” planning would be better templates for progress notes. In other words, it would be great to create templates for common conditions, such as URI, UTI, low back pain, etc. Does anyone know if there is open source software for this??
Our SW does this, so does lh-toolkit, the question is whether it’s really the right model.
1 Like
https://github.com/LibreHealthIO/EHR_Contribs is where we are…or are supposed to be…developing what you are calling “templates”. The term in this context is “Clinical Form”. Template implies text narrative content only.
This repository has not seen much activity. There are 12 pull requests that have been sitting there for months, and nothing in the master repo as yet.
1 Like
@tony @yehster @teryhill @aethelwulffe @r0bby
I want to post the few responses I got from universities regarding host fees:
I do think I can convince the University to fund a small amount to have student access to the entire package – given a little warning time to grease the skids!. Actually, an annual charge would be easier to get approved than a monthly charge for this program…Dr. Mitchell, Harrisburg University
At small institutions like ours (New England College), hosting our own installation simply isn’t workable. A hosted solution would be a far better solution for us. I’m fairly certain NEC would be willing to pony up for a hosting/support fee. The college already cheerfully funds a similar fee for a course I teach in the CIS program. Finally, I agree with Glenn that an annual charge would be easier all around.
So far, we’ve only used Libre for the activities Bob posted for all of us. Those have gone very well overall. I believe that our college would pass the cost of the Libre subscription model on to our students. If that is the case, we’d probably need to sell a user Id and password to the student and it would be active for 12 months. Our program is small, though, so I’m not sure we’d have enough students for the cost to be affordable per student. We only have about 20 students per year and they would have assignments in about 4 of their courses per year… Brandy Z. LSSC
1 Like
Great feedback. We should do a simple annual subscription model. MI2 could do that and share some of the revenue back to the project as a possibility. Or the project could do it directly, but that seems like burden on free resources (ie: Robby)
@tony tells me that you have good connections to Google, which is unusual but potentially very helpful. I wonder if they might support the notion that we would like LibreHealth EHR to be an educational and research sandbox and that is why we could use some financial support for future enhancements of:
- Open API so we can create SMART Apps on FHIR, Apple Research Kit, etc.
- Incorporate NLP as part of the platform. @tony believes he can secure de-identified progress notes we could mine patient encounters. BTW, the machine learning program I use (WEKA) also does text mining but there are lots of open-source NLP programs.
This Sunday I will attend an AMIA workshop on NLP so hope to know more
Hi @rhoyt. Google does offer research grants for folks, and do of course love open source. See https://research.google.com/research-outreach.html#/research-outreach/faculty-engagement for examples of their faculty research awards and focused research awards.
It’s usually best if an applicant has a fairly clear & concise mission statement, and then works to get a sponsor within Google. I believe @sunbiz has some experience with this in the past? And a colleague of @ada too, I think.
I certainly have contacts, but they are fairly limited in terms of being the actual folks who give out the funds. They could certainly help with referrals, if need be.
Has anyone else here worked to get funding/grants on projects through Google?
1 Like
Thanks for your reply and I hope others respond as well. I was also interested in the Chan/Zuckerberg Initiative that seems to fund educational and technological advances. While they don’t explain whether they accept unsolicited grants, I saw that several of their board members have Facebook sites so perhaps could be contacted that way. The idea here would be to say that an educational EHR could be of benefit to multiple students in multiple fields.
1 Like
I did ask the universities using LH EHR whether they felt that YouTube instructional videos would be better than just the “starter” PDFs I created. I head back from 3 of them and they were unanimous that they thought the video tutorials would be extremely beneficial. The wheels are in motion to begin creating the videos. Thanks @tony and @htuck
@rhoyt that’s gratifying to hear, though I’ve seen that 'paper’ references definitely have their place. We’ll have to see how the training program shapes up and where written information would fit.
It is my vision (realistic or not) that we have more advanced videos down the road. For example, if Kevin is able to improve the search engine capabilities, that include csv downloads, we can come up with some simple scenarios for visual and descriptive analytics for clinical and non-clinical students.
I see no reason many of the health sciences would not benefit from our initiative. This is also why I am contemplating asking a large publisher such as Pearson, if they have any interest in supporting our project. They have about 10 textbooks that theoretically could benefit from having hands-on exercises using LibreHealth EHR.
I probably should mention that I plan to discuss our initiative at the AMIA Education Working Group this Sunday night. It is from 7:30 to 10 PM so not sure how many people will attend. Did see that AMIA is pushing the Library of Medicine to promote Data Science Research so our project does promote data science
1 Like
There was good interest in the LibreHealth EHR project but no suggestions regarding industry partners or grants.
Attending every lecture I can on NLP, because we discussed using our EHR for that purpose down the road. Unfortunately, NLP looks like a full time job for anyone entering the field. Supposedly, Harvard developed a simpler information extraction engine, known as Canary that is free to download, so at some point, I will evaluate in detail. Here is a recent article on Canary. I wonder if we have any graduate students who would want to investigate. aci_2017-8-2_27489.pdf (534.2 KB)
1 Like
There are a few, cTakes, Canary, Banner are the ones I know about. If we can put together a infrastructure to receive and parse notes we can add as many NLP parsers and take the “best” data responses. Hadrian Zbarcea and I have been working on the backend services for cTakes already.