One could easily argue that the diabetes registry should be part of the production LibreHealth EHR and the educational version, given the incidence of diabetes in populations at risk and the frequent users of open source EHRs worldwide. I ran into this Merck grant that seems attractive because they believe technology is a piece of the transformation of good diabetes care. Unfortunately, the letter of intent is due January 24th with formal applications due in April
This is just a different angle for funding: adding features to the EHR to improve medical care for"populations at risk" with expensive chronic disease such as diabetes. Iāll try to see what else is out there that might fit our needs over the weekend
The letter of intent is about 5 pages. I wish it wasnāt the weekend because if we did it I would have to devote the next few days doing nothing but writing the LOI. On the positive side we have loads of buzz words that might support a grant. Because it is the weekend I canāt bounce this off my own research folks. What say ye?
@tony is the plan to move the next iteration of librehealth EHR to the toolkit database ? If so we have had talks with sunbiz on adding spring data that makes integration with analytics easier
I am using a version of toolkit as a radiology registry so I like the idea of adding the diabetes registry to the EHR platform
I really think the approach to add a state of the art diabetes registry and an analytics package is a slam dunk, BUT re-reading the Merck grant I ran into the following that is disallowed "Capital or building campaigns, including new construction, renovation of facilities, or health information technology (HIT) installation or improvement"
I donāt know if there is a phone number I can access on Monday to see if we would be excluded but this would only give us a day to write the LOI. Given our recent publication, it wouldnāt be hard to cut and paste stuff but I think today Iāll look for other organizations supporting diabetes with more lead time to submit a proposal
I am assuming that if we get a grant and name the Software Freedom Conservancy there is no problem directing funds to them and they will disperse them to LH for this project? Many of the grants I looked at allow 501 c 3 orgs to be the main recipient of grants. Then would the two universities be consultants on the project?
I just stumbled onto a 2013 paper Data Mining NHANES.pdf (838.1 KB)
that really supports our approach. These authors combined data from multiple NHANES studies and used machine learning (WEKA) to analyze the data. They pointed out that in 2008 AMIA sponsored a data mining contest using NHANES data.
Another thought regarding this paper would be whether to invite the second author of this paper who is a machine learning expert at BYU. Would he still have the NHANES data he would be willing to share as that might save us time/money. Just a thought
Modifying the production and educational/research version of LibreHealth EHR would be a very exciting step and one likely to be supported by many funding agencies. Iāll make some phone calls tomorrow, speak to my research department and Sunbiz will do the same with IUPUI. Exciting stuff
I wonder if this Bloomberg Data Science grant might not help support our initiative, along with other grants. Seems to emphasize open source projects that can be shared. Only $40-$70K but might work to either support Sunbiz at IUPUI or a computer scientist here at UWF who would like to be involved. Application is very short but for this cycle is due February 3rd. Please review and provide input
Judy,
This is a very very important idea. Mind you, it is a heck of a bit of work, but one that I think I can help a lot with.
My silly approach is to simply attempt to convert the existing EHR data structure to the other oneā¦if it is possible.
SO
I would like to start mapping one DB to the other. Does anyone have a dump of a toolkit database with inserts that have reasonable patient data that I can start asking questions about?
As it stands I have a one page presentation we can present at InSpire 2017 and the deadline is next week. It doesnāt appear that we have enough contributors for a panel. Once again here is the link to the presentation that I would like to upload in the next few days
Keep in mind that the emphasis of this one-pager is to present survey data to prove an educational version of LibreHealth EHR is justified and that the first step will be to populate it with adequate patients
I have worked all afternoon trying to find appropriate grants for this project. While I like the SBIR grants, the due date was Feb 22nd and it might take weeks for the PI to register on multiple sites to be eligible.
HHS/AHRQ has an interesting option Small Research Grant Program (R03) which covers 1-2 years for feasibility studies. The application is only 6 pages but there is a limit of $100 K (direct and indirect costs). There is no deadline which is helpful. I decided to email AHRQ because they have a HIT funding email so I asked them what might be most appropriate grant be for our proposed project. It might take 1-2 weeks for a response. I did mention that two universities are involved, the Software Freedom Conservancy and so forth. I donāt need any compensation personally from the grant, but there may be some need for Sunbiz and a computer scientist here at UWF. I know this is a difficult question but is there some feel for how much money we might need to support the integration of the diabetes registry and the machine learning software?
@rhoyt the planned presentation summary looks great.
Re: Funding ā¦ SBIT grant did you post a link on this ? I am meeting Saptarshi next monday to review funding for this on our end
The small Research grant program looks feasible given 6 pages .
I briefly looked at the Diabetes registry paper you shared - but this is dependent on whether we keep the current LibreHealth EHR backed for tables ā or move to the toolkit backend for tables. The openmrs database that we adopted tends to be more scalable because of the obs table - but i will let Tony and Art comment on this ā¦ Still we should try and get a quick win.
I havenāt looked at weka on the integration side but i can try and glance over it over the next few days to see the feasibility of integration ā¦
@judywawira
Judy, Our gang talked a good bit about the database mapping bit yesterday. We are really hot to get on that.
Toolkit->EHR->Toolkit mapping is the first stage. This will give us a gap analysis for both data structures. Next is migrating the EHR database towards the Toolkit structure as much as possible, then looking at what is missing on the Toolkit side (there will be a good bit), creating that in the most consistent way possible, then finishing the transformation of the LHEHR database.
API integration is great and all, but 1:1 relationships trump that any day.
-Still need that data sampleā¦