Educational Goals

On a copy a long long time ago. I have done it for other databases recently. Yours, frankly, is huge, and I can’t actually do that project until the beginning of April due to work commitments.

Bob,

Do you have a class or graduate students that could contribute time helping with a manual review and redaction after the first machine pass or two?

We would need a small cadre that we could set up a business associates agreement with. They would need to swear in blood and cross their hearts to be good people since they would be dealing with live confidential data. (Just kidding about the blood of course). It seems that we could work this out legally and having some dedicated manpower would help a lot in terms of creating a larger and more meaningful database for current and future students.

Sam Bowen, MD

@sam-bowen @aethelwulffe

Sam, I’m no longer associated with the University of West Florida so I don’t have graduate students to help. I’m certainly willing to donate some time, particularly if we start small as an alpha test to see how labor intensive this might be Bob

@sam-bowen, @rhoyt our Health informatics program at IUPUI can definitely help. I need to talk to the lawyers, and school-level administrators but we are definitely interested. We are already doing an RCT using LibreHealth EHR with NHANES and synthetic data.
So real data will be very useful!!

@sunbiz @sam-bowen

That would be great if IUPUI could contribute some man/woman power. I suspect they will need HIPAA training. I certainly agree that some real patients and longitudinal encounter notes would make the EHR more useful. Still feel the NHANES data can be used for a myriad of clinical and non-clinical student exercises

@rhoyt, they do take the HIPAA training and know all of the security risks and mitigation strategies. I think it will be valuable to the grad students as well.

@sam-bowen After talking to a few administrators at Indiana University now, do you have any existing BA contract that I could share with the university lawyers? I think we can use the remaining part of Spring semester or the summer semester to get the extraction + deidentification work done.

2 Likes

Try http://suncoastconnection.com/Downloads/BPUCBPA.odt

1 Like

The link I posted should be pretty much set up for a BAA for redacting data for an educational purpose. I am sure that despite it being CMS boilerplate for the most part, your legal eagles will want to edit. :slight_smile:

1 Like

I also posted a link to the latest version of the synthetic data tool. Oh, I guess I will paste it here too… http://suncoastconnection.com/Downloads/PrintaPatient_A7_2_0.zip

Thanks, @aethelwulffe I did use it to generate some 100 odd patients with many visits for the RCT that I mentioned above. And then added some more to that using those generated using Synthea.

New version creates large numbers of users and facilities so that these can be used for load testing of admin UI stuff.

1 Like

@sunbiz

I’ll be giving a presentation on LibreHealth at the AMIA Informatics Educators conference in June in New Orleans. Do you plan to attend? If so, do you want to talk about the non-NHANES approach?

@rhoyt, I’m not attending the Informatics Educators conference. But we should collaborate on writing in more detail about the student feedback. And also the trial that we are doing in the course.

@sunbiz Ultimately, I would like to write an article on the “academic” or “educational” EHR and hope you can contribute. In June I will be on a panel at the Academic Medicine conference discussing “Educational EHRs out of the box” which is the desire of Harold Lehmann to coerce the big boys to add educational features to their production EHRs. In a paper, we would mention that approach and the Regenstrief project and the SIM-EHR at OHSU, but my desire is a little broader to include the potential to support analytics and research in a open source EHR

Coercing the big EHR vendors to release their educational EHRs is a pipe dream. The advantage with open-source, particularly for informatics training, is that you can explore, modify, test and improve the internal workings of the EHR system. The flexibility and extensibility, at least of lh-toolkit is also beyond anything that the big EHR systems can offer. So, open-source allows for a different kind of teaching-learning pedagogy that just using a product doesn’t allow.

I think we are quite close to getting the BA agreement ready. Do we have @sam-bowen permission to have access to it? How are we going to get the data? as a MySQL dump? can we discuss this over email?

@sunbiz

I could not agree more about the pipe dream. Harold thinks the big boys are going to create a “learning health system EHR” with constant feedback to the student and faculty, using AI, etc. I told him if there was no business case there will be no educational version.

I recommended an open source EHR more than 15 years ago so we all had a “sandbox” to test functionality, usability and new innovations. Never happened but our work with LibreHealth is about as close as anyone has come.

It is ready within two days of you folks saying “push the button”.

1 Like

Art would you forward the proposed BAA for me to look at?

Sam Bowen, MD

@tony @sunbiz @judywawira @teryhill @r0bby @downey @aethelwulffe @sam-bowen

AMIA is interested in supporting LibreHealth EHR in anyway that they can, because they see the need for an educational EHR. I have a meeting at noon on June 20th with Doug Fridsma and Jeff Williamson at the Informatics Educators Conference in New Orleans. Jeff had mentioned that they have access to some grant organizations that are closed to most people (i.e. they don’t accept unsolicited proposals).

Can we please put together some concrete proposals about how grant money could be used and an associated price tag.

  1. Sunbiz has mentioned completing the API and Open Web Apps but I don’t know if grant money is needed
  2. We could expand the patient database with more data fields. For example, NHANES can provide “social determinants of health” such as income and education. We could also expand the lab section to include more results.
  3. Add de-identified patients to the EHR so we could follow someone longitudinally (Sam’s data)
  4. Add YouTube videos on simple analytics, using Adminer and other tools to mine LibreHealth data
  5. Another round of beta testing with more universities
  6. More long-term funding to support hosting and maintenance

Please give me some ammunition to present with the AMIA staff, along with financial estimates. Thanks.

1 Like