Over the past 10 years during my conversations with smaller universities and colleges they seemed to have struggled finding a non-commercial EHR for teaching. Unfortunately, I am unaware of a published needs assessment to determine what type of EHR (inpatient? ambulatory? both?) would suit their educational needs. In addition, it is my belief that an EHR that also offers the ability to upload de-identified real patient data could be used for EHR training, predictive analytics, database management, research, etc. I have feelers out to Oregon Health Services University to see if they can supply synthetic patient data but have not heard back. In the meantime I have closely examined the NHANES files 2011-2012 (almost 10,000 patients) and extracted 60 tables containing important clinical attributes: NHANES 2011-2012 tables.pdf (35.3 KB)
If we combined multiple study periods we would have data on about 65,000 outpatients.
I would like to consider sending out a survey to health informatics instructors to determine their needs for EHR education. I could use the AMIA Clinical Informatics Communities of Practice (CICOP) user group or I have a contact list of 400 health informatics instructors who have requested a download of my textbook who might be willing to take the survey. I wanted to run this by others for their opinion and see if anyone would like to assist
@aethelwulffe and I have been chatting with Bob via email about this and we think the technical challenges are not too great. We can create a distribution of LibreHealth EHR with selectable test data sets, for instance. Interested in what can also be done in lh-toolkit.
I know @sunbiz has been gathering up folks who are interested in using LibreHealth products for education purposes. There are several people at his university (Indiana) and also a few other faculty who have approached us both before & after our public announcement. So we already have a few universities on board that want to use our software as part of their curriculum, and others who are interested in learning more.
We do have a category #community:education available & probably should start some more significant conversations from everyone to make sure we understand all the desired goals. I think there is some overlap and there are also some unique needs that we should try to capture before starting too much work.
Hi @rhoyt ,
This seems like a very nice idea. At the AMIA academic forum, there have been numerous requests for a teaching EHR and to share the data we have at IU with others. It would be nice to get your efforts going, and I know quite a few places that would be onboard with this too, as they have been in discussion with us.
Beyond just the data, I think an academic EHR also needs to have important workflow related modules, as well as particular reporting/analysis features that can allow looking inside the EHR. We’ve debated a few times and put this under the broad umbrella of “simulation”, such that clinicians, as well as informaticians can see changes in outcomes, when programs are changed. lh-toolkit seems like a useful place to put this into, since we can build upon FHIR resources for interoperability, as well as be closely aligned to HL7 concepts in the data model.
Would you like to see questions about the modules in the survey? Could you provide an example of the type of outcome that might change with program changes?
I would hope that this platform could be multi-purpose. Not only could it support EHR training and education, it could support usability testing, SMART apps, NLP, Machine learning, etc
I would also plan for lab and radiology systems beyond EHR - most informatics training programs focus on EHR , yet there are other domains critical to caring for patients. Also would include things like CDA , FHIR and clinical decision support. I would be willing to work on a terminology component for this and i am pretty sure @akanter would.
It would be great if we can spend some time writing all possible data sources with patient information (Feels like a good #GCI task)
My initial feeling would be to begin with attainable goals and then steadily add features. I see tremendous potential in this type of project but I think it will take us quite a while to add all of the features everyone wants. Do others disagree?..Bob
I agree we need to start with what we should call the minimum viable product (MVP), and then expand on the features. I think an NIH P30 or even an R25 like grant would be the best, since we are talking multiple universities.
We should probably get started with a formal needs assessment, using a survey which helps capture the top 3 features that everyone needs, and then details of those features, that will help us construct the MVP… and then release and get feedback.
I began a survey using Qualtrics that I exported to Word and stripped out the skip logic. Obviously, it needs additional questions and fine tuning…Bob
Survey_Educational_EHR_with_Data.pdf (23.0 KB)
Is there a Google doc version of this, where we can provide suggestions or edits? e.g. Q.15 option 3, doesn’t quite apply I think? or may be you are asking if they want to switch to an alternative?
May be we should have a question whether they already have data, that they want to use in this teaching EHR or they will use our synthetic data.
Another question to ask is whether they have the resources (human, technological, capital) to host their own teaching EHR or would they prefer a hosted/tenancy one that they can connect to? Real deidentified data is much more easier in such environments I think.
We should also think if this teaching EHR is meant for teaching clinical aspects, informatics aspects or both?
Good points. What I started was just a very rough draft to get the process going. I would be happy to place this on Google Drive and invite those who join this forum to add questions or comments…Bob
I like the questions. To help us better shape our offering it would be important to understand something about the learning outcomes the users expect from their students. We can compare notes on the responses. We could learn a tremendous amount from the responses which would enable to direct our attention appropriately. Obviously, the outcome descriptions would also provide insight into the background of the users - computing experience, health-related background, etc.
Just a few thoughts.
Here’s a Google Doc with world editing permissions for everyone to add in their brainstorming. Feel free to paste in anything you’ve already written up, or add new comments/ideas … please make it clear who you are, and provide some white space or lines between contributions, so we can tell who said what!
Thanks, I moved the questionnaire to the Google doc and added some of the questions that I wrote. @rhoyt do you have a timeline in mind when you want to distribute the survey? I think the AMIA academic forum as another place where this could be distributed.
Thanks for posting this on Google Docs and I like the additional questions. Some of the issues are sorta chicken and egg issues. Art and Tony seem to think that we can give installers the option to automatically upload data as part of a standard installation ( I don’t know if that means both local instance and remote hosting). I guess we don’t want to guarantee any technical detail at this point.
I would like to get input from instructors reasonably soon with a survey. While it would be great to have an extensive and validated survey, that might delay our need for early input. We all would benefit from knowing how important a hands-on EHR is for Informatics Education. I certainly appreciate the input from others so we can have concise, well thought out survey questions.
Another challenge will be the fact that survey participants will probably know nothing about the LibreHealth EHR and it would be asking a lot to have them test drive the demo before taking the survey, but it would be hard for participants to have an opinion without looking at the software. Any thoughts about this?
I dont think this will have much impact. Users will be able to describe their needs without using the product. I think EHRs have been around for long enough for people to know, the kind of features they could use.
I have added the additional questions and modified some of what I had already created. Please take a look at this document that also includes the skip/display logic. It is always a delicate balance between asking enough questions versus asking too many and participants having “survey fatigue”
Educational_EHR_with_Data (2).pdf (34.1 KB)
Is there a version of this that is a fillable form? Or perhaps it only works with one particular adobe product or another? I can convert it to a widely usable format or to an HTML5 form/survey if you need/like.
I exported this from the survey tool Qualtrics. Here is a link to the actual survey. Feel free to take it and be sure the skip logic works. I’ll delete the responses prior to sending the final survey
https://uwf.co1.qualtrics.com/SE/?SID=SV_b4vGBHN4GcWMHnD
