I’m a university professor/researcher in the Artificial Intelligence domain. I’m working with applications of AI on healthcare, like chatbots (see e.g. Redirecting ). I’ve just started to learn about openEHR. I’m interested in topics like mapping textual data to structured openEHR using NLP techniques. If you have some good pointers so that I can learn more about openEHR and such existing applications, I would be glad if you tell me.
Yours,
Sergio Queiroz
Centro de Informatica - Universidade Federal de Pernambuco, Brazil
Welcome to the community! I’m highly interested in your work.
I myself am working at a vendor on adding structured data from free text using ML at input time. For some related discussion see Semi structured narrative data
Learning openEHR is not easy, and it’s proven hard to produce easy “getting started” materials. Because there’s a lot of different backgrounds (clinical, informatics, policymakers) and levels of understanding of health informatics, and due to the complexity of healthcare data openEHR is highly complex as well. And depending on your experience some basics of openEHR can be counterintuitive.
If you search this forum or openEHR.org or YouTube some good material will come up.
And please don’t hesitate to ask for more help or to test your assumptions and understanding, people are very much willing to help here.
The highmed project from Germany may be interesting to you as well, afaik it’s not doing (much) ML, but it is transforming data into openEHR for (currently) research purposes. Ping @birger.haarbrandt
The best way to understand the specs is the specs themselves, though it’s not a shortcut, requires a lot of dedication and coming here to ask questions on things that might not be clear. For data processing and NPL stuff, I would suggest to start from the data types spec, specifically from the text package Data Types Information Model