Fantastic Work! It’s great to know the work related to Covid-19 and a shared project specially for Covid-19.
My team has also been working on modeling Covid-19 for weeks. I thought our work can also been align with the project and we’d like to contribute.
As you all knew, China is the place with the earliest and the most patients in the world. Lots of efforts have been put by Chinese during these months, together with many achievements. So far, the diagnosis and treatment guideline for Covid-19 has been updated to 7th edition，And it worked well in Chinese hospitals and saved more and more patients. With the hope of sharing evidence-based knowledge on the diagnosis and treatment of COVID-19 gained in China in the past two months, our team is currently working on modeling these knowledge and planned to share it with the public.
What we are engaged in including:
Developing a template for interchange Covid-19 patient data set between applications, the use cases can include some decision support tools for diagnose and treatment,the dashboard view to show the patient data related to Covid-19, and patient registry database collecting patient data for further research. The original Chinese narrative text guideline was segmented,structured, marked up, and then translated into English. The Entities captured from the guidelines were were modeled to the openEHR templates. The dashboard view are also under development and the prototype has already been tested in one hospital in Wuhan. The work has been nearly completed and will be shared soon. From my understanding after a quick glance of the modeling results you have shared, our work may be some bit more with the purpose of supporting diagnose and treatment applications in the hospital, while yours are more with the risk assessment. We can discuss more after we have the first draft after a few days.
Authoring the Chinese COVID-19 diagnosis and treatment guideline with Guideline Definition Language (GDL). We have already completely encoded the 6th edition of the Chinese guideline with GDL and tested the rules with patient cases. A diagnosis and treatment decision support system prototype has also been developed for demonstrating the GDL rules. Now, we are working on upgrading the rules to the latest edition (i.e., 7th edition) of the guideline. Our work is about to be shared with the public on GitHub within a few days, and a journal publication is under preparation.
Considerably modified and tidied up Covid-19 screening template. I have tried to minimise any dependency on English - rermoved ad-hoc sections, added some specialisations, used symptom-sign specialisation, and removed the list of countries/regions so that this can be inserted or served in local languages.
I’ve had an issue uploading to CKM for some weird reason that I’m going to ask Sebastian Garde to look at ASAP.
Very good to hear from you Xudong. The rest of the world is catching up to learn from China’s experience and
your own team’s work. The GDL stuff sounds fascinating- do you use GDL 1 or 2?
I’d also be interested in your registry /exchange template - we have started work on something simialr based on the WHO reporting dataset. I hope to have a fist cut of that visible tomorrow.
I have uploaded the Covfefe templates and archetypes to that new account and created a Covfefe repository . For now, until we get some goverance rules/arrangements can I ask you to treat it as read-only, but feel free to replicate the repository if you want to work e.g on translation or local adaptation.
Over next week, we can try to come up with something more democratic!! But for now it’s mine - hands off! (Change requests welcome, of course!!).
We could possibly change that to a GitHub repo instead of a local folder but I have found that can cause some issues.
@sebastian.garde has kindly stepped in to explain where I was being stupid about CKM. THe upshot is that I will have to change a s couple the new V1 archetype to V0 and break the template. I will also take the opportunity to make sure that the symptom-covid extension is specialised from the existing published V1 symptom archetype, not the version that has the breaking change on CKM.
I need to go to the shops before the ‘HamsterKaufen’ buy up all the food and toilet rolls but will get back onto this in a couple of hours.
Ian McNicoll
openEHR International Board member
7 March
Very good to hear from you Xudong. The rest of the world is catching up to learn from China’s experience and
your own team’s work. The GDL stuff sounds fascinating- do you use GDL 1 or 2?
I’d also be interested in your registry /exchange template - we have started work on something simialr based on the WHO reporting dataset. I hope to have a fist cut of that visible tomorrow.
Xudong Lu
openEHR International Board member
7 March
Hi Ian, Fantastic Work! It’s great to know the work related to Covid-19 and a shared project specially for Covid-19. My team has also been working on modeling Covid-19 for weeks. I thought our work can also been align with the project and we’d like to contribute. As you all knew, China is the pla…
Sorry @bna - I might have messed up or lost some of your translations - I had to do quite a bit of work on the specialisations. Let me know if something is missing and I’ll do my best to correct it.
Thinking out loud about the assessment and symptoms:
Postulate 1: Symptoms should be mandatory
Postulate 2: Add the “story” element to make it possible to record a story without any symptoms (Question is: How to capture a non-symptom story?)
The only element used is “date of return”. What is the semantic of this date? Is it a travel to a region with a known breakout? Is it a travel trip to my friend in the neighbour city?
To answer this question about ‘Covid-19 screening negative’ now I’m back on-board. I have sent an email to the person who I think is closest to this decision, Dai Evans ant PRIMIS UK. One of the problems of UK GP system information models is that though they are good at handling new SNOMED codes, they are not good at contextualising that, particularly for querying. This raises an issue for ‘Covid-19 screening negative’ - when, where , what context, so I suspect a decision may have been made only to record ‘positive’ results/actions.
I will look for a generic SCT code to add for our purposes since we control context very easily-
THere are some codes for e.g ‘low risk of Ebola infection’ but nothing on Covid-19 or even a generic ‘low risk of infection’ code - this is where SCT creaks badly.
I am going to add “723505004 | Low risk (qualifier value) |” which is definitely wrong but currently not possible without using post-coordinated grammar whicxh is barely supported out in the real world
I did wonder about this and the field is in the underlying Story archetype but I’d like to keep really close to what people are actually doing, asking for in real usage.
My thinking was that this particular template is about a rapid screening assessment and that the detailed story is not critical (here) .
This template is definitely not about a full clinical assessment, or a diagnostic report (after a positive result) or an outcome report - we will need all of these to help with reporting but these are different use-cases. Hopefully we can use the same archetypes and SCT code to help pre-populate and pull through data.
We need to think generally about governance because there will be requests (like this) which do have merit to be added to the ‘master’ template even if it is very likely that local adaptations must be made to fit local policy/ circumstances.
I propose that any ‘formal’ PRs/CRs be done through the CKM mechanism and we set up a small editorial team to decide on what is reasonable to add to the master templates.
I suggest everyone adopts the templates so that they get notifications of changes.
