Dutch COVID-19 template for elderly care

Community Covid-19
,
1

@ian.mcnicoll there are plans coming up to use this by the association off Dutch elderly care doctors(Verenso):wink:
Do you need any help from a medical perspective?

1 Like
2

Absolutely - I would welcome having a small clinical team to review the screening template before it gets updated publicly. It probably only needs 3 or 4 of us but might mean late night webcalls .

A new update will go out today that I got some clinical feedback on, so all ok right now but almost certainly this will change again in a few days - so yours and any other input would be really helpful. It is going to be tricky to stay aligned internationally but NL input most welcome.

Can we try and have a chat?

Ian

2 Likes
3

The association off Dutch elderly care doctors(Verenso) has requested us to localize the template to match this form COVID-19 registratie formulieren.pdf (60.2 KB)
How should I approach this? In the end it needs to be adl2 to work in our system btw. But I could get internal help for that.

2 Likes
4

Could you or someone else do a quick translation - google is fine, just to figure out any significant differences?

5

Form 1: “COVID-19 Initial” Suspicion Date: [date] Symptoms:

  1. influenza-like symptoms (present / absent / unknown)
  2. coughing (present / absent / unknown)
  3. fever (present / absent / unknown)
  4. shortness of breath (present / absent / unknown)
  5. sore throat (present / absent / unknown)
  6. other symptoms (free text)
    all fields optional and fields 1 to 5 “single select” answers, default “unknown”)
    Body temperature (number, 1 decimal)
    Problem / diagnosis (Suspicion COVID-19 / COVID-19 confirmed / no suspicion
    COVID-19 ")
    Other (free text)
2 Likes
6

made two adl2 archetypes to test, they are not technically based on the supplied files (yet) nl.joostholslag__openEHR-EHR-OBSERVATION.covid-19_followup.v0.0.1.adls (3.0 KB) nl.joostholslag__openEHR-EHR-OBSERVATION.covid-19_initieel.v0.0.1.adls (8.6 KB)

7

The only real discrepancy there that I can see is that the suggested ‘risk assessment outcome’ for us is

low risk => which means “No suspicion of Covid-19”
high risk=> which means “Covid-19 suspected”

The problem diagnosis currently only covers ‘positive diagnoses’ e.g. Suspected or confirmed as SNOMED codes.

We have already had issues in the UK where ‘exclusion of of COVID-19’ has been recorded as a SNOMED code in GP systems and is being misinterpreted as ‘Negative testing for COVID-19’ so we very deliberaelt left this out of problem diagnosis.

I have just ‘hidden’ Other symptoms but equally happy to put it back in.

Not sure about the value /safety of allowing other diagnoses in the context of the screening assessment but it is easy to support.

Worth a brief conversation?

The differences are very slight.

Do you know that CKM has an ADL2 export facility? You might want to see if that works.

Ian
Happy to discuss

1 Like
8

Yes, the goal of verenso was to take the template and use it for research/analysis purposes with as little as possible unnecessary questions as possible. So it was meant to match the template:)
Us the risk assessmenent calculated btw? Or just filled out by the clinician?
I know about ckm adl2 , but it’s buggy and doesn’t work for templates:)
Happy to discuss, pm.

1 Like
9

Not calculated - it is actually very simple but depends on the exact questions used, which themselves vary depending on the setting - e.g self-assessment vs secondary care vs primary/community care, and of course national policy which may differ.

We primarily refer to Public Health England guidance (correct as of 17 March)

For that reason we felt it better to leave any algorithm to the app. @rong.chen did build a wee CDS GDL app which he might be able to share but was very much a POC.

10

In the end we decided our new openEHR application was not advanced enough to meet our clients requirements for COVID-19 registration. So we modelled it in our legacy application:s. I’m still willing to help out where I can.

1 Like
11

Thanks Joost, your input was very helpful and we always expected that some people would use the data model as a specification to be implemented ‘natively’.

Please continue to share experiences and requirements even though you might actually deploy in your legacy system?

Ian

3 Likes
12

The missing requirements were not limitations in openEHR but just not implemented in our new application: a way to fill out a follow-up questionaire after 14 days, a customer accessible way to export all covid-19 registration data, and not all customers are familiar with our openEHR based app yet:) It was a good test case for us btw to see how far along our app development was. And it made us/me (even) more enthousiastic about all the possibilities openEHR offers!

4 Likes