Capillary refill archetype


The archetype openEHR-EHR-CLUSTER.capillary_refill_time.v0 is in draft state, but we’d like to translate it to Swedish and use in a project. However it seems to belong to another CKM ( so we can’t start translation in the international CKM (see image below).

How would you suggest that we go on in order to translate it in a way that can be shared via the CKM? Or can capillary refill be recorded in any other existing archetype that we have missed?

@ian.mcnicoll is the original author of the archetype.


1 Like

Hi Erik,

Maybe you could download it first and try to translate it locally into your language using the Archetype Editor from Ocean.

Hi Erik,

So the process is that you go to the editor in the original 'owning 'CKM (Apperta) and put in a CR - no real reason why we can’t do the translation there in the first instance. either directly in that CKM of perhaps just do the translation locally in AD or AE. If you are signed up to the Apperta CKM you should be able to upload the translation to a branch that the editor (me!) can fold in.

Of course that begs the question as to whether we should just move this archetype to the international repo.I am actually using this archetype in anger in the Uk and ama ware of a gap, (missing term in the list of capillary times) so my suggestion is that we work together in the Apperta CKM (including the translation) then propose uploading to int-CKM.


@Erik - Can I ask what datapoints/values you actually need in your project. This archetype was created a long time ago for Better’s paediatric application and things may have moved on from there.

e.g in a UK setting I cam being asked to represent refill time > 2 seconds (as a term), which is not in the current list. One option is to just use ‘Increased capillary return time’ as per the snomed code 50427001 | Increased capillary filling time (finding) |

What is your actual data value?

1 Like

This currently a cluster - should it be an full OBSERVATION?. There are some discussions about recording at different sites.

I am minded to remove the Capillary return field and only have Capillary filling time with a choice of duration in seconds or internal coded_text of

Less than 2 seconds => SNOMED-CT::45332005 | Normal capillary filling (finding) |
Greater than 2 seconds SNOMED-CT::45332005 | Increased capillary filling time (finding) |


Btw, for everyone else, this is the archetype in question:

I support making this an OBSERVATION archetype. I can’t see which OBSERVATION it’d be relevant to nest this CLUSTER under. It’s definitely event based information, so it belongs in an OBSERVATION.

I also support removing the Capillary return element and making the Capillary fill element a choice of Duration (seconds) and internal coded test with the following values and terminology bindings:

Normal capillary filling (45332005 | Normal capillary filling (finding) |)
Increased capillary filling time (50427001 | Increased capillary filling time (finding) |)
Possibly also Decreased capillary filling time (27157002 | Decreased capillary filling time (finding) |)?

I’m not sure we can extrapolate those two first SNOMED CT terms to mean exactly less than or greater than 2 seconds? If we need those, we need to add them as separate values IMO.

1 Like

Those SCT terms are interpretations of Capillary fill time. So maybe we need a nominal/coded (for formal concepts and their codes) or even narrative (for free text) data point for such an interpretation in addition to the quantitative/more specific data point. Another reason to do so is that archetypes are a kind-of maximal data set.

1 Like

Normal capillary filling (45332005 | Normal capillary filling (finding) |)
Increased capillary filling time (50427001 | Increased capillary filling time (finding) |)
Possibly also Decreased capillary filling time (27157002 | Decreased capillary filling time (finding) |)?

Yes, that also makes sense. I guess that is one of the questions - is there now (was there ever!!) consensus that > 2 secs, represents ‘increased’ and what does ‘decreased’ mean?

I will adapt this an an OBSERVATION. I wonder if it might be reasonable to carry ‘Filling time’ separately as an Interval of Quantity(seconds) (to allow for fractions of a second) and leave these codes as ‘Interpretation’. That would allow us to carry more specific ranges like 1-2 seconds, 1.5 seconds, > 2 seconds alongside the interpretation

1 Like

Can you measure this more specifically than 1 second? Are there machines that measure capillary refill that can do this? Because I don’t think I’d be able to judge by eyeballing it in anything more detailed than whole seconds…

Edit: Also, could you upload the OBSERVATION to the international CKM? We’d like to federate it to the Norwegian CKM too. :smile:

Yes - agree - we probably should phrase th archetype as capillary fill time, as I think that is now a more accepted.

I’m thinking

OBSERVATION.Capillary fill time

Fill time (INTERVAL of QUANTITY) (Either as the ‘raw’ observation or to document the Interpretation)

Interpretation (DV_TEXT/DV_CODED_TEXT) with Silje’s suggestions.


and the usual confounding factors, Extension slot etc.

1 Like

There are different ways of meassuring Capillary Refill Time (CRT), also by Medical Devices.

E. Kviesis-Kipge, E. Curkste, J. Spigulis, L. Eihvalde, “Real-time analysis of skin capillary-refill processes using blue LED,” Proc. SPIE 7715, Biophotonics: Photonic Solutions for Better Health Care II, 771523 (18 May 2010);

So we need Protocol, also SLOT for Device.


Sure - all of the standard observation stuff!!

1 Like

Do we need partial seconds - I don’t really know but I think quantity is more appropriate than Duration in any case, and I have certainly seen fill time expressed as 1-2 seconds, so someone thinks they can estimate it!! and as @varntzen says there may be devices that can measure more accurately. Perhaps there is no real need of INTERVAL but possibly better to put it in alongside QUANTITY for discussion purposes.

This article actually talks about recording of rwsults - amazing!!

  • An abnormal CRT in infants and children over 7 days of age is 3 seconds or more; a normal CRT is 2 seconds or less. A CRT measurement of between 2 and 3 seconds may be considered ‘borderline abnormal’, but some healthy children may have CRT as long as 2.5 seconds.
  • Record measurements using the actual number of seconds (for example, ‘4 seconds’ or ‘2 seconds or less’) rather than using terms such as ‘prolonged’ or mathematical symbols.

Candidate uploaded as a proposal to CKM …

1 Like

Imported, with some minor modifications:

1 Like

I just :heart: love the speed, skills and knowledge of the openEHR community! From a simple question to a revamped, improved and even SNOMED-terminology-bound archetype in four days (even with a weekend included). I did not even have time to answer questions before it was finished…

I believe this redesign will be perfect for the emergency ward application we are working on.

Thank you very much everybody!