MedicationStatement with multiple Dosages

Hi All,

I am trying to find an example MedicationStatement with multiple dosages. For example, if a patient in Inpatient ward was prescribed with omeprazole oral capsule 20mg, 2 times a day (8AM and 9PM) for 7 days, then there are 14 doses.

If I need to represent this in Medication statement, Should there be 14 dosages in MedicationStatement ?


No. The dosage includes the concept of a single pattern that works out to be multiple dosing events.

1 Like

Thank you for the response,
if the dosage has single pattern like 2 times a day for 7 days.

  1. how the time is calculated - for Inpatient medication time depends on the Hospital / ward.
  2. If the dosing events has different times how this can be communicated in the Medication Statement ? for example some this medicine could be administrated 8AM ?


Have you looked at this

I’ve not read it in detail but it sounds similar

That’s correct - still in development but should cope with that dosage instruction ok. Bear in mind that all current GP systems only carry a single dose/text string, and the use of structurted dosage is very inconsistent in secondary care systems. We are very close to figuring out how to do this in a system-neutral way in FHIR and we have proved it can be done in openEHR :wink: The dose/timing models are very similar.

@Tharma- your example is a single dosage instruction, actioned by 14 administrations. You would not normally carry all of these adminstration events with a MedicationStatement (though you can as linked resources) much more common is to carry e.g. details of the last administration.

I don’t want to give the impression that FHIR is waiting for a new dose syntax before it can be used.

If you have a good understanding of all your own system’s data items you can map them onto the data items of FHIR now.

There are many examples of complex FHIR dosage already in the standard.

But mapping between complex models is never easy, and the new dose syntax will help if you are trying to standardise the way dose is represented across a wide range of settings and systems - many of which don’t capture that data in a structured way currently.

I do know there is a need for that work in the UK, to solve a wider problem, but I am not sure that is what the questioner is trying to do here.

To add to Ian’s answers about the specifics:

1 FHIR doesn’t calculate times. If you have that requirement you need to manage that in your own system code. However FHIR does let you represent the planned time, if there is one, and the eventual time also.

2 FHIR allows you to specify the time of the dose in the request. You can specify the actual time that it happened in a MedicationAdministration. However as we have discussed already MedicationStatement is a more general purpose, higher level summary, and would not normally cover all the times of doses. That is not its job.

Hi Rik,

“I don’t want to give the impression that FHIR is waiting for a new dose syntax before it can be used.”

At one level, of course I have to agree since the resources exist … however… this is an area where the value only comes if all potential stakeholders agree on what parts to implement and how. It is such a complex area and is also heavily dependent on the particular terminology in use and national prescribing patterns. So I would argue that practical deployment has to take place within a community of practice around agreed implementation guidance.

That work is taking place in the UK and very close to being ready for deployment, albeit within the restricted context of common GP/community prescriptions.

I agree with your other answers …

Specific times/dates can be carried in the timing structures but how these are derived from more abstract timings like one tablet three times per day depends on the care setting/institution, which may have specific ‘drug round’ times.

Hi Ian

I meant to say that while obviously everyone will need analysis, mapping, consensus etc, you are not bound to this particular dose syntax in order to use FHIR.

The question wasn’t UK specific as I understood it (although this is a UK hosted board).

HI @riksmithies ,

Are you saying that MedicationAdministration resource can be used to represent already administrated dosing events and planned / future dosing events as well?


HI All,

There seems contradicting views among the FHIR community on this? found some answers in Zulip as well.
do you think the NHS / CareConnect profile should address this? and the status of each dosing event whether its given or not giving delay reason etc should also be considered.

for both medication Request or medication statement this issue is there.



Hi Tharma

No, MedicationAdministration resource is not intended for future events.

If you look at the effectiveDate attribute it says “A specific date/time or interval of time during which the administration took place”

That is in the past.

Also see here:
where it states “A record of a patient actually consuming a medicine, or if it has otherwise been administered to them”

If your system captures this level of detail, you use a request to say when to say when medication should be given in future, not an administration. Using a future date in an administration would be incorrect, as far as I am aware.

I can’t see any ambiguity, but if you can, please give references so that we can take steps to improve the FHIR documentation.

Also note that, as mentioned on these threads, not all systems will use the full medication resource pattern (request, dispense, administer). Some systems, in the UK GP sector for instance, are expecting to just “summarise” things in MedicationStatements - not covering the taking of every tablet. Few systems outside of secondary care track every dose that is taken by the patient themselves, for instance.

Also re: >and the status of each dosing event whether its given or not giving delay reason etc should also be considered.

If this level of detail is supported (which it would not, for many systems) this would be captured with MedicationAdministration.statusReason (“a code indicating why the administration was not performed”)

Btw we still don’t know the scope or setting of the system you are developing, which would be helpful for us to guide you better.


1 Like

this is coming from a charting system

As it’s going via transfer of care it would be a MedicationStatement but how to represent the dosage choice for the paracetomol?

Is support for this in MedicationRequest using intent = option.
One way of doing this would be to have 1 MedicationStatement which is basedOn two MedicationRequests both of which have intent=option.

Hi Kev
That’s a possibility yes. It seems that is what “option” is for but I haven’t seen an example of it in use. I am not sure where you express the condition. This sounds like one for zulip if you haven’t asked there already. Btw Is ToC limited to MedicationStatements as with GPs?

This is actually raised be myself and it is for ToC for eDischarge, and to answer your question, yes ToC is limited to a MedicationStatement which causes the problem.

I’ll check with the clinician responsible for ToC. (he’s at the rewired hackathon)