FHIR R4 decimal implied precision

[Erik Brakke]

Hello!

Reviewing the R4 specs here: https://hl7.org/fhir/R4/datatypes.html#decimal

The precision of the decimal value has significance:

• e.g. 0.010 is regarded as different to 0.01, and the original precision should be preserved
• Implementations SHALL handle decimal values in ways that preserve and respect the precision of the value as represented for presentation purposes
  

And

Quantity.value
Element Id
Quantity.value
Definition

The value of the measured amount. The value includes an implicit precision in the presentation of the value.

When creating an Observation resource where the valueQuantity has trailing zeroes, I expected the FHIR store would persist and return the trailing zeroes, but I'm having trouble getting that result.  In this example I create 9.2000 and get 9.2 in return.

The FHIR viewer presents the same result

Are my expectations misplaced?

Thanks!

-Erik

[Erik Brakke]

Ah, I didn't think to check the reference implementation, but I see now it aligns with my interpretation of the specs.  Can CHAPI team review please?  Thanks! -Erik

[Chaoqun Li]

Hi Erik,

I did a quick test, the precision keeps well, see attached image.

I think the precision occurs during the display process. If you use the curl command or read from code, that precision is kept.

If you use the postman, you may investigate how to do the settings or contact them for the display.

For our GCP FHIR Viewer, I'll contact with our FE team about it.

Best

Chaoqun

[Brakke, Erik M.]

Ah thanks Chaoqun, you're right.  I thought my client wasn't parsing JSON, and I should have double checked.  Thanks for taking the viewer feedback!

I checked PubSub notifications, and they correctly include the implied precision zeroes.

The BigQuery projections do not, and I see the SQL on FHIR spec already calls this out.  It will be interesting to see if the SQL on FHIR v2 implementation will let us map to a STRING if we really care about maintaining the precision.

Thanks again -- 

Erik

---

From: Chaoqun Li <chao...@google.com>
Sent: Tuesday, September 30, 2025 11:08 AM
To: Brakke, Erik M. <Brakk...@mayo.edu>
Cc: GCP Healthcare Discuss <gcp-healthc...@googlegroups.com>
Subject: [EXTERNAL] Re: FHIR R4 decimal implied precision

 

[Joost Pielage]

Not to be rude or anything, but we have found that the GCP fhir viewer is lacking a lot of things. We went as far of building our own viewer that better represents the actual API. Just so you know never trust the viewer, use the API 

Kind regards,

Joost Pielage

0645351694

--
You received this message because you are subscribed to the Google Groups "GCP Healthcare Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gcp-healthcare-di...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/gcp-healthcare-discuss/CY1PR01MB9224FA52847BED7FC53B48CA951AA%40CY1PR01MB9224.prod.exchangelabs.com.

[Vaughan Hilts]

Hi Joost,

Thanks for your feedback. We plan to roll out a number of improvements to the FHIR Viewer this year.

Did you have any core functionality you feel would help a lot?

Vaughan Hilts | Software Developer | vhi...@google.com

To view this discussion visit https://groups.google.com/d/msgid/gcp-healthcare-discuss/CAOaPFsDtcR26tB9XfrRu7mYno%3Dqi_M_xERqeNctDosMzwNeZ6Q%40mail.gmail.com.

[Vaughan Hilts]

Hey Brakke! 

I think as you noticed, this is one of the few small changes between the JSON document variant of a FHIR resource and the actual resource that is passed down (when parsed). We're showing the JSON parsed variant on the web UI (and many other tools do the same). I'll investigate whether we can maybe have an option to show both.

Let us know if you have feature requests and/or questions about this. 

--

You received this message because you are subscribed to the Google Groups "GCP Healthcare Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gcp-healthcare-di...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/gcp-healthcare-discuss/CY1PR01MB9224FA52847BED7FC53B48CA951AA%40CY1PR01MB9224.prod.exchangelabs.com.

--

[Joost Pielage]

I will ask my team over the years owe have encountered a number of things that don't work. 

Two things for. The top my head:

The telecom search param on patient resources doesn't work. 

I believe _revinclude doesn't work not 100% sure on that one. 

There are some others more complicated search params that dont work. I would need to dig around 

We are on FHIT R4. Happy to have a call if it helps you. 

In our fhir viewer we have also enabled post and put operations. It can be helpful especially in testing context when you submit complicated bundles. Obviously it is a product decision if you want to go that way as it is quite risky on. A production environment. 

Kind regards,

Joost Pielage

0645351694

To view this discussion visit https://groups.google.com/d/msgid/gcp-healthcare-discuss/CAOzuF15honin5VaknUwEV%2BQssm%3DbKWNYKecF8eFORMuOFvET-w%40mail.gmail.com.

[Brakke, Erik M.]

The CHAPI front end team have been very responsive to our feedback.  They implemented a fix for the resource compare tool to flip the differential highlighting within a week or two.

Thanks to the Google team for being open to feedback, and hosting this venue to discuss!

Best --

Erik

---

From: Joost Pielage <joostp...@gmail.com>
Sent: Tuesday, September 30, 2025 1:10 PM
To: Vaughan Hilts <vhi...@google.com>
Cc: Brakke, Erik M. <Brakk...@mayo.edu>; Chaoqun Li <chao...@google.com>; GCP Healthcare Discuss <gcp-healthc...@googlegroups.com>
Subject: [EXTERNAL] Re: Re: FHIR R4 decimal implied precision

 

[Vaughan Hilts]

Thanks for the feedback, Joost!

I did try Patient?telecom=xxx-xxx-xxxx (some phone number specifically) and it does work on the "Search" tab for me. I also included Patient?telecom=xxx-xxx-xxxx&_revinclude=Claim:patient and this does return a set of Patient's and Claim's as well. Maybe your team has stumbled upon a configuration that does not work? If so, we would like to know the steps.

You can report feedback right in the Cloud Console (top right corner) or via our support channels. Of course, you can reach us here too. If you run into these cases, please let us know. Our team wants everyone to have a great out of the box FHIR experience.

For FHIR creation, I can't promise anything right now but we're looking into that while still preserving operational safety.

& Erik, thanks for the kind words. A similar offer is extended, please get in touch if something is not working as it should. 

[Brakke, Erik M.]

Thanks for the update! -Erik

---

From: Vaughan Hilts <vhi...@google.com>
Sent: Tuesday, October 14, 2025 1:10 PM
To: Joost Pielage <joostp...@gmail.com>
Cc: Brakke, Erik M. <Brakk...@mayo.edu>; Chaoqun Li <chao...@google.com>; GCP Healthcare Discuss <gcp-healthc...@googlegroups.com>
Subject: [EXTERNAL] Re: Re: FHIR R4 decimal implied precision

 

We rolled out a number of improvements to the FHIR Viewer, which includes inline resource creation. Thanks for your feedback. 

We're exploring things we can do on the precision front as well, but it's a bit more tricky since it could be a breaking change for some users. 

[Vaughan Hilts]

We rolled out a number of improvements to the FHIR Viewer, which includes inline resource creation. Thanks for your feedback. 

We're exploring things we can do on the precision front as well, but it's a bit more tricky since it could be a breaking change for some users. 

On Tue, Sep 30, 2025 at 7:04 PM Vaughan Hilts <vhi...@google.com> wrote: