# Modeling reference ranges
**URL:** https://discourse.openehr.org/t/modeling-reference-ranges/14933
**Category:** Clinical (archive)
**Created:** [12 October 2009 22:32 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933 "2009-10-12T22:32:03Z")
**Posts on this page:** 10
**Page:** 1
### Author:  [@pablo](https://discourse.openehr.org/u/pablo)
#### Post date: [12 October 2009 22:32 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/1 "2009-10-12T22:32:03Z")
Hi,
I’m playing around with archetypes trying to model an observation and its reference ranges,
I mean something like “blood pressure” and some range to define what is “hypertension”, but
I can’t found an archetype that defines a reference range for an observation.
Any one has experience in modeling something like this?
An archetype is the correct place to define a reference range for an observation value?
Any ideas?
Thak you!
Cheers,
Pablo Pazos Gutierrez
---
### Author:  [@ian.mcnicoll](https://discourse.openehr.org/u/ian.mcnicoll)
#### Post date: [12 October 2009 22:59 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/2 "2009-10-12T22:59:32Z")
Hi Pablo,
The Quantity datatype in the Reference model has built-in support for Reference ranges so these do have to be modelled overtly in archetypes.
See [http://www.openehr.org/svn/specification/TRUNK/publishing/architecture/rm/data\_types\_im.pdf](http://www.openehr.org/svn/specification/TRUNK/publishing/architecture/rm/data_types_im.pdf)
This makes sense for lab tests etc where each test will report a reference range (which are often lab/analysis method dependent) along with the results themselves.
However, you are talking about something different. There is really no such thing as a reference range for blood pressure, which might indicate hypertension. The definitions of hypertension vary, over time and by locality and the diagnosis will depend on many other factors than just the blood pressure itself.
I think what you may be trying to capture is some thing more like a ‘trigger blood pressure’ which displays an alert to the clinician or initiates some other action if a set of criteria have been reached e.g 3 readings with a diasstolic \> 100.
This is more akin to a guideline or care pathway. You might want to have a look at the work Rong Chen has been doing using archetypes within computerised guidelines for chemotherapy. In this case the archetype does not represent actual patient data but an abstract of ALL patients who might fall within the guideline.
See [http://www.hst.aau.dk/~ska/MIE2009/papers/MIE2009p0653.pdf](http://www.hst.aau.dk/~ska/MIE2009/papers/MIE2009p0653.pdf)
Hope this helps,
Ian
Dr Ian McNicoll
office / fax +44(0)141 560 4657
mobile +44 (0)775 209 7859
skype ianmcnicoll
[ian@mcmi.co.uk](mailto:ian@mcmi.co.uk)
Clinical Analyst Ocean Informatics [ian.mcnicoll@oceaninformatics.com](mailto:ian.mcnicoll@oceaninformatics.com)
BCS Primary Health Care Specialist Group [www.phcsg.org](http://www.phcsg.org)
2009/10/12 pablo pazos \<[pazospablo@hotmail.com](mailto:pazospablo@hotmail.com)\>
---
### Author:  [@system](https://discourse.openehr.org/u/system)
#### Post date: [12 October 2009 23:12 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/3 "2009-10-12T23:12:33Z")
Hi Pablo
The issue is that you do not see the reference model attributes in the archetype editor. A Quantity data type has a normal range and other reference ranges built in.
We do not set the reference ranges in archetypes as these vary and archetypes are the absolute statement about things (what could possibly be true ever, anywhere).
So it is in the form or data that you will get access to the reference range. You could set it in a template (not possible in our tools as yet). Generally the reference ranges come with the results from the lab or a dynamic depending on gender, age etc.
I hope this is helpful – have a look at the data type specs for clarification. The UML is at:
[http://www.openehr.org/uml/release-1.0.1/Browsable/\_9\_0\_76d0249\_1109599337877\_94556\_1510Report.html](http://www.openehr.org/uml/release-1.0.1/Browsable/_9_0_76d0249_1109599337877_94556_1510Report.html)
You will see an optional normal\_range and 0..\* other reference ranges as part of a root abstract class DV\_ORDERED
Cheers, Sam
---
### Author:  [@ian.mcnicoll](https://discourse.openehr.org/u/ian.mcnicoll)
#### Post date: [13 October 2009 00:11 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/4 "2009-10-13T00:11:57Z")
Thanks Sam,
That was helpful but would you agree that is does not make much sense to use a reference range for blood pressure in the same manner as you would for a lab test. I have suggested that if Pablo is trying to set trigger conditions e;g a series of BPs over a particular level, then this properly belongs in the guideline/pathway space, rather than as ref ranges?
Ian
Dr Ian McNicoll
office / fax +44(0)141 560 4657
mobile +44 (0)775 209 7859
skype ianmcnicoll
[ian@mcmi.co.uk](mailto:ian@mcmi.co.uk)
Clinical Analyst Ocean Informatics [ian.mcnicoll@oceaninformatics.com](mailto:ian.mcnicoll@oceaninformatics.com)
BCS Primary Health Care Specialist Group [www.phcsg.org](http://www.phcsg.org)
2009/10/13 Sam Heard \<[sam.heard@oceaninformatics.com](mailto:sam.heard@oceaninformatics.com)\>
---
### Author:  [@pablo](https://discourse.openehr.org/u/pablo)
#### Post date: [13 October 2009 13:48 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/5 "2009-10-13T13:48:15Z")
Hi Ian, thanks for the answer.
I see, so the “reference range” is only for lab test results.
Yes, what I like to do is something that for certain observation values it display something to the physician.
Blood presure was just an example, the observations I have are:
- Glasgow Comma Scale: \<15 is a problem
- Cardiac Frequency: \<60 or \>100 is a problem
- Breath Frequency: \<10 or \>20 is a problem
I’ll look to Rong’s work.
Thanks a lot!
Cheers,
Pablo.
---
### Author:  [@Alain\_Maskens](https://discourse.openehr.org/u/Alain_Maskens)
#### Post date: [13 October 2009 15:06 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/6 "2009-10-13T15:06:31Z")
One suggestion is to use for some of these observations the concept of “critical” reference range, which can then be exploited for implementing ‘triggers’ or rules.
Obviously, the values to be used as critical reference range will depend on appropriate clinical knowledge and guidelines.
Alain Maskens
**CLASS REFERENCE\_RANGE\**
**Purpose**
Defines a named range to be associated with any ORDERED datum. Each such
range is particular to the patient and context, e.g. sex, age, and any other factor
which affects ranges.
**Use** May be used to represent normal, therapeutic, dangerous, critical etc ranges.
**ISO 18308** STR 3.13
**Attributes Signature Meaning**
**1..1**
**meaning** : DV\_TEXT Term whose value indicates the meaning
of this range, e.g. “normal”, “critical”,
“therapeutic” etc.
**1..1** **range** : DV\_INTERVAL The data range for this meaning,
---
### Author:  [@pablo](https://discourse.openehr.org/u/pablo)
#### Post date: [13 October 2009 17:58 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/7 "2009-10-13T17:58:27Z")
Hi Sam,
Thanks for your answer.
That’s exactly what I’ve tried to do, look for a constraint to the reference range of a Quantity.
I think I can extend my templates to support this requirement, may be using the Assertion model. I’m using a custom template model.
I have seen the datatype model, that’s exactly what I whant to define with archetype/template model (if it’s the right place to do it, now I see the archetypes are not a good place to do so).
Thanks a lot!
Cheers,
Pablo.
---
### Author:  [@thomas.beale](https://discourse.openehr.org/u/thomas.beale)
#### Post date: [13 October 2009 22:36 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/8 "2009-10-13T22:36:18Z")
pablo pazos wrote:
> Hi Ian, thanks for the answer.
>
> I see, so the “reference range” is only for lab test results.
Actually, it isn’t; it’s just that that is the most common use. The ‘reference range’ idea in the openEHR quantitative types is not for representing general reference range data, i.e. for all patient types / ages / sex / other circumstances (e.g. diabetic, transplant recipient), but just those values that correspond to the patient in question - i.e. as if the specifically applicable normal range values had been plucked out of the tables. This applies to any other ranges - sub-critical, critical etc. You can have multiple ranges for a given actual value.
- thomas beale
---
### Author:  [@system](https://discourse.openehr.org/u/system)
#### Post date: [14 October 2009 00:39 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/9 "2009-10-14T00:39:16Z")
I agree Ian – though they are always triggers in reality. Australia made more progress on Lipids when it changed labs from reporting actual norms to target norms. Suddenly everyone had high cholesterols and down they have come!
Cheers, Sam
---
### Author:  [@pablo](https://discourse.openehr.org/u/pablo)
#### Post date: [14 October 2009 14:32 UTC](https://discourse.openehr.org/t/modeling-reference-ranges/14933/10 "2009-10-14T14:32:28Z")
Hi Sam, Ian,
Thanks for your tips, I think we are far away from defining guidelines soon, but I think I can extend my templates to define this conditions/triggers there while we don’t have a formal guideline to do so.
The objective of this project (it’s my degree thesis) is: to build the first OpenEHR 100% medical record implementation in my country Uruguay, so if this is successful we can spread the word about OpenEHR works and how other can adopt it to build EHR systems. Now there are many initiatives to build EHRs but not much experience in how to do it, IMHO OpenEHR is the answer we are looking for, but I need something working to show it to the others 😃
Cheers,
Pablo Pazos Gutierrez
[http://pablo.swp.googlepages.com/](http://pablo.swp.googlepages.com/)