# Valid values
**URL:** https://discourse.openehr.org/t/valid-values/12697
**Category:** Clinical (archive)
**Created:** [23 May 2011 12:43 UTC](https://discourse.openehr.org/t/valid-values/12697 "2011-05-23T12:43:22Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@yampeku](https://discourse.openehr.org/u/yampeku)
#### Post date: [23 May 2011 12:43 UTC](https://discourse.openehr.org/t/valid-values/12697/1 "2011-05-23T12:43:22Z")
I have a doubt about valid values (or ranges) for clinical data. Has
clinical sense to define a range with discontinuous domain?
I am thinking if something like 'critical range' with [-100, 0] and
[100, 200] makes sense at all. Or maybe something like 'any positive
or negative number but zero' ]-inf,0[&]0,+inf[
I am not talking about normal or reference ranges, but more on the
kind of ranges that are currently being archetyped
---
### Author:  [@thomas.beale](https://discourse.openehr.org/u/thomas.beale)
#### Post date: [24 May 2011 18:08 UTC](https://discourse.openehr.org/t/valid-values/12697/2 "2011-05-24T18:08:45Z")
At the moment there is the possibility to define intervals like the following (from the ADL spec):
Examples of 2-sided intervals include:
- length matches {|1000|} – point interval of 1000 (=fixed value)
- length matches {|950..1050|} – allow 950 - 1050
- length matches {|0..1000|} – allow 0 - 1000
- length matches {|0..\<1000|} – allow 0\>= x \<1000
- length matches {|0\>..\<1000|} – allow 0\> x \<1000
- length matches {|100+/-5|} – allow 100 +/- 5, i.e. 95 - 105
- rate matches {|0..infinity|} – allow 0 - infinity, i.e. same as \>= 0
Examples of one-sided intervals include:
- length matches {|\<10|} – allow up to 9
- length matches {|\>10|} – allow 11 or more
- length matches {|\<=10|} – allow up to 10
- length matches {|\>=10|} – allow 10 or more
The Interval type doesn’t cope with sets of intervals, i.e. disjoint intervals, and I am not sure that it should directly. If there was a need, we would probably define a new kind of type allowing disjoint intervals. But… is there a real need?
- thomas
---
### Author:  [@yampeku](https://discourse.openehr.org/u/yampeku)
#### Post date: [24 May 2011 18:51 UTC](https://discourse.openehr.org/t/valid-values/12697/3 "2011-05-24T18:51:57Z")
That was exactly my question 🙂
I can see something like the 'critical range' that I defined earlier
being used. If you want to build archetypes about "wrong" status it is
quite possible that you want to define disjoint intervals (if you have
a range of "healthy" values, then the inverse disjoint interval lets
you define the "unhealthy" range)
---
### Author:  [@thomas.beale](https://discourse.openehr.org/u/thomas.beale)
#### Post date: [24 May 2011 19:38 UTC](https://discourse.openehr.org/t/valid-values/12697/4 "2011-05-24T19:38:02Z")
it can be done in current ADL/AOM as follows:
some\_attr matches {
DV\_QUANTITY [at0001] matches {
magnitude matches {|-200..-100|}
}
DV\_QUANTITY [at0002] matches {
magnitude matches {|100..300|}
}
}
The above means that the two DV\_QUANTITYs are alternatives for the single-valued attribute some\_attr (and effectively the same even if some\_attr is a container). So values like -50, 0, + 50 will all fail the above constraints.
Is this close enough to what you want?
- thomas