Interval events - "change" math function semantics

Technical (archive)
1

Hi, another question about interval events.

I’m having issues understanding how to use the “change” math functions. From the spec http://www.openehr.org/releases/RM/Release-1.0.2/docs/data_structures/data_structures.html#_change_data

  • “change”: this means that the value recorded is the difference between the value now and the value some time previously. It can be positive or negative;

  • “increase”: the value recorded for the change is positive. The name (i.e. ELEMENT.name) chosen for the item in an archetype carries the semantic of positivity e.g. “increase of …​.; rise of…​.; …​.gain” etc;

  • “decrease”: the value recorded for the change is positive. But the name chosen for the item carries the semantic of negativity e.g. “decrease of …​.; fall of …​.; …​. loss”.

Since “change”, “increase” or “decrease” are based on a previous value, shouldn’t the record of the change have a link to the previous value that served as reference to record the change?

I’m not sure in which context a change or increase on some value would be useful without the value that change or increase is compared with.

Checking the terminology for math functions, we have absolute (can be calculated just from given values), and relative (needs a given value and a previous value) math functions. https://github.com/openEHR/terminology/blob/master/openEHR_RM/en/openehr_terminology.xml

These are the ones that I don’t understand, in terms of how to use them in software:

<concept id=“147” rubric=“change”/>

<concept id=“148” rubric=“total”/>

<concept id=“149” rubric=“variation”/>

<concept id=“521” rubric=“decrease”/>

<concept id=“522” rubric=“increase”/>

Does anybody have some example that uses any of those math functions?

Thanks!

2

it is based on a previous value of the measured datum, but that previous measurement is not necessarily in the data set. The use of these codes indicates that the difference is being recorded, not the absolute value. Agree; to make it sensible, one would expect the first value in a HISTORY of EVENTs to be an absolute instantaneous value, with some or all of the other values being differences.

3

Hi Thomas,

I think I’m missing something:

“…previous measurement is not necessarily in the data set.”

From that ^ I understand the reference value from which the “change” is measure & recorded is not on the HISTORY time series.

“…one would expect the first value in a HISTORY of EVENTs…”

And that ^ seems to point on the other way. confused

I’m inclined to have an initial absolute / reference value, then the “changes” (deltas, increase, decrease, …) to be recorded in the same HISTORY.data time series. IMO this needs clinical validation, but my assumption os if “reference” and “changes” are not recorded together, there is missing context for the “change” records that should be resolved in software (if the specs don’t provide a recommendation), and might be an issue for patient safety (also this needs clinical validation).

If we can validate that, and agree on a criteria, we might need to add a clarification/recommendation on that area to the specs.

4

It's probably a good idea to do it like this, but it would be good to get some input from people workig with operational systems, to see how they are using these kinds of math function indicators.

- t