ISM_TRANSITION

Technical (archive)
1

Hi,

I started wondering about different representations of ISM_TRANSITIONs of the ism_transition attribute in the ACTION archetypes. I found that the below example:

ISM_TRANSITION matches {
current_state matches {
CODED_TEXT matches {
code matches {[openehr::524]}
}
}
careflow_step matches {
CODED_TEXT matches {
code matches {[local::at0001]} – Planned
}
}
}
ISM_TRANSITION matches {
current_state matches {
CODED_TEXT matches {
code matches {[openehr::524]}
}
}
careflow_step matches {
CODED_TEXT matches {
code matches {[local::at0004]} – Requested
}
}
}

could be compressed into this:

ISM_TRANSITION matches {
current_state matches {
CODED_TEXT matches {
code matches {[openehr::524]}
}
}
careflow_step matches {
CODED_TEXT matches {
code matches {[local::at0001, – Planned
at0004]} – Requested
}
}
}

The examples have the same semantic meaning. However, at data entry the second example would mean that a choice must be made concerning the careflow_step, i.e. either at0001 or at0004. Should the archetype editors support both examples? The first example is the only one I’ve seen in archetypes…

Regards,

Mattias

2

Hi again,

I should probably also mention that we need some document that provides recommendations about how archetypes should be created when there are more than one way to do it like the example below. Otherwise there is a risk that there are too many archetypes with different representations that essentially mean the same thing.

The rules that Thomas talked about in this thread http://www.openehr.org/advice/implementers-priv/msg00311.html , i.e.

- if there is a plug-in C_DOMAIN_TYPE type available, always use it to 

represent constraints for the corresponding RM type
- if there is a syntax equivalent for this type, then use it
- otherwise use generic ADL

Should also be documented somewhere.

Regards,

Mattias

2006/10/31, Mattias Forss <mattias.forss@gmail.com>:

3

Mattias Forss wrote:

Hi again,

I should probably also mention that we need some document that
provides recommendations about how archetypes should be created when
there are more than one way to do it like the example below. Otherwise
there is a risk that there are too many archetypes with different
representations that essentially mean the same thing.

The rules that Thomas talked about in this thread
http://www.openehr.org/advice/implementers-priv/msg00311.html , i.e.

- if there is a plug-in C_DOMAIN_TYPE type available, always use it to

represent constraints for the corresponding RM type
- if there is a syntax equivalent for this type, then use it
- otherwise use generic ADL

Yes, such rules would certainly help. They can be even implemented in
the software. For example the editor should be able to recognize
different ways of representing the same archetype and "compress" it with
the best approach.

Cheers,
Rong