>> actually , looking at the book again, change-current-medication, and issue-script >> would be instances of an action type class which an action attribute is an >> instance of, since actually subtypes of action >> class would be proposed action, abandoned action, implemented action, >> completed >> action, implemented-but-abandoned action. Probably inheritance is for the >> later >> heirarchy, because there might be additional attributes in the subclasses >> e.g. >> when abandoned, when implemented, when completed. >> >> The outcomes of an action are observations (a subtype of observation, that >> has a link to a precedent action), according to the book. > >All this sounds like what OpenEHR does. We don't, however, have >a metamodel, that is to mean we *implement* actions instead of > >Karsten you could stick boolean is_action, ref_action on clin_root_item action ,action_type, action_intent tables , and voila , metamodel of actions in gnumed , no?
_______________________________________________ Gnumed-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnumed-devel
