Stream: IG creation
Topic: Structure definition: conditional
Brian Reinhold (Sep 16 2018 at 12:12):
What is the best way to make it clear to an implementer in the profiled Observation structure definition the following conditions:
A Personal Health Device may send a measurement status along with a measurement. The measurement may contain a special value.
If the measurement status indicates a certain set of conditions, a dataAbsentReason with reason 'error' shall be populated and value[x] will not exist.
if the measurement contains a special value, a dataAbsentReason with reason 'NAN', "PINF', or 'NINF' shall be populated and value[x] will not exist
If the measurement status indicates another set of conditions, a meta.security.code shall be populated with 'HTEST' and the value[x] will exist.
If the measurement status indicates another set of conditions, one or more interpretation.coding.code elements will be populated and the value[x] will exist
If multiple conditions are sent, the value[x] is not present 'wins'.
There are likely many ways to do this and end up with the correct result. In the text part of the IG the details are spelled out but it would be nice to see it in the struct def diff
Lloyd McKenzie (Sep 16 2018 at 15:54):
Most of those sound like they're best expressed with the constraint element. Not sure what's meant by "the value[x] is not present 'wins'."
Brian Reinhold (Sep 16 2018 at 23:27):
Most of those sound like they're best expressed with the
constraintelement. Not sure what's meant by "the value[x] is not present 'wins'."
@Lloyd McKenzie What i mean by that is if the measurement status reports multiple events, one (if by itself) would still have a value[x] element and another that would not, the would not case would take precedence and there would be no value[x]
Lloyd McKenzie (Sep 17 2018 at 00:00):
That you can't easily do except by hand-wavy description. Because the instances might be passed around all by themselves. The notion of what "counts" is harder to communicate than "here are the rules around what data can exist".
Last updated: Apr 12 2022 at 19:14 UTC
