Stream: implementers
Topic: Coding without a code
Elliot Silver (Jun 21 2016 at 19:40):
Is there a reason why we allow a Coding without a code value? I see the description has a case where a Coding with a system but no value means "no appropriate value in the system," but that seems like an edge case. How valuable is it to be able to express that compared to making the code value manditory?
Jeremy Walton (Jun 21 2016 at 19:55):
@Elliot Silver: I would say very. I've worked with a system where everything had to be coded, and practices were constantly asking for new values to be added even if the spec didn't provide it. Unfortunately, the industry moves faster than the standards community. Currently, we represent uncoded data as "ADHoc" and warn the user that the uncoded data will not interact with other data in a meaningful way. If this were a perfect world, you would only allow coded data. A small example edge case, but happens more frequently, is a doctor gets samples of a new drug and offers them to his patients but the drug database the doctor is using doesn't offer a code for that type of drug (Drug database data can be update weekly or monthly etc)
Elliot Silver (Jun 21 2016 at 20:04):
@Jeremy Walton Thanks, that makes sense. And I suppose adding a separate flag for "no suitable value" wouldn't help with the optional code value issue.
Jeremy Walton (Jun 21 2016 at 20:06):
@Elliot Silver Isn't the lack of a value/entry, a "flag"?
Stephen Royce (Jun 22 2016 at 04:29):
Depends on whether you believe in 5th normal form or not. In theory, it's never safe to assume a meaning from the lack of a data value. However, hardly anyone ever designs their systems like this.
Lloyd McKenzie (Jun 24 2016 at 02:47):
There could also be cases where a system might have the display name but not the code - but still know the system
Last updated: Apr 12 2022 at 19:14 UTC
