Seeking expertise on addiction medicine...

Thomas Beale thomas.beale at
Fri Feb 19 03:06:08 EST 2016

On 19/02/2016 04:05, Heather Leslie wrote:
> And the valueset for smoking status are these SNOMED codes:
> (1) Current every day smoker. 449868002
> (2) Current some day smoker. 428041000124106
> (3) Former smoker. 8517006
> (4) Never smoker. 266919005
> (5) Smoker, current status unknown. 77176002
> (6) Unknown if ever smoked. 266927001
> (7) Heavy tobacco smoker. 428071000124103
> (8) Light tobacco smoker. 428061000124105

I hope these are not under the same parent in SNOMED - 1 & 2 are current 
temporal pattern ; 3, 4, 6 are past temporal pattern ; 7 & 8 are current 
(we assume) usage pattern. And 5 doesn't make any sense with respect to 
the rest.

> Mixing frequency and consumption levels with whether they are currently smoking or not - sends shivers down my spine... Can two values be recorded concurrently? What if they fit both (1) 'current every day smokers (I assume of tobacco) AND (2) heavy tobacco smoker. What if they are former heavy smokers but only on some days...?

exactly... if the above codes are from a single parent in SNOMED, they 
can't be used computationally in any safe way. So then they are only 
potentially useful for human consumption and coding, but clearly due to 
the poor design of these codes, they won't provide values that would 
actually enable a physician to correctly characterise past status, past 
usage, current status, and current usage.

Two data points could help, but that's the tail wagging the dog...

- thomas

More information about the openEHR-clinical mailing list