Re: [Gnumed-devel] EMR tree display of allergy

[James Busser]

On 25-Sep-08, at 8:20 PM, James Busser wrote:

> ... despite the temptation to ask it "at least once", it does  
> little good to have asked it *only* once

I may be here onto something...

The value in having asked about allergy is really only (ever) in  
relation to any treatment being given. So yes, it is true, if a 20- 
year old patient has had a lifelong history of penicillin allergy,  
and at age 18 (in the year 2006) joins your praxis, you can argue  
that it would be useful -- in *case* there would be a need for an  
antibiotic to be prescribed -- to have this information in the database.

However if the patient had no need for antibiotics until now (in  
2008), it is only now that it becomes important to have documented  
anything.

If we now turn it around to suppose we had two patients, both of whom  
had been in the praxis x 5 years:
- both, at entry, had their allergy information asked: = no allergies
- both, over the next 5 years, receive various prescriptions
- allergy information for one was re-asked only last week, but since  
the information is unchanged, let us suppose that this re-asking was  
not captured in the EMR

Now I see one of these patients (normally looked after by somebody  
else) and I must prescribe something. I can only know that the  
allergy question was asked as long as 5 years ago. Does that help me?

If I now prescribe something, and there is an adverse reaction, and I  
am asked "did you ask if the patient had allergies?" and I say "well,  
the EMR said they had none" my answer would evade the question.  It  
is possible that I did not re-ask, or --- if I did re-ask --- I  
cannot prove that I re-asked.

Would it make any difference if I could know that the patient had  
never in 5 years had the question re-asked? Or at the other extreme,  
if I could know that it was asked only 1 week ago, when some other  
prescription was given?

Ideally, I would be able to say that I *confirmed*, at the time that  
I made the prescription, any allergy information present or absent  
from clin.allergy.

This resolves a gap between what Karsten needed -- the ability to  
document that allergy state was *ever* asked -- and what I felt this  
left missing, which was the capacity to see when it was last re- 
affirmed, and/or myself to re-affirm it.

I would expunge from clin.allergy_state a value for "has_allergy"  
because data here is properly derivative from (yet risks allowing a  
value discordant from) what is in the clin.allergy rows.

Up until the time that clin.allergy_state was ever for the first time  
asked, this table would contain no row for the patient. As soon as a  
row is created, its value modified_when would record the date and  
time of the asking of the allergy details. The *content* and  
*certainty* of the allergy details we would simply record in  
clin.allergy (not in clin.allergy_state).