gnumed-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnumed-devel] gmdrugs structure


From: Karsten Hilbert
Subject: Re: [Gnumed-devel] gmdrugs structure
Date: Wed, 20 Nov 2002 11:28:07 +0100
User-agent: Mutt/1.3.22.1i

Hi,

> 1. table adverse effects: 
>  severity should be a foreign key pointing to a table
> "severity_description". Reason: many sources of drug data use different
> (ordinal) scales. It will be difficult to translate one scale to another.
Agreed.

> 2. tables conditions, subsidized_products:
>  those are certainly country specific and should IMHO be marked as such
> (i.e. conditions_au or similiar). I'm not quite sure if those tables
We could also add a field "country" denoting the region to
which this condition/subsidy applies.

> should be part of the standard gnumed drug database, on the other hand
> these tables can be just ignored by a client running in a different
> country.  
Or one could split the SQL into gmdrugs-generic.sql and
gmdrugs-regulations-*.sql which would, of course,
live in country/specific/*/ and which would be selected upon
database configuration.

> mean. I would suggest a three-level hierarchy here:
Sounds good to me.

>  a) the substance itself (i.e. active substance, adjuvant or whatever,
> e.g. paracetamol, fentanyl etc.). This entity  has a name and belongs
> to one class (e.g. NSAID). I might be linked to further information
Depending on the crispness of the definition it may belong to
more than one class (pain killer + antiphlogistic,
antiarrythmic + anaesthetic).

> (interactions, indications, adverse effects) that are supposed as an
> default value if no preparation-specific information is supplied. 
That's, methinks, an excellent way of putting things.

>  b) the specific form of appearance of substances defined by formulation,
> (i.e. tablet, solution etc.), amount of drug, adjuvants etc. This is IMHO
[...]
> level without being ambiguous ('preparation' ?).
Isn't this a "drug" (as in German Medikament vs. Stoff) while
a) would be "substance" ?

>  c) the package-level (containing x units of level b), having a brand
> name, manufacturer and so on. 

I suspect that in 95% of cases there's going to be a 1:1
mapping of level b) to level c) apart from the number of units
contained in a package. But that's probably OK.

This whole 3-level structure makes a lot of sense to me.

> 5. table link_drug_adverse_effects:
[...]
> frequency (one would not expect to have frequent severe events in
> available drugs, though). This is how adverse events are usually handled
Well, not if we _know_ about them.

BTW, we need a way to "lock out" particular
substances/preparations/... from current use (sometimes
temporarily) due to things like newly suspected adverse
reactions. We recently had two children admitted to ICU here
in Leipzig after having been administered MCP by Stada (*not*
with anaphylaxia, that is, but with heart failure in one and
another severe condition in the other).

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346




reply via email to

[Prev in Thread] Current Thread [Next in Thread]