gnumed-devel
[Top][All Lists]
Advanced

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

Re: [Gnumed-devel] Re: GNUmed substance schema was Re: GNUmed and FreeDi


From: Karsten Hilbert
Subject: Re: [Gnumed-devel] Re: GNUmed substance schema was Re: GNUmed and FreeDiams
Date: Tue, 4 Jan 2011 21:10:24 +0100
User-agent: Mutt/1.5.20 (2009-06-14)

On Tue, Jan 04, 2011 at 11:37:05AM -0800, Jim Busser wrote:

> I am hoping:
> 
> 1) that I can cleanly populate ref.branded-drug from my FreeDiams hybrid 
> output table

Why not ?   Perhaps a need exists to define "cleanly".

> 2) that substances prescribed in GNUmed can keep a linkage (even if indirect) 
> to
> 
>       ref.branded-drug.pk

Why, clin.substance_intake.fk_drug_component ?

> **********
> re #1
> 
> Canadian brand names are sometimes (but not always) specific to their drug 
> strengths, for example the second manufacturer names all strengths ""ZYPREXA 
> ZYDIS"
>       
>       "02229250"|"ZYPREXA  -2.5MG"|"N05AH03"|"2.5"|"MG"
>       "02229269"|"ZYPREXA - 5MG"|"N05AH03"|"5"|"MG"
>       "02229277"|"ZYPREXA  -7.5MG"|"N05AH03"|"7.5"|"MG"
> 
>       "02243086"|"ZYPREXA ZYDIS"|"N05AH03"|"5"|"MG"
>       "02243087"|"ZYPREXA ZYDIS"|"N05AH03"|"10"|"MG"
>       "02243088"|"ZYPREXA ZYDIS"|"N05AH03"|"15"|"MG"
> 
> conflicts with
> 
> >     From: Karsten Hilbert <address@hidden>
> >     Date: January 2, 2011 6:15:10 AM PST
> >     To: address@hidden
> >     Subject: Re: [Gnumed-devel] Re: GNUmed and FreeDiams
> >     ...
> >     Brands across different strengths must have (ever so
> >     slightly) differing names, say
> >             "MetoPharm comp 100/12.5"
> >             "MetoPharm comp 100/25"

I would, even if just for clarity, expand on the names:

        "02243086"|"ZYPREXA ZYDIS (5)"|"N05AH03"|"5"|"MG"
        "02243087"|"ZYPREXA ZYDIS (10)"|"N05AH03"|"10"|"MG"
        "02243088"|"ZYPREXA ZYDIS (15)"|"N05AH03"|"15"|"MG"

> why, in branded-drugs, do we need UNIQUE (description, preparation) meaning 
> UNIQUE (brand_name, form) ?

Hm, let me think. Likely because there's no more fields
to sanely be unique on:

 description 
 preparation
 atc_code
 is_fake
 fk_data_source
 external_code
 external_code_type

One *really* wants to be unique on the name itself but then
people will complain about not being able to enter
"Paracetamol" as tablets as well as solution or
suppositories.

> **********
> re #2
> 
> >     From: Karsten Hilbert <address@hidden>
> >     Date: January 2, 2011 6:15:10 AM PST
> >     To: address@hidden
> >     Subject: Re: [Gnumed-devel] Re: GNUmed and FreeDiams
> >     ...
> >             ... is there some way that the addition of a
> >             substance to a patient drug list would keep resolvable the
> >             [specific-to-strength] external_code? 
> > 
> >     I am not sure I follow this question.
> 
> 
> My question "is there some way" is maybe better understood by ignoring 
> [specific-to-strength] …  In table clin.substance_intake, there is a warning 
> on column
> 
>       fk_substance    (FKEY ref.consumable_substance.pk)
> 
> which warns
> 
>       DO NOT TRY TO USE THIS TO FIND OUT THE BRAND. IT WILL BE WRONG.
> 
> is it wrong because the referenced substance may have been linked-to multiple 
> brands, and therefore is unable to completely resolve which one?

The meaning of the

clin.substance_intake.fk_substance
 -> ref.consumable_substance.pk
  -> ref.lnk_substance2brand.fk_substance
   -> ref.lnk_substance2brand.fk_brand
    -> ref.branded_drug.pk

chain is

        those are the registered-with-GNUmed brands containing
        the substance the patient happens to be taking

> *But* the column ref.lnk_substance2brand.pk *will* resolve it --- provided 
> the user *bothered* to specify/distinguish the Brand, right?

It may technically resolve it but it will mostly be wrong.
If you want to find the brand of a substance intake you need
to look at fk_drug_component rather than at fk_substance.

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]