Re: [Gnumed-devel] (FreeDIams) Re: Interactions engine available with Canadian drugs database

[Eric MAEKER]

Le 25 juil. 10 à 15:02, Karsten Hilbert a écrit :

> > I should keep the link primary key in the Interactions table, because
> > there are many couple of ATC1 ATC2 that link to the same knowledge.
>
> Now that I come to think of it -- no patients ATM -- I start to
> see what you mean:
>
>
>
> table interaction_knowledge
>   pk integer primary key
>   xml varchar - you said that would be language independant
>
>
> table interaction_knowledge_text
>   pk
>   fk - links to interaction_knowledge.pk
>   lang - language of risk and management
>   risk varchar
>   management varchar

Yop read my previous mail.

> table interaction_by_atc
>   pk
>   fk - links to interaction_knowledge.pk
>   atc1
>   atc2

Ok

> table interaction_by_inn
>   pk
>   fk - links to interaction_knowledge.pk
>   inn1
>   inn2

Nop, FreeDiams *does not* provide a direct INN interactions search.
All known INN are stored inside the ATC tree. I've added a specific  
FreeDiams category to store the molecules without any ATC code at "Z00".
These molecules must be translated to english and deutsch (about 40  
mols).

INN search should be :
- ATC 7 chars Code  ==  INN (select french, english, deutsch)
- get the id of the ATC
- look for interacting classes in iam_tree (does my ATC is referenced  
in an interacting class - eg IBUPROFEN -> NSAID)
- look for interactions
        between classes,
        between classes and inn,
        between two inn

> maybe even:
>
> table interaction_by_drug_uid
>   pk
>   fk - links to interaction_knowledge.pk
>   uid1
>   uid2

Nop.

> Karsten
> --

Eric