[Top][All Lists]

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

Re: [Gnumed-devel] gmDrugs

From: Hilmar Berger
Subject: Re: [Gnumed-devel] gmDrugs
Date: Sun, 6 Oct 2002 12:50:38 +0200 (CEST)

On Sat, 5 Oct 2002, Ian Haywood wrote:

> On Fri, Oct 04, 2002 at 08:38:55PM +0200, Hilmar Berger wrote:
> > 
> > Second, it could be difficult to translate available drug databases to the
> > hierarchic class design. The process of adaption would probably mean
> > changes of the supplied information, which in turn means that you can't
> > rely any more on the fact that the drug information is the one supplied by
> > the manufacturer. That bears legal dangers, IMHO.
> That is not any intention. Databases which already have done it 'the
> hard way' by having, say, seperate entries for the interaction between
> rifampicin and every oral contraceptive on the market, can import
> that data as is, and not link to classes at all.
AMIS has class based data on interactions, too. The rest of the data is,
however, not class based and I wonder how it is possible to derive the
full charasteristics of substances in one class from the characteristics
of the 'parent' class. IMHO there are a lot of examples where the
number of differences between the single substances in one class is larger
than the number of similiarities. That would necessitate some way of
'overriding' class properties or at least assigning different levels
of expression of a property to different substances.
> However, AMH and WHO/BNF use class-based monographs, their interaction
> tables are also class-based. MIMS, remarkably, also has class-effects in
> its database.
I was not aware of the existence of any database using class-based
information - is any of those publicly available so that I could have a
look at it ?
> In any case, AMH and WHO/BNF are *not* manufacturer-based (which is why
> we what to use them ;-)
I don't care where some database is coming from if it is only reliable and
complete and doesn't impose to much restriction on use and distribution.
If it is completely free, even better - but that's not a precondition.


reply via email to

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