Re: [Gnumed-devel] pharm information as XML

[Horst Herb]

On Thu, 3 Feb 2005 15:00, Ian Haywood wrote:
> > http://sourceforge.net/projects/ioe
>
> This looks very interesting. Especially as I don't recognise the data as
> coming from drugref.org. Where is this data from? What is its licence?
> These questions are all-important if it is to be of use to drugref.org or
> gnumed.

This looks like a BNF (British National Formulary) data excerpt; maybe from 
the Mercy ships or WHO EDF data sets which are loosely based on the BNF?

PS to all: sorry for the delay - Jay Gallagher (from OSCAR) and I are working 
full steam on releasing the new drugref datasets and software; I made a 
conceptual change (yet again) which required some data mangling, but we are 
almost done. 

We stared day before yesterday uploading the first bits and pieces to CVS - 
our "ATC dictionary" which allows to match drug/class/brand/concept names to 
one or more ATC codes for the purpose of decision support (e.g. 
interactions), and a cute functional web interface to manually match 
unmatched terms. Plan is to release every day whatever we find already 
functional under the new concept.

Main changes:
- all data is now stored in (mostly) it's original format instead of imported 
in a monolithic drugref database
- the monolithic drugref database survives in it's current form as the free 
core dataset, unpolluted by possible 3rd party copyrights
- all "external" data lives in separate namespaces (schemas) making it 
possible to host them either within the same database as the drugref core, or 
in separate database (systems) if necessary or practical.
- all databases (both "external" and "core" are accessible via one single 
drugref API
- the drugref API is extremely simplified, I'll explain in an extra email how.
- there is a drugref service server, which acts like a discovery service for 
available databases

Why all this? - It allows using both proprietary, semi-proprietary, 
license-restricted and free data sources to coexist and to be accessed in a 
single consistent way. It also makes database maintenance / updates much, 
much easier than before. It will prevent proprietary providers from making 
malicious claims that we "steal" their information for integration within the 
core. E.g. "MIMS" will be provided as the full and unaltered MIMS via drugref 
API, and MIMS data will not be altered or mixed with non-MIMS data.

The discussions at the Hawaii conference, especially with IP lawyers of which 
several were present, made it clear that this step was absolute necessary 
lest we end up in never-ending court battles.

Done (as in you can really tick it off):
- import of *all* drug interaction data from the Holbrook tables (Thanks to 
the OSCAR team under David Chan).
- import of the full & newest WHO essential drug formulary (a few glitches 
still have to be ironed out manually, but don't affect functionality)
- import of all WHO interaction data
- import of all WHO renal/ hepatic impairment & dosage data
- import of all WHO breastfeeding data
- import of Australian PBS data as provided by HIC
- import of Canadian DPD database
- import of MIMS annual and MIMS abbreviated into separate databases, and 
about 80% of the work done for the new drugref API to access it's *full* 
content / functionality (you can prescribe with it is it is right now the 
same way as you would use it from MDW)

Horst