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

[Jim Busser]

On 2011-01-03, at 4:56 PM, Eric Maeker wrote:

> You should considere to store the route as well. FreeDiams should be
> able to identify/switch drugs using unique sets of:
> - inn name (not atc code)
> - dosage + dosage references
> - route
> 
> Without the route FreeDiams Will not be able to identify/switch
> dexamethasone for eg...

Even if an EMR would also provide FreeDIams with "route":

        INN drug name, concat (strength, units), route

the addition of route does not provide any unique match, because even the above 
combination can exist more than once in FreeDIams (say, multiple commercial 
drug providers).

Route is a combined property (an intersection) of:
1) how a given commercial drug may *possibly* be used (even if not designated 
"officially" for that use) AND
2) how the doctor *instructed / prescribed* the molecule to be used

I am looking to better understand what FreeDiams can handle from an EMR:

1. we know it can handle demographics, and receive clinical data
2. we know it can pass back to the EMR a "final list" of prescribed 
medications, with granular data
3. we know it can warn (inside FreeDiams) of
        interaction between prescription selection <-->  allergy/intolerance
                even when the allergy/intolerance is passed as
                        inn
                        atc
                        uid
        interaction between
                prescription selection <-->  prescription selection
4. we know it can handle drugs that have been "passed" to FreeDiams
                --> for prescribing (but only if passed with UID)
                --> for interaction testing (but only if passed with UID)

I agree that FreeDiams could not easily _build a prescription_ from drugs 
passed without UID.

But couldn't FreeDiams test interactions using just exchange-in ATCs?