Re: [Gnumed-devel] Re: Schema question re labs

[James Busser]

On 7-Feb-08, at 9:14 AM, Karsten Hilbert wrote:

> > In BC
> > Canada, it is however possible for the lab to accept --- on the paper
> > requisition that we complete and provide to the patient --- a
> > reference number that we ourselves can assign, and which the lab will
> > return in field ORC 004 (subject to confirmation).
> Sounds like the perfect candidate for request_id to me.
>
> > A candidate reference number that may be good to send would be the pk
> > which is automatically generated in the table ... any reason it would
> > be any bad idea to use this?
> Absolutely. The PK is a) an implementation detail which shouldn't  
> leave
> the innards of even the local client, b) is illegal to use outside  
> the praxis in some jurisdictions, c) is not intended to carry business
> meaning, d) is not guaranteed to not change during lab request/results
> retrieval (though this is unlikely and best avoided).

In Germany, there is no default value for request_id which must be  
scanned in from the label or manually entered. Are you using GNUmed  
presently to record lab_requests and what workflow do you use to do so?

In Canada and other places where the request_id could be free to be  
decided by the requesting praxis, there is presently no mechanism to  
be able to manage what should go in here. One option would be to  
autocopy (pk + a constant) into the request_id which should get  
around the legal issue if it would have applied. Something like that  
would best be applied as a trigger but I don't imagine triggers are  
easily configurable as a site preference so how best to implement a  
counter or serial function for some installations of GNUmed but not  
others?