Re: [Gnumed-devel] basic test types - Sample type prefixes

[scaredycat]

> Complicating matters - but at some point something we may want to 
> factor into design - is that the labs often use a different set of 
> codes by which to ORDER tests as opposed to codes by which to REPORT 
> results. Among other reasons, a single order code can designate 
> multiple tests whereas that single code would be a clumsy way by which 
> to manipulate/search results. Also there are some tests which although 
> you may order a single test (for example serum iron saturation), I 
> think the lab must do a pair of tests (transferrin and iron) and even 
> if that is wrong there are other situations where they must do a 
> "ghost test" (my term) in order to generate a result.


so what are you trying to say ?  It sounds like "this relationship is 
1-many , 1 order code to many results code,  however only one result 
code relates to an order code.
(i.e. there is a functional dependency between result to order code), so 
I need 2 tables , not 1 , for test_type. (test_type_order, 
test_type_result). Are you wanting to
put version fields in the both tables ( to accomodate for lab fashion 
trends)? What about just posting a change as
tablename1(attributes) -1- relationship-name -*-  
tablename2(attributes). Please include in gnumed schema, sir.

Or maybe dig in CDA and see how they do it , and then rename it so it 
belongs to gnumed. (i.e. re-invent the wheel,  that's what a path friend 
said I was doing
when I started my interest in programming, joining the profession that 
re-invents the wheel for a living , instead of getting a life ).