Re: [Gnumed-devel] clinician input wanted: how to implement "coding"

[Jim Busser]

On 2009-11-18, at 1:07 PM, Karsten Hilbert wrote:

> How do you conceptually envision coding to be
> implemented in the client ?
>
> What would you like to use coding *for* ?  (I am not asking
> what you *have* to use coding for currently.)
>
> How do you currently use coding ?

We only presently use ICD9 for billing. In the billing program that I  
wrote, but no longer use, I also used it as the "key" from which to  
reference, from a Patient2Diagnoses link table, the reference to the  
diagnosis table which had only 2 columns (icd9 and description).

> How would you *like* to use coding ?
> Do you see any application of coding ?
> If so, how would you like GNUmed to support that ?

I assume the above are closely related :-)

What quickly comes to mind...

- per patient, a coding system that had a nice population of language  
translations could be invaluable, judging from the (at best) limited  
translations we achieve when a visiting patient brings a foreign, text- 
based document with them. Typically we have no local access to someone  
who is both clinically familiar with what might be the patient's true  
issue AND confidently able to determine what was meant by the foreign  
clinician

- per patient, the above even applies to native speakers within the  
same language, whose meaning of text-based notes can still be hard to  
interpret

- per patient, a coding system could offer a semi-automated index of  
within-praxis diagnostic certainty (or at least of precision),  
depending on the precision that the coding clinician chose to apply,  
for example with ICD9 three or four or five character

- assisting an organ-based or system-based auto-tree within a  
patient's health issues list

- reducing small keyboard errors (mistyping words) albeit at the risk  
of substituting an item that wildly alters the clinical meaning, based  
on mistyping a code character. The UI in this context would take on  
critical importance!

- across-patient searches and groupings of somewhat-similar conditions

- defining potentially-useful language-independent queries that could  
be shared among GNUmed users internationally

- linking to internal or external clinical guideline engines that will  
surely depend on coding systems

- transcoding between coding systems, if that proved safely possible