Re: [Gnumed-devel] Measurements selection, formatting, behaviour

[Karsten Hilbert]

On Mon, Jul 06, 2009 at 08:30:58PM -0700, Jim Busser wrote:

>> It must because the column labels are granular to the day.
>
> Ah... I see... this is a constraint on the measurement grid's columns? 

By design decision, yes.

> Such that if in the backend we had, for a single patient:
>
> July 1 2009 07:42:57  glucose, random   8.2 mmol/L
> July 1 2009 08:44:33  glucose, random   10.2 mmol/L
> July 1 2009 08:44:33  glucose, random   12.1 mmol/L
>
> then these could not be presented in 3 different columns, to display  
> that these had been sampled asynchronously... they would be presented in 
> a single column labeled
>
>       July 1 2009 (or alternative date format) with no time

That is right. The date format is adjustable via the
translation message catalog.

> and the cell would contain some concatenation of the three values?

one per line prefixed with "hour:min"

> Could you maybe model this in test data in the public database which  
> could also be future-included in the test data bootstrap?

Done :-)   I added a second CRP reading on the same day.

> PS if the first result would have been a fasting glucose level and thus 
> typed as having been a different test than the other two results, we 
> would get one cell with the fasting glucose, and another cell with the 
> two in-common random glucose values.

Unless, in the future, we would have aggregated them via a
meta test type.

> What are the implications for gnuplotting such things?

The clinician must know what they want to plot and whether
that is what they are, in fact, plotting. Software cannot
take that responsibility off some human being in the chain.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346