gnumed-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gnumed-devel] Measurements (was: New developer searching for a task)


From: Christof Meigen
Subject: [Gnumed-devel] Measurements (was: New developer searching for a task)
Date: 05 Dec 2002 13:39:03 +0100
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2

Hi,

Karsten Hilbert <address@hidden> writes:
> The other area would be input and storage of patient related
> measurements. Actually this is just a subproject of storing
> clinical "observations" but has some of its own requirements.
> With measurements I mean things like height, weight, lab
> results, etc. Basically everything that can be expressed by a
> number or pseudo-number. Things that are not clinical
> narrative. 

The two main problems I see with measurements, are in my
opinion

1) There's a huge variety of parameters that can be measured, and 
   these parameters can be measured with various kinds of equipment
   or methods, which might make a difference. So, a "kind of measurement"
   table must at least contain these colums. 

   Now, either you ship with a huge table containing all parameters
   you can possibly measure and how you can possibly measure them
   (similar to a standard drug database or a ICD-10 reference) which
   causes trouble when people want to enter stuff not in that database
   ---
  
   Or you let users enter their own stuff which makes it hard
   to exchange data between different installations of gnuMed.

2) Normalization: Either you stick a measurement to a clinical
   transaction, which might be not appropriate if people get
   measured elsewhere or do (like diabetis patients?, or 
   people with high blood pressure?) measure themselves and
   tell you, like (I' thinking up something, I'm not a doctor)
   "Two days ago I had 180/130, yesterday I had 210/150, 
   so I thought I better come here" or whatever ... "I was on
   a health resort and the doctor there measured this and that..."
   
   It seems to be against the intention of clinical_transaction
   (with it's involved doctor_ID) to let these things be 
   "clinical transactions". 

   On the other hand, if you make a seperate table 
   "measuring_action" you have to save a date there and since
   most of the measuring _is_ bound to a clinical transaction,
   you save the date two times, which for sure will one day
   create consistancy problems. 


So, if people who are working in this area, vote:

1) 

[ ] There is a comprehensive standard of codes for
    types of measurement, or we can make up such a 
    standard
[ ] We'll simply give a set of some basic parameters 
    and let every gnuMed-User extend/edit this list
    according to his/her needs

2) 

[ ] We extend the meaning of a clinical transaction
    so that no doctor needs to be involved, and neither
    did the patient have to have had any contact with
    clinical personnel at the entered time

[ ] We treat measurements seperately from clinical
    encounters (of course remaining the ability
    to connect them, or attach them to clinical
    episodes etc.)

----

... I would be happy to come up with appropriate 
sql-tables and interface stuff.

           Christof







reply via email to

[Prev in Thread] Current Thread [Next in Thread]