Re: [Health-dev] Accuracy in lab test criteria value

[Luis Falcon]

Hi Erika !
On 11/06/2013 01:48 PM, Erika Rocha Aguilar wrote:
> On 06/11/13 09:01 -0300, Luis Falcon wrote:
>>>/ Hello Cedric/
>>>/ On 11/06/2013 08:07 AM, Cédric Krier wrote:/
>>>/ > On 06/11/13 07:44 -0300, Luis Falcon wrote:/
>>>/ >>> How can I assure the accuracy of the measure of the/
>>>/ >>> biochemichal instrument?/
>>>/ >> We can use high precision representation using fields.Numeric,/
>>>/ >> you can specify the number of decimals to be shown./
>>>/ > /
>>>/ > No Numeric fields doesn't increase the representation precision. /
>>>/ > Both Numeric and Float has the digits parameter that define the /
>>>/ > precision. You have to use Numeric only when float computation is/
>>>/ > not right./
>>>/ Thanks for the clarification !/
>>>/ /
>>>/ My interpretation from Erika's mail, is that the analyte result is/
>>>/ already computed and is given as a number, with n decimals (there is/
>>>/ no computation). She wants to make sure there is no rounding in the/
>>>/ decimals needed./
>>>/ So, in this case, we should use the current Fields.Float that we have/
>>>/ for the result, and specify the decimal precision./
>>>/ /
> 
>>>/ My main concern with this though is the initial comment. We don't want/
>>>/ to clutter the result with extra digits that can actually lead to/
>>>/ misinterpretation from the (human) reader./
> 
> /Luis and Cédric/
> 
> /That’s right, I don’t want a human misinterpretation. In one hand, as Luis 
> says, I can change in tryton "script" the field type from float to numeric in 
> Health/Lab/Test results.value./
> 
> /In this case I could have 0,0098000000000…/
> 
>>So you would like to strip the non-significant digits.
>>There is no such option in Python formatting [1] but it could probably
>>be an option in the tryton client. The default behavior is good because
>>it allows correct alignment of numbers when stripping non-significant
>>will break the alignment.
> 
I'm actually thinking that the best way to have a general solution is to
add, per analyte, the precision (along with the Unit of Measure). So,
with this info, we can have the representation for any analyte depending
on the value, for the reports.

If no precision value is given, we keep the default fields.Float with no
specific digits parameter.

What do you think ?
> 
>>[1] http://docs.python.org/2/library/string.html#formatspec
> 
> 
> /An in the other hand, as Cedric says, I can change the ‘default’
> behavior of tryton client allowing the correct alignment of numbers. In
> this case I could have 0,009800 ( 8 char) or 0,0098000 (10 char) . /
> 
> /If I want to do both things, Where can I change it? /
> 
> /Thanks/
> 
> 
> Erika Rocha
> 
> 
> 


-- 
GNU Health : The Free Health and Hospital Information System
http://health.gnu.org
Twitter: @gnuhealth