[EpiData-list] double entry verification : problem when entering numerical fields

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Sat Apr 11 03:24:11 CDT 2009


Dear Valérie,

There is perhaps a way to get round your problem (and I think you looked in
the right direction). Let's try this:

1- define a character field with 3 spaces ('fieldchar')

2- define a numeric field of 3 digits ('fieldnum')

3- in your check file, state that the numeric field is filled from the
character field by simply writing something like:

FIELDNUM 
 BEFORE ENTRY
  fieldnum=integer(fieldchar)
  goto nextvariable
 END
END

4- in the checkfile for the 'nextvariable' field write:


NEXTVARIABLE
  BEFORE ENTRY
    IF (fieldnum <> integer(fieldchar)) THEN
      CLEAR fieldnum
      CLEAR fieldchar
      GOTO fieldchar
    ENDIF
  END
END


In this case, when the operator enters a string of 3 numbers in the
character field, everything is OK; when the second operator will enter
another string of 3 number, the error message should be prompted. And if the
operator enters a string of 3 characters amongst which one is not a number
(even in the middle of the string), such as "1a1" in your example, the
prompt will remain stuck in the fieldchar until a correct value is entered.

Of course, your fieldchar variable will not be of any use later on, but that
shouldn't be a problem.

Hope this helps.. and have a nice Easter break!

Yves


Yves MARTIN-PREVEL, Epidemiologist, MD, MSc, PhD
'Nutripass' IRD (WHO Collaborating Center for Nutrition)
UMR204 IRD/Montpellier1/Montpellier2/SupAgro
Institut de Recherche pour le Développement (IRD)
BP 64501 - 911, avenue d'Agropolis 
34394 MONTPELLIER CEDEX 5 (France)
yves.martin-prevel at ird.fr
Tel: +33(0)4.67.41.61.70 - Fax: +33(0)4.67.41.61.57 
https://www.mpl.ird.fr/ur106/index.htm
http://www.cilss.bf/nusapps/


-----Message d'origine-----
De : epidata-list-bounces at lists.umanitoba.ca
[mailto:epidata-list-bounces at lists.umanitoba.ca] De la part de
epidata-list at lists.umanitoba.ca
Envoyé : vendredi 10 avril 2009 16:29
À : epidata-list at lists.umanitoba.ca
Objet : Re: [EpiData-list] double entry verification : problem when entering
numerical fields

Currently I only see this solution - assuming you have your data for 
entering now:
 
1 Based on your first file go to tools and use "copy structure".
2 In the copy your system enter the data separately (without the 
simultaneous comparison).
3 Use Document->Validate duplicate files
4 for the errors that show up you can compare with the original recordings.

Just now the priority is to work on the long term sustainable changes, 
which leaves it quite uncertain how soon any detailed problem will be 
solved. (Notice the "how soon" - not "if").

Also in particular in regards to clinical trials there will be a 
complete "package" assisting clinicians to follow the Good Clinical 
Practice Guidelines. I participated in a working group within the ECRIN 
network (Http://www.ecrin.org/). The report defines how we see GCP 
defined in practice, - and EpiData Software will follow these 
guidelines. ECRIN is an industry INdependent researcher network.

ECRIN: European Clinical Research Infrastructures Network
The report: GCP-compliant data management in multinational clinical 
trials. ECRIN-2, Deliverable D10, Version 1, 15 September 2008. is 
avaialable at:
http://www.ecrin.org/index.php?option=com_docman&task=doc_download&gid=64&It
emid=68
 

regards Jens Lauritsen
EpiData Association

epidata-list at lists.umanitoba.ca wrote:
 > We intend to use EpiData for a double entry verification.
 > Unfortunately, we have found a problem when entering numerical fields.
 > If the keyboarder types  001  in a 3 digit numerical field, Epidata
 > stores 1 (whitch is quite normal). But during the verification step, if
 > they enter 001 again ... there is a message to remind them that there is
 > a difference. Of course, the difference is between the stored value 1 et
 > the first typed digit 0.
 >
 > This is a problem for us.
 > In the framework of Clinical Trials, the operators have to enter exactly
 > what is written on the CRF.
 > A solution could be to define the field as a text one. But of course
 > this allows to enter letters instead of digits, and that is not enough
 > secure.
 >
 > We also tried to associate a numerical range (000-999) to the text 3
 > characters field (in a .CHK file). With this kind of control, typing
 > a11  is denied, but typing  1a1  is allowed...
 >
 > Do you have an other solution or idea to solve this problem?

_______________________________________________
EpiData-list mailing list
EpiData-list at lists.umanitoba.ca
http://lists.umanitoba.ca/mailman/listinfo/epidata-list



More information about the EpiData-list mailing list