[EpiData-list] Longitudinal data validity checks

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Mon Jul 9 16:13:09 CDT 2007


One very efficient approach for overview of consistency. E.g. counts of 
id numbers at various times is to use the "count by id" function in 
document menu of EpiData Entry. (or for batch procedures with EpiC).

The idea is to count in different files how many times a unique 
identifier was found:. Example output:

Count of records based on field ID

Selected files:
File 1 = key.rec
File 2 = data\idkey.rec

120 different values for ID found

              Files
ID            1    2
---------------------
0             1    1
1             6    6
2             6    6

The key variable could as here be a simple integer, but could also be a 
string, e.g. of the type:
0-1   0-2    0-3  
constructed as a combination of the id (here 0) and the time of visit 
(here either 1 2 or 3).

In the check file:
id
mustenter
end

timevar
  comment legal
    1 "Initial visit"
    2 "6 months follow-up"
    3 "2 year follow up"
  end
  mustenter
  after entry
     idkey = string(id) + "-" + string(timevar)
  end
end

idkey
 key unique
 noenter
end

 .........................................
Notice that the "key unique" status for idkey is checked at creation of 
the contents, that is in
the "after entry" block of the time variable.

regards

Jens Lauritsen
EpiData Association


More information about the EpiData-list mailing list