[EpiData-list] test version of analysis ready (read epx file - new Odds Ratio calculation)

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Wed Oct 27 07:02:55 CDT 2010


A new test release of EpiData Analysis has been placed for download on 
the testing page.
http://epidata.dk/testing.php

The zip file contains a complete working copy, including updated help 
file. When you unzip please maintain folder structure. An EPX formatted 
version of bromar data is included in the zip file.

Changes from previous releases are:

A. Ability to read xml formatted files created by EpiData Manager (EPX) 
files
           (NOT EPZ zipped files)
B. The default principle of Odds Ratio calculation has been changed
     from: MLE based
     to: Standard simple with Robins, Greenland, Breslow 95% CI

to get the new default Odds Ratios use:  /o , to use the same as 
previously: /oex

Output is slightly changed, example:

. tab runok sex /o /oex
Outcome:Marathon race completed ?
Sex of runner    1    0    Total
2    3119    418    3537
1    437    53    490
Total    3556    471    4027
Exposure: Sex of runner = 2Outcome: Marathon race completed ? = 1
Odds Ratio = 0.90 (95% CI: 0.67-1.23)    (Robins,Greenland,Breslow CI)  
Exact
(MLE) Odds Ratio = 0.90 (95% CI: 0.66-1.23)    pORex<=1.0 = 0.2872

With this test release users are more than welcome to challenge the 
whole cycle of creating files in Manager (including import fra Stata, 
Rec, QES files, pasting via clipboard or editor etc), entering data with 
the EntryClient and analysing the data in Analysis. You may use the 
Manager to export data to Stata for comparisons of results.

Anything not working - or comments to the way we have implemented 
aspects - please report to this list or if you find reproducible errors 
it is easier for us, if you can report to the flyspray system - see link 
on the testing page.

regards
Jens Lauritsen



More information about the EpiData-list mailing list