[EpiData-list] Final test version of rel 5 v2.0.5.148 placed on testing.php page

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Wed Oct 1 14:58:37 CDT 2008

The final build for testing of release candidate of v2.1 EpiData 
Analysis has been placed on the page:

After testing of build 148 by users this will be released as v2.1 
replacing current general download version, which has a number of known 
problems/bugs. Please report any problem with b148 to the list or 
discuss principles here.

There will be NO further implementation of function before release of 
v2.1 - only fixing any bugs reported.

Many issues have been solved - among these unification of commands, new 
lifetable and stattable command and fixing titles for graphs. Among many 
other extensions the possibility of making box plots for several 
variables in the same graphs and further user "set" specifications of 
headers in table and output can be mentioned. The principle of /by 
variable has been implemented where relevant, and a new reporting 
oriented command stattable has been introduced, where summary measures 
can be used.

Unification means that all tables or graphs working with an outcome 
variable will have this variable first, examples:
epicurve outcome timevar
lifetable outcome timevar
tables outcome exposure
ciplot outcome variables

The lifetable takes on two forms with explicit time variable and the 
possibility of giving as well "date of entry" and "date of exit", plus 
an option /MT "missing time" to indicate how to handle missing data on 
the "exit" variable, e.g. when follow-up is still ongoing.  Lifetable 
implements the principles of confidence intervals published by Altman et 
al in "statistics with confidence" e.g. estimates at specified levels of 
survival (median and quartiles) or survival with confidence intervals at 
specified times. The plan is to implement other "Statistics with 
confidence - CI principles" in the coming year.

All commands are documented in the commands.htm file, which will be 
shown when hitting the F1 button with an empty command prompt. (But I am 
quite sure, there could be some deficiencies in the documentation, so 
please report these).

I wish to give credit for the current extension and intensified 
development, which is only possible by funding of full time academic 
programmer for the project by various projects credited on the 
www.epidata.dk/funding.htm page and the willingness of people to take on 
tasks. More about the funding sources and staff listing will be revealed 
when v2.1 is released.

Work now continues with implementation of further SPC (statistical 
process control) graphs, these will  be included for general use 
accompanied by full implementation of translation in v2.5.

Jens Lauritsen
EpiData Association

More information about the EpiData-list mailing list