- EpiData-list - University of Manitoba Mailing Lists

RE: Sorting of 2x2 tables (Oswego example discussion)
by epidata-list＠lists.umanitoba.ca 09 Nov '07

09 Nov '07

The problem of the /sd option in relation to tab ill cakes vanilla chocolate /oa /sd /ex versus tab ill cakes vanilla chocolate /oa /ex is related to how boolean variables are sorted. Currently in the version 2.0 released there is a bug in sorting of boolean 2x2 variables. The error will be fixed as soon as possible. For numerical variables coded 1/0 or 2/1 there is no problem. The reason I did not catch the boolean Y/N error is, that the tab command gives correct estimates for boolean as well as numerical variables: tab ill cakes vanilla chocolate /oa /ex But if the user adds the sorting (as was needed earlier) : tab ill cakes vanilla chocolate /oa /ex /sd reveals the bug. The way development is done is that before release a number of tests are made, currently around 1250, e.g. after the mentioned tabulation 24 tests were applied for the tabulation of oswego rr values with different options (but not the /sd) option. The oswego rr test has now been extended to cover this issue as well. (18 more tests). A given test can be like this using result variables: tab ill vanilla rolls /o assert samenum(round($OR*100)/100,22.15) assert samenum(round($ORLL*100)/100,5.22) assert samenum(round($ORUL*100)/100,138.39) assert samenum(round($OR1*100)/100,12.26) assert samenum(round($ORLL1*100)/100,1.76) assert samenum(round($ORUL1*100)/100,150.86) assert samenum(round($OR2*100)/100,38.19) assert samenum(round($ORLL2*100)/100,4.06) assert samenum(round($ORUL2*100)/100,1983.85) assert samenum(round($SumOR*100)/100,20.61) assert samenum(round($SumORLL*100)/100,4.93) assert samenum(round($SumORUL*100)/100,126.46) The "hidden" bug arising by the sort option is a good example of why many tests are needed, but I think also that it shows that even with a number of test releases the policy of "silent release" is appropriate. regards -- Jens Lauritsen Coordinator and initiator of EpiData Project http://www.epidata.dk

1 1

Congratulations to all epidemiologist and THANKS to Jens
by epidata-list＠lists.umanitoba.ca 08 Nov '07

08 Nov '07

Dear all, The new version is wonderful. I have the oportunity to make a demo at the begining of this year to a cohort of EPIET fellows. They were using Stata and when they saw EPiData Analysis doing the, them a draft version yet, of tables ..... /oa and epicurve. They got completly breath taken I think Jens deserves all our acknowledge Pedro

1 0

Analysis V2 - Congratulations
by epidata-list＠lists.umanitoba.ca 08 Nov '07

08 Nov '07

Although there are a few minor bugs, these will usually be very obvious and only in the unusual situation. My quick testing on favourite troublesome data sets shows this is a great release. If you know anyone working on outbreak investigations who has not seen Analysis, do them a favour and show them how to do the following on the OSWEGO.REC data: tab ill cakes vanilla chocolate /oa /ex gen i onset=(onsettime div 100) + iif(onsetdate = "04/19",24,0) epicurve onset ill You will make someone very happy, I'm sure. Many of us working in the field have been waiting for something as simple as this (I have been waiting over 25 years). Sure, you could almost do this in EPI6, but only with a lot of advanced knowledge - and forget the exact test except as table by table. Many thanks to Jens & team!! Jamie Hockin Public Health Agency of Canada

1 2

Chris Sellers/FHS/ADPH is out of the office.
by epidata-list＠lists.umanitoba.ca 06 Nov '07

06 Nov '07

I will be out of the office starting 11/05/2007 and will not return until 11/13/2007. I will check emails periodically while I am away. If an issue needs immediate attention, please, contact Fred Grady at FredGrady(a)adph.state.al.us or 1-800-338-8374. Otherwise, I will respond to your email when I return.

1 0

Re: EpiData Analysis - first "silent" non-test release
by epidata-list＠lists.umanitoba.ca 05 Nov '07

05 Nov '07

The bugs in Chi Square values in stratified analysis of 2x2 tables and the correlation command error reported today has been solved. Therefore a first version v2.0 for analysis placed on the general download page. The work continues solving minor problems only affecting special situations, such as percentiles with small numbers of observations (n=10-25) or final adaptation of all help documents to the new options. Only the general help file has at this point been uploaded. When all of these points have been settled a general announcement will be made. This is expected to take 1-2 weeks. Thanks to many users for testing and commenting. regards Jens Lauritsen EpiData Association

1 0

CORRELATE command
by epidata-list＠lists.umanitoba.ca 05 Nov '07

05 Nov '07

Dear Jens, After years of hard work you have brought Epi Analysis to the current status. There are almost all the features of EPI6 and much more. Apart from the pending issues you have outlined, I wanted to draw your kind attention to CORRELATE command (build 126). According to the HELP file, correlate var1 var2 [var3 ...], it should work.. When I run this command, the following happens; table design=box . CORRELATE V1 V2 Invalid floating point operation Probably some thing is wrong in the syntax I am attaching a small sample corr.rec file. Please help. Waiting for the release of this marvelous software. With regards, Dr. Shavinder Singh

1 0

EpiData Analysis - status release candidate
by epidata-list＠lists.umanitoba.ca 04 Nov '07

04 Nov '07

Three aspects needs clarification and further exploration : 1 Chi Square values in stratified analysis of 2x2 tables 2 Saving datafiles with "Epi6 naming convention". Only "First Word" works completely. 3 Percentiles have problems with small numbers of observations Except for these three problem areas I think all major issues have now been solved. The table command syntax has been simplified to: tables ........ (default) will do stratified analysis or a n*k table. tables ....... /oa will create a condensed outbreak table with attack rates and rr tables ........ /cc a condensed table with odds ratios tables ....... /fv will do a "first against all other variables", but currently is limited to 2x2 tables It is my intention to release current as a "silent starter" for v2. By silent I mean without a lot of news due to the limitations mentioned above. A warning is given when the analysis commands (tables, describe, stattable) affected by the problems are used. I expect the problems to be solved within about a week or two at most. The reason for the "silent release" is that some courses are due, which do not use these commands. What could change my decision of release as is would be strong disagreement from users on this list. Please discuss, but if no major concerns show up, the release will happen tomorrow. regards Jens Lauritsen EpiData Association

1 0

Almost final EpiData Analysis v2.0 placed for testing
by epidata-list＠lists.umanitoba.ca 27 Oct '07

27 Oct '07

Two aspects needs clarification and further exploration : 1 Chi Square values in stratified analysis of 2x2 tables 2 Saving datafiles with "Epi6 naming convention". Only "First Word" works completely. After finding the cause for the CHI Square problems and final testing of MERGE command it is my intention to release current test build as first v2.0 of EpiData Analysis. Among the latest changes are implementation of a MERGE command, which replaces relate in analysis and removal of more or less all reported major problems on the Mantis system. Some wishes are still there, but will be looked at later. The MERGE command is a replacement for relate. An example: read patients rename name to ptname merge surgeon /file=surgeon /table rename name to surgname merge hospid /file=hospital /table rename name to hospname the /table option makes sure, that several patients can get information from one record in the hospital or surgeon file. "Table lookup". Without the /table option, the key must be unique in the files. A key consists of the variables (as many as you like, e.g. sex age city) mentioned after the merge command. Notice that the file to merge now must be given with the /file parameter. Users finding problems please discuss here or for definite errors report to the Mantis database. See www.epidata.dk/testing.php Kind regards Jens Lauritsen

1 0

Specification of output in EpiData Analysis version 2
by epidata-list＠lists.umanitoba.ca 18 Oct '07

18 Oct '07

Shavinder Sing attached a larger html file (77kb) and asked : ........................................ The Htm output file of Epi Analysis (build 122) behaves differently from the other htm files. I tried opening one htm file (after running a PGM in the Epi Analysis) in the Open Office 2.2, 2.3 (Writer) and Microsoft Word 2000, 2003, 2007. The data figures are visible. The outline of the tables is faintly visible in the white background. In the print out only data figures are visible, not table outline One has to individually select and choose the appropriate table outline from the table properties box (which is quite tedious for long reports). Is there any solution? I set table design=box. I also tried looking into the “epiout_w for any hint but no luck. Is it not possible to have the same table output for printing, as it is visible in the Epi Analysis window after generating the report? With regards, Dr. Shavinder Singh ........................................ I rejected the mail, since we do not want large attachments (some users are using slow modem lines). But the question is relevant. The output definitions in EpiData Analysis conforms to W3C standards of html and requires also knowledge of the use of css style sheets. The definitions and further description of how to handle this can be found in the document: http://www.epidata.dk/php/downloadc.php?file=epidata_analysis_output_defini… The problem of the lacking frames described is related to the use of style sheets, which are files describing output. For use of output in external browsers users should experiment with: a. copying the style sheet files to an appropriate place b. including the style sheet information in the output file. These two set commands: set STYLE SHEET "D:\EPIDATASTAT\TEST\EPIOUT.CSS" set STYLE SHEET EXTERNAL ON/off Simply: set style sheet external=off will secure that all styles are defined in the html output file. Jens Lauritsen EpiData Association

1 0

Alyshia Landry is out of the office.
by epidata-list＠lists.umanitoba.ca 13 Oct '07

13 Oct '07

I will be out of the office starting 10/13/2007 and will not return until 10/17/2007. If you require immediate assistance, please contact Stephen Drew at extension 5330. Thank-you, Alyshia

1 0