- EpiData-list - University of Manitoba Mailing Lists

Lifetable - Error on P50 calculation
by epidata-list＠lists.umanitoba.ca 30 Mar '09

30 Mar '09

Hello, I am very disappointed about error for the calculation of p50 in lifetable. Probably, Ill stop the use of this command (I had very hoped in this program..). I join an example with 3 errors. I can address the file. Thank you. Bernard BRANGER, Nantes, France . lifetable arret durall /p50 /by=codemat /t /noci Life Table: ARRET by codemat Total N Cases n Expected Cases Time Minimum Maximum Sum time Hazard Ratio Time S=0.50 Time Diff. 1 19 16 11.56 1 29 249 Ref. 10.00 Ref. 2 10 6 5.97 1 29 127 1.07 11.00 1.00 3 10 5 6.81 1 28 147 0.02 15.00 5.00 4 10 6 3.04 1 14 77 842818.30 8.00 -2.00 5 22 14 16.74 1 29 344 0.09 14.00 4.00 6 18 13 13.09 4 27 284 0.91 26.00 16.00 8 14 13 7.27 1 29 160 87399.95 23.00 13.00 9 18 15 10.93 0 29 239 223.32 12.00 2.00 10 20 17 15.50 1 28 333 4.12 30.00 20.00 11 7 6 3.22 1 22 72 235936.51 11.00 1.00 12 21 11 14.84 1 29 312 0.02 18.00 8.00 13 20 15 13.92 1 28 303 3.09 13.00 3.00 14 28 15 20.54 3 29 437 0.02 23.00 13.00 15 20 8 16.57 0 29 344 0.00 27.00 17.00 Outcome: ARRET = 1 Log Rank test of equality of survivor function: Chi2(13)= 21.092 P= 0.0711 LR test of homogeneity among groups Chi2(13)= 17.547 p= 0.1755 Interval NAt risk Deaths Lost Survival 1 1-2 19 1 0 0.95 2-3 18 2 0 0.84 4-5 16 1 0 0.79 6-7 15 2 0 0.68 8-9 13 1 0 0.63 9-10 12 1 0 0.58 10-11 11 2 0 0.47 13-14 9 1 0 0.42 14-15 8 1 0 0.37 16-17 7 1 0 0.32 19-20 6 1 0 0.26 21-22 5 1 0 0.21 25-26 4 1 0 0.16 27-28 3 0 1 0.16 28-29 2 0 2 0.16 2 1-2 10 1 0 0.90 4-5 9 1 0 0.80 8-9 8 2 0 0.60 9-10 6 0 1 0.60 11-12 5 1 0 0.48 15-16 4 1 1 0.36 28-29 2 0 2 0.36 3 1-2 10 2 0 0.80 7-8 8 1 0 0.70 9-10 7 0 1 0.70 13-14 6 1 0 0.58 15-16 5 1 0 0.47 24-25 4 0 1 0.47 25-26 3 0 2 0.47 27-28 1 0 1 0.47 4 1-2 10 1 0 0.90 5-6 9 1 1 0.80 6-7 7 1 0 0.69 8-9 6 2 0 0.46 9-10 4 1 1 0.34 13-14 2 0 2 0.34 5 1-2 22 1 0 0.95 3-4 21 4 0 0.77 5-6 17 1 0 0.73 8-9 16 1 0 0.68 10-11 15 2 1 0.59 11-12 12 1 0 0.54 14-15 11 1 0 0.49 16-17 10 1 0 0.44 25-26 9 1 0 0.39 27-28 8 0 2 0.39 28-29 6 1 5 0.33 6 4-5 18 1 0 0.94 5-6 17 1 0 0.89 6-7 16 1 0 0.83 7-8 15 1 0 0.78 8-9 14 1 0 0.72 11-12 13 1 0 0.67 12-13 12 1 0 0.61 13-14 11 1 0 0.56 15-16 10 1 0 0.50 16-17 9 1 0 0.44 17-18 8 1 0 0.39 19-20 7 1 0 0.33 21-22 6 1 0 0.28 26-27 5 0 5 0.28 8 1-2 14 1 0 0.93 3-4 13 2 0 0.79 7-8 11 2 0 0.64 8-9 9 2 0 0.50 9-10 7 1 0 0.43 10-11 6 1 0 0.36 15-16 5 1 0 0.29 18-19 4 2 0 0.14 25-26 2 1 0 0.07 28-29 1 0 1 0.07 9 0-1 18 1 0 0.94 4-5 17 1 1 0.89 6-7 15 2 0 0.77 7-8 13 1 0 0.71 9-10 12 2 0 0.59 12-13 10 2 0 0.47 14-15 8 2 0 0.36 16-17 6 1 0 0.30 20-21 5 1 0 0.24 25-26 4 2 0 0.12 28-29 2 0 2 0.12 10 1-2 20 1 0 0.95 3-4 19 1 0 0.90 5-6 18 1 0 0.85 7-8 17 1 0 0.80 8-9 16 2 0 0.70 10-11 14 1 0 0.65 14-15 13 1 0 0.60 15-16 12 1 0 0.55 18-19 11 2 0 0.45 22-23 9 1 0 0.40 23-24 8 2 0 0.30 25-26 6 1 0 0.25 26-27 5 2 0 0.15 27-28 3 0 3 0.15 11 1-2 7 1 0 0.86 4-5 6 1 0 0.71 10-11 5 1 1 0.57 11-12 3 1 0 0.38 15-16 2 1 0 0.19 21-22 1 1 0 0.00 12 1-2 21 1 0 0.95 3-4 20 1 0 0.90 4-5 19 0 1 0.90 7-8 18 2 0 0.80 9-10 16 2 1 0.70 10-11 13 1 0 0.65 13-14 12 0 1 0.65 14-15 11 0 1 0.65 16-17 10 2 0 0.52 17-18 8 0 1 0.52 18-19 7 1 0 0.45 20-21 6 1 0 0.37 27-28 5 0 1 0.37 28-29 4 0 4 0.37 13 1-2 20 1 0 0.95 4-5 19 0 1 0.95 7-8 18 1 0 0.90 9-10 17 3 0 0.74 12-13 14 3 0 0.58 13-14 11 3 0 0.42 15-16 8 1 0 0.37 16-17 7 1 0 0.32 25-26 6 1 0 0.26 26-27 5 1 1 0.21 27-28 3 0 3 0.21 14 3-4 28 2 0 0.93 5-6 26 0 1 0.93 6-7 25 1 0 0.89 8-9 24 1 2 0.85 9-10 21 1 2 0.81 10-11 18 0 1 0.81 13-14 17 0 3 0.81 14-15 14 0 1 0.81 15-16 13 1 0 0.75 16-17 12 1 0 0.69 19-20 11 0 1 0.69 21-22 10 2 0 0.55 23-24 8 1 0 0.48 24-25 7 1 0 0.41 27-28 6 1 0 0.34 28-29 5 3 2 0.14 15 0-1 20 1 0 0.95 3-4 19 1 0 0.90 5-6 18 0 1 0.90 7-8 17 0 1 0.90 11-12 16 0 1 0.90 13-14 15 1 0 0.84 14-15 14 0 3 0.84 15-16 11 2 0 0.69 19-20 9 1 0 0.61 21-22 8 1 0 0.53 26-27 7 0 1 0.53 27-28 6 1 0 0.45 28-29 5 0 5 0.45 _____ Jutilise la version gratuite de SPAMfighter pour utilisateurs privés. Jusquà présent SPAMfighter a bloqué 568 courriels spam. Nous avons en ce moment 6 millions dutilisateurs de par le monde entier. Les utlisateurs payants nont pas ce message. Vous pouvez télécharger la version gratuite <http://www.spamfighter.com/lfr> _____ Jutilise la version gratuite de SPAMfighter pour utilisateurs privés. Jusquà présent SPAMfighter a bloqué 594 courriels spam. Nous avons en ce moment 6 millions dutilisateurs de par le monde entier. Les utlisateurs payants nont pas ce message. Vous pouvez télécharger la version gratuite <http://www.spamfighter.com/lfr>

1 0

testing of quick guide for SPC graphs
by epidata-list＠lists.umanitoba.ca 23 Mar '09

23 Mar '09

Dear All As part of telling of the most recent implementation of SPC Control graphs while attending a quality assurance conference/forum in Berlin (http://internationalforum.bmj.com/) last week I produced a "Statistical Process Control Graphs in 5 minutes" - a link is placed near the top on epidata.dk. Please comment whether you find the text appropritate. (e.g. whether it should be 10 rather than 5 minutes). See front page of EpiData or this direct link: http://www.epidata.dk/downloads/spc5min_epidata.pdf regards Jens Lauritsen EpiData Association

1 0

recoding numbers....
by epidata-list＠lists.umanitoba.ca 13 Mar '09

13 Mar '09

Hello all, Im facing a little problem which so far I could not find a work around. Any sugestions would be welcome. I have tow datasets which come from different sources... I intended to join them side-by-side with epidata entry to keep on going wiht the data update. I know that the second one would update some stuff in the firts one, at least for the majority of the observations. Both of the have a unique identifier which is a variable called num. However in the first set num is like this... 133.01.2007 244.01.2007 while in the second one num is like this... 2007.01.133 2007.01.244 For any software these would be different numbers but given that I know that 2007 is the yaer and 1 is january and 244 or 133 is the sequence within the month I know that the second dataset have num in the inverse order. Concatenating the datasets like this would be just as append because no observatiosn would match. I tought in recode the numbers but this would be same as digitizing all again. Nay sugestions? -- Abraço forte e que a força esteja com você, Pedro Emmanuel

1 1

v2.2 of EpiData Analysis - released
by epidata-list＠lists.umanitoba.ca 13 Mar '09

13 Mar '09

Version 2.2 has been released for public downloads today. You can get it from http://www.epidata.dk/downloads.php The file is provided as a setup exe file, but also as already installed, where all files are contained in a zip file. Since last public release in november 2008 quite many aspects have been improved or adapted. In particular the whole SPC (Statistical Process Control) graph system has been extended with more options and more efficient layout. Also a grid for creating SPC graphs from scratch has been implemented. - is available from the graph menu. Some set commands have changed, such that users can be forced to update any specified ini file that was created with the former version. An extension is the possibility for the user to create preference options for all commands, example: set option tables = "/t /r /c" would define that all tables done with the command "tables" will be added the options /t /r and /c Just today an error with stratified tables and small sample sizes was reported - the error is now fixed. The problem was that some strata would be combined incorrectly. Analysis done previously should be repeated with the new version and proper control. Please report any remaining problem. To see all changes go to the file: http://www.epidata.dk/analysisversioninfo.htm Commands are all as previously documented in the command reference, which will open when you press F1. Please notice that the release of this depends on donated funding, as documented on the page: http://www.epidata.dk/funding.htm Users should update to this version of analysis, since some of the bugs removed could affect results. The development continues with further online documentation and teaching materials as well as core module restructuring, such as announced on the list recently. regards Jens Lauritsen

1 1

SPSS v17 export problem
by epidata-list＠lists.umanitoba.ca 12 Mar '09

12 Mar '09

Hello. I have recently "upgraded" to SPSS v17. My exports from EpiData to SPSS format no longer work. The error messages I get in the SPSS syntax editor are: 14 DATA LIST Unrecognized text appears on the DATA LIST command where a variable name was expected. Text: TO 69 VARIABLE LABELS This command is not permitted before the beginning of file definition commands. 137 execute This command is not valid before a working file has been defined. Can someone suggest a resolution? Thank-you kindly. __________________________ Stephen Drew Health Data Analyst Region of Waterloo Public Health

1 1

naming of files (htm or html)
by epidata-list＠lists.umanitoba.ca 09 Mar '09

09 Mar '09

Dear all As part of the transition it was suggested to name all help files which has the type "html" to be released with extension html instead of htm - which is now used for EpiData Analysis. eg. commands.htm would be commands.html, start.htm would be start.html etc. If this poses problems to anyone let us know on the list. regards Jens Lauritsen EpiData Association

1 0

Autobackup-feature and current test of EpiData Analysis b160
by epidata-list＠lists.umanitoba.ca 24 Feb '09

24 Feb '09

The development is now at a crucial point in time getting to a certain "gate". Before the gate is opened EpiData Analysis will be released as v2.2 and some updates to EpiData Entry will also take place, e.g. hopefully fixing the reported "auto backup" problem reported. After the gate there will be a common core system for reading, writing and interpreting files for as well EpiData Entry, EpiData Analysis and EpiC. This will not be seen immediately by the end user, but will assure that all functions are doing the same regardless of which part of EpiData Software and errors only needs one place for bug fixes. The translation process will also be adapted and released in common for the three software parts, since after that all functions and commands in relation to reading and writing data files and chk files will have the same number code. The list and translators will be informed when possible on all of this later this quarter or beginning of next quarter. Meanwhile EpiData Analysis b161 has been placed for testing for download from the testing.php page of the main site. (by a mistake the first were called b160, but actually were b161) - if you report problems report the date of the build if this was b160). One new aspect is that one commented by J. Hockin in the Mantis issue number 335. (A template for SPC graphs in a menu like system): The technical principle is the following: A: An htm file is created by the user outside of analysis with a combination of EpiData Analysis commands and ordinary html coding. (here called spc.htm) B: This file is activated in the help window via an entry in the graph-spc toolbar menu/dialog. (the same as the command: helpview "spc.htm") C: When the user clicks on a given link in the file, certain commands are issued and information can be given in the output window - without closing the helpview file. J.Hockin argues in the Mantis issue that the strange window is difficult to close again and that it would be preferred to have the spc.htm file opened in the main results window. I experimented a bit with both solutions and it turns (to me) an easier approach to use the help window. This is the same approach used if one uses the menu "Windows->screen font". Users can develop other htm files for such menu systems and load them either into help window (helpview spc.htm) or result window (view spc.htm) and if you wish an easy return to such a file issue the "set START PAGE " command after which you can open your menu file with F8 or have the system open it at start of the exe file. For inspiration on creation of such files, see the startfont.htm and the spc.htm files. Apart from testing that new menu/command button principle it is important for us to get tests of the Statistical Process Control Graphs and the new features introduced. I will contact a number of users directly, but all are more than welcome to participate. The optimal is if users would : a1: Create simple datasets using the built in system accessible via the toolbar-graph-spc menu/dialog in such a way that the following can be documented: - freeze (where estimates are based on the first n data points - breaks (where the time series are split in sub-parts) - exclude (where individual data points are excluded from estimates) and the tests (t1-t5), which will show different tendencies/aberrations. The optimal is to send such datasets in a zip file to "info at epidata.dk" consisting of the data and the pgm file to run the tests, including expected results. Disregard the mail notion of the mail not being received - we will let these through the spam filter and contact you directly. regards Jens Lauritsen EpiData Association

1 0

auto backup!!
by epidata-list＠lists.umanitoba.ca 21 Feb '09

21 Feb '09

Batch files are fine and easy to automaticly do stuff. Improving the last coment, I would sugest to put a Warning to press F10 before exit in after file block or in a before file block in the check file! Although this may not work 100%, it may help educate and remind the training!!! -- Abraço forte e que a força esteja com você, Pedro Emmanuel

1 0

serious backup problem in epidata entry
by epidata-list＠lists.umanitoba.ca 21 Feb '09

21 Feb '09

Hello Friends I create many datasets at my hospital and instruct physicians on how to use it. I always add a backup section to my chk files. The problem is that after the user clicks 'save record to disk' and wants to exit, he/she usually clicks the top 'x' on the upper right side of the screen, which is the way most software works. Unfortunately, the backup is not done unless the user clicks the lower 'x'. There is no warning, and backups are not made. Am I doing something wrong? If not, is there a workaround. Failure to make backups has caused me many problems . I am using ver 3.1(120306). I would really appreciate any help with this problem. Thank you all Steve Blum Bronx-Lebanon Hospital Center Department of Medicine Bronx, NY

1 1

New Analysis - release candidate as v2.2
by epidata-list＠lists.umanitoba.ca 16 Feb '09

16 Feb '09

Since previous release of EpiData Analysis quite a few changes have been implemented. Most of these fall in two groups: a. Bug fixes of minor (spelling, parsing certain data sets etc) and major importance. Of major importance are a few in relation to aggregating data (aggregate and Stattables) and some if statements. Also some parts of other graphs, e.g. epicurves which would show incorrectly as connected rectangular bars when outcomes were separated by long periods. b. Enhancements and further implementation of SPC (Statistical Process Control Graphs) New chart types have been added (xbar-s, uchart , cchart), and further possibilities of doing proper analysis, e.g. freeze mean and limits based on first part of data, further warning tests (t4 and t5) and the setup of variables has been simplified. Additionaly a new "spc from scratch" menu has been added. You will find this on the graph toolbar submenu. c. Translation is now possible. But further documentation and contact to translators is needed. Users will find the translation tools on the testing page (in windows and linux versions). Please also test these. After test and removal of identified bugs this build will be released as v2.2. Known issues are: some datasets will give floating point error with xbar-s - please report on this. The internal validation routines used here are being updated to a new structure. Users are highly urged to test the new build and report problems here or in the feature and bug reporting database. The test versions are available from: http:\\www.epidata.dk\testing.php Other changes are expected soon in the documentation with a shift of core documentation (technical documentation) to a "wiki" based principle. After v2.2 a quite long period should be expected without new builds (apart from removal of identified serious bugs). The core parts are being rewritten to a common form for as well EpiData Entry, Epic and Analysis. And next test release will be based on this new core for all three parts. regards Jens Lauritsen EpiData Association

1 0