[EpiData-list] Condensing several Y/N results into one table

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Fri Oct 10 08:56:34 CDT 2014


Two commands in EpiData Analysis are created to enhance condenced tables:

E.g. to see the proportion of ill persons among those taking orangejuice, water whisky,
ciplot ILL bread asauce stuff cslaw
or without the graph:
ciplot ILL bread asauce stuff cslaw /ng

You will get a table for each with proper Confidence intervals

To see the odds ratio for being ill among several use instead:
tab ILL bread asauce stuff cslaw /ct /o
where option /ct indicates compact table and /o Odds ratio

or to see the proportion ill among those exposed:
tab ILL bread asauce stuff cslaw /ar /ct

this will show an attack rate, which is the proportion.

I think we need another compact table for your specific use, but please explore these possibilities.

I could be be possible to programme an output like you ask for using "user written commands" and result variables which comes out after commands, I will elaborate on this later.

regards

Jens Lauritsen
EpiData Associations

Den 10. okt. 2014, epidata-list at lists.umanitoba.
> Dear list,
> 
> I try to condense the frequency of several yes/no variables into one table:
> 
> In a patient file I register 12 specific procedures, where each patient get
> a variable number of "yes" i.e.
> 
> Intravenous
> antibiotics
> N %
> Y 126 27.8
> . 328 72.2
> Total 454 100.0
> 
> Pain relief
> N %
> Y 113 24.9
> . 341 75.1
> Total 454 100.0
> 
> etc
> 
> I want to transform the result into:
> 
> Treatment summary
> N %
> Intravenous antibiotics 126 27.8
> Pain relief 113 24.9
> etc
> Total 454 100.0
> 
> In the old Epi Info v6 this was relatively simple via the REPORT command,
> where a template .RPT file could pick specific values from tablecells
> across several variables.
> 
> I can copy and paste values from individual tables into a spreadsheet, but
> it would obviosly be more user-friendly to a routine user to be able to
> present the result in EpiData.directly.
> 
> Any suggestions?
> 
> Best regards,
> Vegard Hogli
> CMO, Skien, Norway
> _______________________________________________
> EpiData-list mailing list
> EpiData-list at lists.umanitoba.ca
> <http://lists.umanitoba.ca/mailman/listinfo/epidata-list>
> 


More information about the EpiData-list mailing list