Re: [EpiData-list] Graphing results of outcome analysis showing CIs
Thanks Jamie
I tried this and it worked well. It was quite easy to copy the Attack rate table into Excel. I just highlighted the data in the htm file and pasted it into a black worksheet. Then I 1. used data/text to columns to separate out the confidence interval colum into two columns (low and high) 2. Used edit replace to take out the brackets 4. Change the order of the columns to LowCI HighCI RR 3. Selected the data to make a chart 4. Chose stock chart 5. I also chose the option to include the data table within the chart
The graph does not look as good as CIplot but does look pretty good.
many thanks
Christine
Christine Roseveare Analyst Regional Public Health 570 9194 027 495 9671
epidata-list@lists.umanitoba.ca 6/10/2009 2:28 a.m. >>>
Hi Christine,
CIPLOT is showing proportions, not RR. I think you will have to create your own chart in another package. You can import the attack rate table into excel and with a bit of manipulation, get the columns in the right order for a "stock chart" - stock chart uses the order Close, High, Low, where Close will be RR, High will be UCL and Low will be LCL.
However, you will have to do some formatting to do to get the chart looking as good as CIPLOT.
Jamie Hockin
Christine wrote
Does anyone have a tip for the best and easiest way to graph the results of the attack rate table for an outbreak analysis including confidence intervals? I have run the outbreak analysis and generated the attack rate table (TAB CASE 1 OYSTERS PRAWNS CHICKEN etc OA/ex), This is great. BUT...
Now I have been asked to show all the exposures with relative risks higher than 2 and CIs that don't include 1 (of which there are about 10) on a single graph. The graph must show relative risk, lower confidence interval and higher confidence interval for each exposure. I have looked at the CI graph in Epidata Analysis but a) I only want the total and the CIPLOT graph shows Yes and No as well as total b) I want to show more than one exposure on the same graph and I don't think I can do this with the CI.
participants (1)
-
epidata-list@lists.umanitoba.ca