turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- Funnel Plots

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

01-22-2013 03:33 PM

I'm trying to plot a funnel plot to identify performance outliers from a sample of hospitals. I tried to follow Rick's example here: Funnel plots: An alternative to ranking - The DO Loop

But i had some problems understanding the PROC IML part. In my case, I have already have the population mean, standard deviation, my sample hospitals all had their mean and s.e.

Is there another approach to simulate the confidence limits?

Thanks!

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

01-22-2013 05:55 PM

Hopefully, Rick will respond himself. Otherwise, take a look at: http://www.sas.com/offices/NA/canada/downloads/presentations/Edmonton2010/Funnel.pdf

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

01-23-2013 09:01 AM

I think the blog post is fairly complete. I don't have anything to add. Assuming that the sample size for each hospital is large, you don't need to "simulate the confidence limits." Just use the normal approximation to the binomial.

For other posts on the funnel plot, see the "Trackback" section at the bottom of the blog post. In particular, this post

Funnel plots for proportions - The DO Loop

seems relevant, although it uses an adjustment for the binomial limits, rather than the simpler normal limits.