Small Prevalence Rates, Complex Survey, Age Adjusted

[Bob Gerzoff]

I have a complex survey (BRFSS) from which I need to calculate some very small prevalence rates (<5%) and also age-adjust to three age categories from the 2000 Census. Because of the small rates, I need to use "exact", Clopper-Pearson confidence intervals. Is this something that Stata can do ? Other packages will handle two of the three conditions, but not all three at once.
Thanks in advance for your help and direction.
Bob

[Mike Lacy]

This might be a long shot, but what about -bootstrap-? It allow both a cluster and strata option, so if you knew the sampling structure and so forth, and could write a small program to calculate the age-adjusted prevalence, that should in principle be possible.

[Bob Gerzoff]

I would hate to have to go down that route, since the program will eventually wind up being the responsibility of programmers and that might a little more involved than they are used to. It would probably work, however, and is something to think about. Thanks.

[David Jacquier]

Hello,
I join the topic just because my problem seems to fit with the title and I did not wish to start a new topic, even if the previous question-answer is way out of my league...

I have collected survey data about sleep from the general sample (TD) and from a population of interest (PI). There are several significant demographic differences between both group (demo1, demo2, demo3), and I used the distribution of the sleep quality to categorize (dichotomize) my PI sleep data. But my colleages ask me to adjuste the found prevalence in PI for the various demographic factors.
How do I run this? Is it really meaningfull to rely on the GP data to categorize the PI, and then adjust again the PI data for demographic factors?
I have seen in the forum the svy command, but I must confess that I do not fully comprehend.
Thank you a lot, best regards!