Re: st: calculate the exact Poisson confidence intervals in stata

[Steve Samuels <sjsamuels@gmail.com>]

You would have to do one population (pctcode06 HRG) at a time and
supply the expected numbers.   -istdize- save the lower and upper
bounds for the observed numbers in r(lb) and r(ub).  so r(lb)/Expected
 and r(ub)/Expected are the CIs you want. A Google search will find
the formula for confidence intervals for the Poisson in terms of
quantiles of the chi square distribution, and you could implement this
using Stata's probability functions.

 Why do you think you need 500 observed before you can trust the
large-sample approximations? I didn't find that when I  looked at test
power years ago.
 ( SJ Samuels, JJ Beaumont and NE Breslow  Power and detectable risk
of seven tests for standardized mortality ratios. American Journal of
Epidemiology 1991;133:1191-1197.)

Steve

On Mon, Jul 26, 2010 at 5:38 PM, helen bian <helenbian06@gmail.com> wrote:
> I have generated two variables, observed and expected, and the number of
> observed event is less than 500. I would like to calculate the confidence
> interval for the ratios (=observed/expected), i.e. the indirectly
> standardised ratios. Does anyone know how to calculate the exact Poisson
> confidence intervals in stata?
>
> Alternatively, can anyone tell me how to save the results from istdize into
> a datafile? I have tried to use the following code to claculate the
> confidence intervals for a dataset contains variables: vol, pop, sex,
> agegroup, pctcode and HRG, but I could not manage to save the result into a
> datafile.
>
> istdize vol pop sex agegroup using file.dta, by(pctcode06 HRG) pop(vol
> pop)
>
> Many thanks,
> helen
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>



-- 
Steven Samuels
sjsamuels@gmail.com
18 Cantine's Island
Saugerties NY 12477
USA
Voice: 845-246-0774
Fax:    206-202-4783

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/