04-08-2014 12:17 PM
Could someone tell me if another procedure than SDTRATE can calculate age and sex standardized rates (direct method) with confidence intervals. I do not have access to SAS/STAT.
I have a reference population large number of regions that I want to compare with each other.
04-08-2014 12:28 PM
You can do it in a standard data step, just need to know the formula's
Here's a good paper that walks through all the calculations.
04-08-2014 03:28 PM
Thank you for your answer.
But because I deal with very small counts, approximation formulas give me negative numbers for lower bounds. I need exact calculation by Poisson or gamma distributions...
04-08-2014 03:55 PM
There's the probit and gaminv functions to calculate those...again goes back to knowing the formula's. There are macro's for this out there, search lexjansen.com or check the docs for the exact formula's