topic Comparing rates on a non-independent sample in Statistical Procedures

Comparing rates on a non-independent sample

SW2 — Wed, 23 Dec 2015 18:53:47 GMT

Hi, all,

I'm checking the increase rates of outpatient visits for certain diseases on the same sample. As you see,

The number of outpatient visits all increased after a time point. Among them, Disease1 had the highest increase rates(50%)

Is there any way to do a statistical analysis to compare the increase rates of ‘disease1’ 50% to that of 'total diseases' 27%? The sample is non-independent. I appreciate for any suggestion.

	before	after	Increase rates
Number of visits on Disease1	100	150	50%
Number of visits on Disease2	100	120	20%
Number of visits on Disease3	100	110	10%
All diseases	300	380	27%

Re: Comparing rates on a non-independent sample

Reeza — Wed, 23 Dec 2015 21:02:09 GMT

Do you have individual level information or only aggregate information?

Re: Comparing rates on a non-independent sample

SW2 — Wed, 23 Dec 2015 21:33:45 GMT

Yes, we have individual information.

Re: Comparing rates on a non-independent sample

Ksharp — Thu, 24 Dec 2015 01:10:23 GMT

Poisson Regression Model ?

http://support.sas.com/kb/24/188.html

Re: Comparing rates on a non-independent sample

PGStats — Thu, 24 Dec 2015 03:15:18 GMT

If you have the licence to SAS/ETS, you should consider proc countreg to fit your data. Look at the following example:

http://support.sas.com/documentation/cdl/en/etsug/67525/HTML/default/viewer.htm#etsug_countreg_examples01.htm