[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: weird results...

From   Christopher Intemann <>
Subject   st: weird results...
Date   Sat, 6 Dec 2008 18:20:35 +0100

I'm trying to calculate the IRR for multiple events in a longitudinal study dataset.
I therefore stset the data:

stset Event_date if disease_no==0 & cond==1, id(ID) failure(disease_yes==1) origin(time birthdate) enter(time birthdate) exit(Visit==50)
replace _t0=_t0+t_ex

Then I'm splitting the dataset to exclude non important time periods:

stsplit x_split, at(60 761) trim

Now, when I calculate the rate, I receive something like:

strate gene if x_split==60,per(365.25)

GT	Rate	
CC	0.64	
CT	0.53	
TT	0.55	

-> the rate of disesase-attacs is lower for the CT/TT-genotype, right?

Now, when I calculate the inicidence rate ratio (IRR) on the same st- set data, I get:

xi: poisson _events i.gene if risktime>0&x_split==60, exposure(risktime) irr

CC	1
CT	1.11 (1.03 - 1.19)	0.005
TT	1.14 (1.04 - 1.24)	0.005

According to these results, the IRR is HIGHER for the CT/TT-genotype.
This is weird. I guess there is something wrong.
Any help would be appreciated!

Attachment: smime.p7s
Description: S/MIME cryptographic signature

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index