An embedded and charset-unspecified text was scrubbed... Name: not available URL: <https://stat.ethz.ch/pipermail/r-sig-epi/attachments/20080905/f6afc1fe/attachment.pl>
Paired by exposure
3 messages · Pedro Emmanuel Alvarenga Americano do Brasil, BXC (Bendix Carstensen)
An embedded and charset-unspecified text was scrubbed... Name: not available URL: <https://stat.ethz.ch/pipermail/r-sig-epi/attachments/20080905/4b93628d/attachment.pl>
This is not what we would call a paired design, because you are not pairing the OUTCOME, but the EXPOSURE. You are merely taking a sample of the exposed and a sample of the non-exposed, and making sure that they have the same distribution of age at entry / date of birth / date of entry --- which one(s)? Estimating mortality (morbidity?) in each of the two groups and the rate-ratio between them is just a straight-forward cohort analysis, either using a Poisson-model or a Cox-model. You will have to consider what time-scale you use for the description, though. And you must be sure that you are not using any of the follow-up information in your selection of persons to enter the study. Best regards, Bendix Carstensen ______________________________________________ Bendix Carstensen Senior Statistician Steno Diabetes Center Niels Steensens Vej 2-4 DK-2820 Gentofte Denmark +45 44 43 87 38 (direct) +45 30 75 87 38 (mobile) bxc at steno.dk http://www.biostat.ku.dk/~bxc
-----Original Message----- From: r-sig-epi-bounces at stat.math.ethz.ch [mailto:r-sig-epi-bounces at stat.math.ethz.ch] On Behalf Of Pedro Emmanuel Alvarenga Americano do Brasil Sent: 5. september 2008 16:01 To: r-sig-epi at stat.math.ethz.ch Subject: Re: [R-sig-Epi] Paired by exposure
Dear R friens, I will start a project this month wiht a partner which is
disigned as
a quasi-experiment. We are interrested in estimate risk
raios and rate
ratios, possibly survival analysis. Everything is registered at medical charts form the last 22 yaers. So everythig is
retorspective.
However, due to a efficiency matters, we dicided to select the non-exposed matching to the exposed. So for every paciente
that took
this paticular medicine, we intend ot match one that did
not take, and
start the follow-up in the same year and has no more than
10 years older than the exposed.
I have looked up for more recent things than those I have
in Klienbaum
and Rothman but it seems that the very good old friend
ajusted MH for
the pair is the recomended. Does anyone have other sugestions? Does any knows a package or functions to run paired risk or rate analysis?
Best regards
-- Abra?o forte e que a for?a esteja com voc?, Pedro Emmanuel
[[alternative HTML version deleted]]