Re: STATS [was] what constitutes enough exercise?

[Guest guest]

In a message dated 10/4/04 10:36:28 PM Eastern Daylight Time, bpinfo@... writes:

also, as to:

>"had multivariate relative risks of 0.90, 0.81, 0.80, and 0.81, respectively "

In the study I cited, "Walking Compared with Vigorous Exercise...", the effect shown under multivariate analysis was similar (1.00, 0.89, 0.81, 0.78, and 0.72), and also is *less* than the relative risks of 1.00, 0.73, 0.69, 0.68, and 0.47

Why does multivariate analysis give different figures than the relative risk analysis?

--

Ken

well, Ken

after a bit of reading, it seems to go something like this:

- multivariate analysis attempts to correct for confounding factors

- multivariate analysis gives an odds ratio

- risk ratio and odds ratio are *not* the same thing, because:

The risk describes the number of participants having the event in a group divided by the total number of participants

The odds describe the number of participants having the event divided by the number of participants not having the event

http://www.cochrane-net.org/openlearning/HTML/mod11-6.htm

- risk and odds will be close if the outcome is < .1 But, they will diverge more and more as the ratios approach 1.0

(see graph: http://jama.ama-assn.org/cgi/reprint/280/19/1690.pdf)

- conclusion: when comparing the size of an effect, it's important to know if you are dealing with risk or odds ratio(s). If you are dealing with risk ratio in one study, and multivariate odds ratio in another, it is not possible to talk about the relative size of the effect per se across both studies.

-

Ken