[R] meta-analysis, outcome = OR associated with a continuous independent variable

[Viechtbauer Wolfgang (STAT)]

I do not see any major difficulties with this case either. Suppose you have OR = 1.5 (with 95% CI: 1.19 to 1.90) indicating that the odds of a particular outcome (e.g., disease) is 1.5 times greater when the (continuous) exposure variable increases by one unit. Then you can back-calculate the SE of log(OR) = .41 with

sei = (ln(ci.ub) - ln(ci.lb)) / (2*1.96), 

which in this case is approximately 0.12. The sampling variance of log(OR) is then vi = sei^2.

Now you have everything for the meta-analysis, using any of the packages mentioned.

What Thomas already mentioned is that the 'one unit increase' must mean the same thing in each study. Therefore, if the exposure variable is measured in months in one study and in years in another study, then the odds ratios are obviously not directly comparable. If the units are just multiples of each other, then you can easily calculate what the OR would be when putting the exposure variable on the same scale. For example, an OR of 1.5 for a one month increase in exposure is the same as an OR of 1.5^12 = 129.75 for a one year increase in exposure.

Best,

Wolfgang

--   
Wolfgang Viechtbauer, Ph.D., Statistician   
Department of Psychiatry and Psychology   
School for Mental Health and Neuroscience   
Faculty of Health, Medicine, and Life Sciences   
Maastricht University, P.O. Box 616 (VIJV1)   
6200 MD Maastricht, The Netherlands   
+31 (43) 388-4170 | http://www.wvbauer.com   


> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
> On Behalf Of Thomas Lumley
> Sent: Wednesday, April 04, 2012 23:42
> To: Marie-Pierre Sylvestre
> Cc: r-help at r-project.org
> Subject: Re: [R] meta-analysis, outcome = OR associated with a continuous
> independent variable
> 
> On Thu, Apr 5, 2012 at 8:24 AM, Marie-Pierre Sylvestre
> <mp.sylvestre at gmail.com> wrote:
> > Hello everyone,
> > I want to do a meta-analysis of case-control studies on which an OR
> > was computed based on a continuous exposure. I have found several
> > several packages (metafor, rmeta, meta) but unless I misunderstood
> > their main functions,  it seems to me that they focus on two-group
> > comparisons (binary independent variable), and do not have the option
> > of using a continuous independent variable.
> 
> 
> There's no problem in using continuous exposures in meta.summaries() in
> the rmeta package.  For each study, compute your log odds ratio and its
> standard error, and feed them in.
> 
> You just need to make sure that the odds ratio is in the same units in
> each study, of course.
> 
>    -thomas
> 
> --
> Thomas Lumley
> Professor of Biostatistics
> University of Auckland
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-
> guide.html
> and provide commented, minimal, self-contained, reproducible code.