Re: Is constructing a contingency table from odds ratio's (including 95%CI) possible

On Jul 27, 10:12 am, Aniko <aniko123...@xxxxxxxxx> wrote:
On Jul 27, 9:50 am, nyika <NDKr...@xxxxxxxxx> wrote:

For a meta analyses i was able to subtract the following information
from an article:
1) the Odds ratio on poor outcome, with 95% confidence intervals
2) the number of patients with treatment and the number of patients
without treatment. (so, also the total number of patients).
I was wondering if anyone knows if it is possible to reconstruct a
contingency table with this information. (is there any software that
can do this? )

regards
ND Kruyt

It might be possible if the sample sizes are large enough. Do you know
the method for calculating the confidence interval? Assuming that an
asymptotic normal approximation was used (or that the sample size is
large enough that all methods give approx the same answer), the CI
should be symmetric in log-scale. For a table with elements a,b,c,d,
SE(log(OR))=1/a+1/b+1/c+1/d, and the CI would be exp(log(OR)
+-1.96*SE(log(OR))). So you would have four unknowns and four
equations: a+b, c+d are the given sample sizes, ad/bc is the odds
ratio and the sum of inverses can be obtained from the CI. This should
be solvable (perhaps even analytically, I did not try). I am not aware
of software that would solve your problem explicitly.

Aniko

Oops, I lost a square root. 1/a+1/b+1/c+1/d is the variance of
log(OR). But the basic idea stands.
Aniko

.

Relevant Pages