Hi everyone! What is the minimum sample size for odds ratio test in each group? What are the rules? For instance, my colleague (doctor) used odds ratios for the analysis. The one group included 10 patients with 50% of the outcome, another one - 295 patients with 12%. He concluded that the results were significant. Because the sample size in one group was too small, I am not sure if it is correct or not.
Here you’ll see how to do a sample size calculation to be able to estimate an OR within a decent multiplicative margin of error. The sample size needed is pretty shocking.
Thank you so much! It is really useful book! Am I right to understand the method is described in chapter 6.9.2?
If that’s so, we have 2 groups 10 and 295 patients. In the first group, the outcome is 50% (p1=0.5), in the second one is 12% (p2=0.12). Power = 0.8. The r function is round ( bsamsize (0.5 , 0.12 , fraction =? , power = .8 )). What is the fraction in my case?
No, section 6.8. That’s where the multiplicative margin of error is highlighted.
Ok. But it is still unclear how can I get a sample size for my case. I have found the next decision:
library(epiR)
epi.sscc(OR = 2.0, p1 = NA, p0 = 0.12, n = NA, power = 0.9, r = 1,
phi.coef = 0, design = 1, sided.test = 2, nfractional = FALSE,
conf.level = 0.95, method = “unmatched”, fleiss = FALSE)