sample size calculation, planned multivariate regression

#1
The primary outcome is seen in the literature as about 6-10% incidence. I am looking at time of an intervention as it relates to outcome development. What makes it more complex is that I am also looking at multiple other variables to see if there will be association with the outcome as I hypothesise that the intervention being assessed doesn't actually impact the outcome incidence at all. I will be doing a multivariate regression analysis for this. Any assistance appreciated.
I have looked at a couple of calculations. One that follows a basic population estimated sample (I have used 950 as this is surgical cases in a year) and it gives me a sample of about 275 when keeping standard error rates powered at 80%.
I have found also a multiple regression calculation that is N>104 +k which with 10 proposed variables gives me a sample size of just 114. Is there one that I would be more correct in using?
 
Last edited: