Frank wrote:I'm currently reading through the full text of the "liberation procedure" paper and have a problem understanding the math at one point.
On page 75 (figure 3) it is said to show that the relapse rate after intervention decreased more than 4 fold (exactly 4,4). The relapse rates were 41% in the year pre-treatment and 25% in the year post treatment.
I would state that the relative change in relapse rate pre->post is 39.02%:
41 = 100%; The change is 41 - 25 = 16; 16 = x%
x = (100% / 41) * 16 = 39.02
What calculation gets to the 4.4 fold decrease in relapse rate?
I've tried figuring these numbers many different ways, and I still don't understand how the authors got these results. First, I'd like to know how they are defining the rate of relapse (.41 before treatment, .25 after treatment). Most drug studies for MS use mean annualized relapses as a primary measure. Are they totaling the number of relapses for the year, and dividing this by the number of patients?
Then, how are they getting an odds ratio (OR), from this? Are they using these relapse rates as the probability of having an attack? If this is the case, then I calculate the odds ratio as:
Odds of having attack before procedure = .41/.59
Odds after procedure = .25/.75
So the odds ratio = (.41/.59) / (.25/.75) = 2.08
which obviously is different from the odds ratio given in the paper.
The only thing I can figure is that they are using some other correction, possibly to account for the small sample size.