I worked it out using a binomial distribution with a 0.4 chance of surviving.
This comes to (8!/(7! * 1!)) * 0.4 ^ 7 * 0.6 ^ 1
I worked this out to 0.00786 or ~~ 1 percent.
I thought 5 percent was the threshold for statistically significant. If you set the survival chance to the lower 20 percent, its even more significant.
I'm not an expert though, and there may be other problems with the study, but I don't think what you said is accurate.
If normal treatment helped to that degree, would fatality rates be as high as they are? I don't know if the 60-80 percent fatality rate takes into account the inadequacy of the health systems in countries with common Ebola outbreaks.
But yes, I already said there were probably other problems too.
5 comments
[ 3.9 ms ] story [ 26.4 ms ] thread8 patients do not make a statistically significant sample...
This comes to (8!/(7! * 1!)) * 0.4 ^ 7 * 0.6 ^ 1 I worked this out to 0.00786 or ~~ 1 percent.
I thought 5 percent was the threshold for statistically significant. If you set the survival chance to the lower 20 percent, its even more significant.
I'm not an expert though, and there may be other problems with the study, but I don't think what you said is accurate.
The transfusion they received is not the only variable at play.
But yes, I already said there were probably other problems too.
Then again there are quite a few convalescent health workers now.