Answer:
I guess you want to find the relative frequencies:
In total, we have 396,651 children
in sample 1, we have 198,747 children, and of those, only 37 developed the disease.
This means that the vaccine is not effective in:
p = 37/198,747 = 0.00019
For the placebo, we have a sample of 197,904 children in total, and 111 of them developed the disease, then we have a probability of the placebo not being effective is:
p = 111/197,904 = 0.00056
So we can see that the difference is significant, meaning that the placebo is not effective in more than twice the cases where the vaccine is not effective.