AK, the 20% efficacy doesn't correlate with late 2016 interim readout because lower efficacy leads to earlier interim. HD had a brilliant explanation and R-I interim calculation few months back. I hope he can repeat that exercise again here.
We know the following about R-I,
1) the 1612th event will occur end of 2017 2) the 967th event will occur sometimes in 2016 3) the lower the efficacy, the earlier the R-I interim