Death rate for critical cases is still extremely high, and going from 190 to 215 or so there were 5 more deaths, which is totally in line with the 23.3% overall mortality observed to date...

So, the p-value will very much improve with a 33% higher sample size...