We accounted for that by only including cases that either died or recovered in our estimate of the hospital fatality risk.