Neither, both would be skewed. That's because death happens after infection and recovery takes longer than death, and the number of cases is growing. What you want is dead/recovered N (28 for example) days after infection, which you will have to find elsewhere.