zico wrote:
Death rate = 4,034/73,000 = 5.5% in 2 months (equating to 33% per year).
Again, I'm trying not to be picky (honestly!) but you really can't do that statistically, unless you are sure you have a linearly driven probability distribution over time, and frankly that's unlikely. You can't just multiply by 6 to make that an annual rate.
What is probably more likely is that for those surviving 28 days (the measure) there is a higher risk of dying on day 29, than day 365. In the same way that we have mortally tables that show a 90 year old more likely to die "tomorrow" than a 30 year old, you will have a mortality effect where you are more likely to die of a relapse, or complication from the possible effects of treatment/surgery closer to the date of that treatment/surgery.
Statistics, particularly in "real life" are complex.