Confidence interval for person time incidence rate

#1
Dear everyone,

I would like to estimate a 95%CI of person time incidence rate (as 45 death per 10,000 person year)
Anyone knows a way to estimate it ?
Same question for Kaplan Meier estimate..
Thanks a lot for your help!
 

ledzep

Point Mass at Zero
#2
See page 84 of the book. There is an equation for CI calculation.
https://books.google.co.uk/books?id...terval for person time incidence rate&f=false

Regarding K-M, I am afraid you cant generate a CI just based on number. It takes account of the number of events, number at risk and uses Greenwood's method (most software by default), which uses the sum of all the estimates in preceding time interval to compute variance and CI. If you've found any other method, I would be interested to hear.

HTH