multistate survival

["Osman Al-Radi" <osman.al.radi@utoronto.ca>]

Dear s-news subscribers:

 

I am dealing with a small dataset of 40 subjects who can be in any one of 5 mutually exclusive states at any time, one of the states being dead and the others stages of a treatment or alternative treatment. The subjects are followed from birth to death or last follow-up, some are lost to follow-up (censored).    

 

how can one plot the prevalence in each state (y axis) versus time (x axis)?

 

how can one estimate the hazard of death over time in each state?

 

Thank you..

 

Osman