Survival Analysis

[Cagoguy]

Just a quick question on running a survival analysis.

Regarding the Time to Event variable, how is it coded?

For example, if I'm measuring time to death and I measure for up to one year, how do I code people who last for longer than a year. Can a zero be coded? If so, what does it mean and how will a stats program, like SPSS, register it?

Thank you..

 

[WeeG]

if I understand you correctly, you mean if the study ends after a year and someone was still alive at the end of the study ?
if this is what you meant, you code this case as 0 (censored). This is how computer software will treat it.
(in general you code death as 1 and censored as 0)

 

[Cagoguy]

if I understand you correctly, you mean if the study ends after a year and someone was still alive at the end of the study ?
if this is what you meant, you code this case as 0 (censored). This is how computer software will treat it.
(in general you code death as 1 and censored as 0)

I'm sorry maybe I wasn't clear. You are describing the event (yes or no) variable. I am describing the "Time to Event" variable that also has to be entered. I.e., if they die on day 30, the variable is 30.

 

[fed1]

I'm sorry maybe I wasn't clear. You are describing the event (yes or no) variable. I am describing the "Time to Event" variable that also has to be entered. I.e., if they die on day 30, the variable is 30.

You should enter the last time you saw them alive here.

 

[Cagoguy]

OK thanks, so I'm working with an old data set and most of the data makes sense then, but some people are entered as zero. Does that mean they dropped out the first day?