Hi experts,

I know that in STATA we can use the logit command to fit a discrete-time logistic hazards model (logit event timevar indepvars, where event is a binary variable with 0 meaning surviving at a given time and 1 meaning 'died' at a given time), given that our data is in the long format.
My question is: can I use local regression commands such as lpoly/lowess/npregress to fit a hazards model? For example, given a long-format data, can I run:

lpoly event timevar indepvars, logit

or

npregress event timevar indepvars

and then use margin to calculate the probability of experiencing the event at each time point of the timevar?