Could you please explain or provide some reference why we need to aggregate the data when the event is not recurrent or rare? I saw papers using time varying covariates when modeling non-recurrent event incidence rate, so the data is restructured to repeated measurements, with the outcome being 0 or 1 and offset is time between two measurements. Are they doing this in a wrong way?
l am also wondering if there is a minimum number of events requirement for multiple Poisson regression, when the event is really rare? For example, only 5 incidence events out of 500 individuals? Is it possible to do any model with covariates?
Could you please explain or provide some reference why we need to aggregate the data when the event is not recurrent or rare? I saw papers using time varying covariates when modeling non-recurrent event incidence rate, so the data is restructured to repeated measurements, with the outcome being 0 or 1 and offset is time between two measurements. Are they doing this in a wrong way?
l am also wondering if there is a minimum number of events requirement for multiple Poisson regression, when the event is really rare? For example, only 5 incidence events out of 500 individuals? Is it possible to do any model with covariates?