Skip to main content

Table 3 Stata specification for each model

From: Identification of risk factors for hospital admission using multiple-failure survival models: a toolkit for researchers

Model type

Commands to convert dataset as formatted in Extract 2 to a format suitable for model fitting

Description of commands to convert dataset

Commands to fit models

Description of commands to fit models

Time to first event Cox

stset End, failure(Failure==1) id(ID) origin(time Clinic) enter(time Clinic) scale(365.25) time0(Start)

origin(time Clinic) and enter(time Clinic) specify that the individual first becomes at risk after the clinic date

stcox varlist

varlist represents a list of predictor variables in the model

Andersen and Gill

stset End, failure(Failure==1) id(ID) origin(time Clinic) enter(time Clinic) scale(365.25) time0(Start) exit(time.)

scale(365.25) ensures that time is measured in years as opposed to days

stcox varlist, efron robust

efron specifies the use of the Efron method for analysing ties in the data

Prentice, Williams and Peterson Total Time

stset End, failure(Failure==1) id(ID) origin(time Clinic) enter(time Clinic) scale(365.25) time0(Start) exit(time.)

time0(Start) specifies that 'Start' is the beginning of the time interval spanned by each record

stcox varlist, efron robust strata(Risk_set)

robust specifies that robust variance estimation is used to account for within-subject correlation of failure times

exit(time.) ensures that an individual’s failures after the first are also recognised

 

strata(Risk_set) ensures that a stratified cox model is fitted with a different underlying risk for each ordered failure event