I have data where participants became at risk at birth but entered the study on Census day in 2011 and were followed for 9 years. Hence in my stset I am using the option ‘origin’ with the participant’s date of birth and ‘enter’ with the Census date.
I subsequently fit a Poisson model. I want to adjust for participants' age, but I am not sure if it is already done by ‘origin’, or I also need to include the participant’s age (at the time when they entered the study) as a predictor.
I have found this relevant bit in Royston & Lambert, Flexible Parametric Survival Analysis Using Stata: Beyond the Cox Model. Stata Press. (section 2.5.7):
"It has become increasingly popular to use age as the time scale in epidemiological cohort studies. This is an alternative (and often better) way of adjusting for age. With age as the time scale, we need the enter() and origin() options."
From this I would understand that age is already controlled for by 'origin' and 'enter', so I don't need to include age as predictor. But just want to make sure I am correct. Thanks in advance.
I subsequently fit a Poisson model. I want to adjust for participants' age, but I am not sure if it is already done by ‘origin’, or I also need to include the participant’s age (at the time when they entered the study) as a predictor.
I have found this relevant bit in Royston & Lambert, Flexible Parametric Survival Analysis Using Stata: Beyond the Cox Model. Stata Press. (section 2.5.7):
"It has become increasingly popular to use age as the time scale in epidemiological cohort studies. This is an alternative (and often better) way of adjusting for age. With age as the time scale, we need the enter() and origin() options."
From this I would understand that age is already controlled for by 'origin' and 'enter', so I don't need to include age as predictor. But just want to make sure I am correct. Thanks in advance.
