Dear all,
I am analyzing lifestyle questionnaire data collected during follow-up (FUP) and comparing them with baseline assessments with time-to-event analysis from the age of FUP assessments until age of event/censoring. Several participants have missing values, ie. missing age at FUP. While a complete case analysis was performed, I would like to implement multiple imputation with the mice package in STATA. Age at follow-up was imputed with interval regression by setting the lower and upper boundaries as [age at entry (baseline), age at exit (event or censoring age)].
STATA would not want me to use the new imputed variable ‘newagefup’ in the stset. There is certainly something I am "missing" (again) about MI for outcome variables.
Any feedback and references would be highly appreciated.
With many thanks, Pietro
I am analyzing lifestyle questionnaire data collected during follow-up (FUP) and comparing them with baseline assessments with time-to-event analysis from the age of FUP assessments until age of event/censoring. Several participants have missing values, ie. missing age at FUP. While a complete case analysis was performed, I would like to implement multiple imputation with the mice package in STATA. Age at follow-up was imputed with interval regression by setting the lower and upper boundaries as [age at entry (baseline), age at exit (event or censoring age)].
STATA would not want me to use the new imputed variable ‘newagefup’ in the stset. There is certainly something I am "missing" (again) about MI for outcome variables.
Any feedback and references would be highly appreciated.
With many thanks, Pietro
