Dear allstat members,
I am trying to analyse quality of life data (EORTC) by a repeated measures
analysis allowing for non-ignorable missing data from dropouts, missing
forms and deaths. I have obtained a number of papers which suggest methods
of analysis (e.g., Schlucter, 1992; Diggle & Kenward, 1994; Fairclough,
1998) but am unsure (understatement!) of how to program these using SAS.
Proc mixed deals with data missing at random, but, as far as I know, cannot
deal with non-random missing items. I think I need to incorporate survival
data including censoring into the model.
If anyone has performed such an analysis using SAS or knows how to, I'd be
very grateful for their assistance.
thanks
Pete
email: [log in to unmask]
|