Hello everybody,

I hope that there is anybody around this world, who can help me. I study viral infections after bone marrow transplantation. The most important risk faktor for infections is the occurrence of a graft-versus-host-disease (common complication with an onset about 20-60 days after Transplantation, defined as binomial variable (existing/not existing)). Other risk factors are age, underlying diseases and so on. Now, I want to carry out a cox regression analysis. But how can I define graft-versus-host-disease as a time-dependent covariate? If it occurs it contributes to the risk of developing viral infections from the day of onset on, which is different for each patient. I don't know how to program SPSS for this simple (?) job. Does anybody have an idea or is familiar with this problem?

Thanks a lot!

Johannes Schetelig