Survival analysis with proxy variable for time?

[davidkurland]

Hi all - I’ve conducted a time-to-event analysis (accelerated failure time model) using an unconventional approach.

We are looking at the effects of a treatment on the probability of getting a blood transfusion during surgery.

Instead of time however, I've used volume of blood loss during the surgery (and censoring for whether or not someone received a transfusion). Long story short, from everything I’ve read, using this in place of a time variable actually seems to satisfy all of the assumptions for survival analyses but I have been unable to find a reference where someone has used a non-time variable before.

Does anyone have any thoughts about this? Would greatly value advice.