Study question: build a model that predicts if a patient would benefit from treatment A vs treatment B.
Outcome: numerical survey score at 24 months
Other variables: demographics, potentially imaging.
Issue: I want to risk stratify and use something like cox regression to create nomogram (so under this score, recommend Treatment A, above cutoff, recommend Treatment B; however, our outcome measure is improvement (i.e we don’t have an adverse outcome for survival analysis) so I guess there is no “risk” per se. Any way around this? Or should I turn to SVM/random forest (but then this wouldn’t allow me to risk stratify)?
