Dear All,
the patient risk score could be calculated by the weighted linear combination of
variables with estimated coefficients (linear.predictors returned by coxph). if
the linear predictor is greater than 0, the risk score is high and low in the
opposite case. however, I don't find the coherence for my data according to
this definition
Did I take some thing wrong in the definition?
the linear predictors of the first 5 patients
cox.dt$linear.predictors[1:5]
[1] 0.5374888 -4.5403035 1.7174977 0.9651050 0.3022809
the survival time of the first 5 patients > dtSurv[1:5]
[1] 54 52 30 46 8
survival of 54, 46 and 8 months have positive and between 0 and 1 score while 52
months has a negative score. which is contradictory
thanks in advance
carol
---------------------------------
Need Mail bonding?
[[alternative HTML version deleted]]