I am using RMS package of R to validate cox regression model with bootstrap. Please see the R code below. I am predicting 1, 2, 3, 4, 5 years survival. Is the optimism corrected discrimination index Dxy a single index for the model, regardless of which year I am predicting, or does it have a different value for predicting each year survival? When “i” is specified to be 1, 2, 3, 4, 5 (years), the optimism corrected Dxy changes. Is this because of the change in the year? Or is it just due to the different random samples generated in each run?

f = cph(s1 ~ 
          GRADE_f_imp + 
          METAST_f_imp +
          largests_2b_imp +
          lymphnod_f_imp +
          LNPOS_f_imp +
        data=d_OS, x=T, y=T, surv=T, time.inc=i*12)

validate (f, B=200, u=i*12)

When you do not have stratification in the Cox PH model, $D_{xy}$ is simple and treats time as a continuous variable, so you do not get separate $D_{xy}$ for different time cutoffs, unlike for calibration curves. $D_{xy}$ is the Somers' rank correlation between predicted log relative hazard and observed survival time.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.