Section 11.1.3
Subsection: Impact of hospital-acquired pneumonia (HAP) on length of stay in intensive care unit
| Cox model with pneumonia as a time-dependent covariate: |
- cox.hap.tdp <- coxph(Surv(start, stop, status) ~ pneu, icu.pneu) summary(cox.hap.tdp)
|
| With some further covariates: |
- cox.hap.tdp <- coxph(Surv(start, stop, status) ~ pneu + age + sex, icu.pneu) summary(cox.hap.tdp)
|
Section 11.2.3
Subsection: Impact of hospital-acquired pneumonia (HAP) on intensive care unit mortality
| Creation of a new variable which encodes both observed competing event status and censoring status: |
- icu.pneu$outcome <- with(icu.pneu, status * event)
|
| Proportional cause-specific hazards models including HAP as the time-dependent variable: |
- cox.hap.death <- coxph(Surv(start, stop, outcome == 2) ~ pneu, icu.pneu) cox.hap.disch <- coxph(Surv(start, stop, outcome == 3) ~ pneu, icu.pneu) summary(cox.hap.death) summary(cox.hap.disch)
|
| Proportional subdistribution hazards model with HAP as a time-dependent covariate using the kmi package. |
- ## Computation of the imputed data sets: set.seed(4284) imp.dat <- kmi(Surv(start, stop, outcome != 0) ~ 1, data = icu.pneu, etype = outcome, id = id, failcode = 2, nimp = 10) ## Cox models on the imputed data sets: kmi.sh.hap <- cox.kmi(Surv(start, stop, outcome == 2) ~ pneu, imp.dat) summary(kmi.sh.hap)
|
