waldtest for cox proportional hazards

[vincent vinh-hung <conrvhgv@az.vub.ac.be>]

I would like to test the joint significance of two variables
(V2, V3) in a Cox model with three variables V1, V2, V3.
Is it ok to do:

fit <- coxph(Surv(time,event) ~ V1, data=m)
update(fit, ~ . + V2 + V3, init=c(fit$coef[1],0,0),data=m)

or with Therneau's waldtest, such as:

fit3 <- coxph(Surv(time,event) ~ V1 + V2 + V3, data=m)
waldtest(fit3,rbind(c(0,1,1),c(0,1,0),c(0,0,1)))

I'm uncertain of usage,
thanks in advance for any comment,

Vincent Vinh-Hung
Oncologisch Centrum, AZ-VUB
101 Laarbeeklaan
1090 Jette
Belgium
tel +32 2 477 6041
fax +32 2 477 6212
mobile +32 477 788 900
conrvhgv@az.vub.ac.be