 # p-value of cox regression model in survival package “coxph()”

Dear all,
I’m using the package “Survival” to perform Cox regression analysis. Until now, I’ve gotten the results successfully. But I still have a question of the results:
As is shown in the picture below, the overall P-value (0.1122) of Lymnodes_status is different from P-value of “Lymphnodes_status=positive” (0.101). what's the reason of this difference?
In my opnion, the p-value of "Lymphnodes_status=positive" is calculated by wald test while overall p-value is given by likelihood test. which one should i choose?
Moreover, for polytomous variable, such as "surgery=c("no", "no-BSO","BSO")",it seems that p-value of each level can just be given by wald test. In this case, i can only accept the overall p-value of wald test? or are there many methods to perform likelihood test for each level?
I'd appreciate it if ypou can give me any suggestion.
my code:

osfsingle<-coxph(formula = Surv(time, os==1)~Lymphnodes_status,data = ess1 ,x=T,y=T)
osfsingle
Call:
coxph(formula = Surv(time, os == 1) ~ Lymphnodes_status, data = ess1,
x = T, y = T)

``````                 coef  exp(coef) se(coef)  z   p
``````

Lymphnodes_statusPositive 0.3690 1.4463 0.1915 1.927 0.054

Likelihood ratio test=3.46 on 1 df, p=0.06296
n= 1172, number of events= 164

``````summary(osfsingle)
``````

Call:
coxph(formula = Surv(time, os == 1) ~ Lymphnodes_status, data = ess1,
x = T, y = T)

n= 1172, number of events= 164

``````                   coef exp(coef) se(coef) z Pr(>|z|)
``````

Lymphnodes_statusPositive 0.3690 1.4463 0.1915 1.927 0.054 .

Signif. codes: 0 ‘’ 0.001 ‘’ 0.01 ‘’ 0.05 ‘.’ 0.1 ‘ ’ 1

``````            exp(coef) exp(-coef) lower .95 upper .95
``````

Lymphnodes_statusPositive 1.446 0.6914 0.9937 2.105

Concordance= 0.537 (se = 0.018 )
Rsquare= 0.003 (max possible= 0.835 )
Likelihood ratio test= 3.46 on 1 df, p=0.06
Wald test = 3.71 on 1 df, p=0.05
Score (logrank) test = 3.76 on 1 df, p=0.05

Hi Wu Jie

below you will find a nice Wiki where most of your questions should be explained... As Note: You have to distinguish between 'Global statistical significance of the model' and the 'significance of each coefficient'. Best regards