Cox proportional hazard regression - overall model fit and covariate significance

#1
Hi!
I am running a univariate Cox proportional hazard regression and I am not sure how I should interpret the obtained results.
In my model, I have only one single variable (X) and the overall model fit is statistically significant (p<0.05) with the following results:
Null model -2 Log Likelihood
90.738
Full model -2 Log Likelihood
84.473
Chi-squared
6.264
DF
1
Significance level
P = 0.0123


but the covariate "X" is not significant:
b -0.1048
SE 0.07542
Wald 1.9295
P = 0.1648
exp(b) 0.9005
95% CI 0.7768 to 1.0440

How should I interpret these results?
Thank you for your help!
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
All of these output were generated by the program?

Well the -2 LL test is good at accessing nested models. I am not sure where it eventually deviates from the HR estimate. Sorry, for not being more help, but in the interim I would go with the coefficient estimate, since it includes the null. Most of the times one prefers to error on the side of cautions, AKA null. Post back if you find a good resource explaining why the results partially contradict each other.

What is the sample size?
 
#3
Thank you.
Yes, these data were generated by the program MedCal. Similar results are found with R.
The sample size is N=30.
 

hlsmith

Less is more. Stay pure. Stay poor.
#4
How many event were there and what type of variable was the predictors (e.g., continuous, categorical)?