Cox regression survival analysis

[Victoria Nunn]

Hi, I am currently undertaking my dissertation and wondered if you could help me with some data analysis from one of my studies. The study indicates that they use Cox regression for survival analysis
Results indicate that for group differences in depressive episodes the hazard ratio =0.38 95% and confidence interval is 0.19-0.75; p<0.006

I have read up on this and cannot seem to understand what it means. Is there any way you could explain this in simple terms. Thanks X

 

[rogojel]

hi,
Suppose you have two groups, then one group has about 62% less risk then the other group. The risk reduction in he real population (as opposed to your sample) could be as high as 80% or as low as 20% and will, very probably be somewhere between these extremes.

I hope this helps a bit

 

[hlsmith]

Rogojel did a good job describing this. Though, I thought I'd try to add a little. The outcome consisted of a binary variable and so did the predictor. Now given your sample is just one version of all of the possible samples, you can be 95% confident the true risk is within the interval.