outliers in survival probability

[orb]

Hi all,

We use survival data of 90 patients that were divided in two sub groups. (31, 59).
Two patients (out of the 31 sub group) survived 50% longer than all others.

Are they considered outliers? What are the criterions to define outliers?

Thanks for your help,
Or

 

[Miner]

What criteria are you using? Survival data are rarely normally distributed. It is often highly skewed in either direction.

 

[ted00]

I personally don't recall ever hearing/reading talk of outliers in survival analyses. E.g., in this entire book chapter on Cox proportional hazards model the word "outlier" doesn't appear once