You don't understand do you, or are you playing ignorant?
Why was that halabi question asked in the first place? Correct me if wrong, they were afraid that 9902b patient population is sicker, and therefore would not respond to provenge as well, resulting in an attenuated HR, decreasing chance of success, right?
Now, this question essentially implies that the presence of pain symptoms is an effect modifier of treatment.
There is no evidence that pain (defined as <3 on an AScale) is prognostic ie predicting OS. Even if there is, it will only impact the treatment outcome if it is significantly imbalanced.
If balanced, the only way it will affect outcome is that modifies treatment effect, ie patients with pain respond less well than those who do not have pain, but since even Tax237 study did not show significant interaction (when it used an AS >10 to define significant pain), then how likely it is for minimally symptomatic pain when it could compose of no more than 1/3 patients?
Market Data 
Markets 
AI
