Replies to post #69490 on Biotech Values

[mcbio]

? Not sure what you are asking different than I already expressed. If HR=0.60 on a trial with 40 patients per arm then only a small percentage of patients could have died so far. As a greater percentage die the HR and the p value will change.

Ok thanks, I was just trying to ensure that I understood your point and I do.

The more interesting point is that they gave enough data to hint that the HR through, say, 20 months was a lot better than 0.60.

But again, that really doesn't mean much if you don't also get a p value <.05 does it? And it doesn't seem like we have any way of knowing the p value given the data we have so far. Are there instances where hazard ratios still matter even if p value >.05?

[mcbio]

One other point regarding the OGXI PR http://biz.yahoo.com/prnews/081203/to979.html?.v=4

"First-line trial currently shows median overall survival of 27.5 months
for OGX-011 in combination with docetaxel and prednisone and a 16.9
months overall survival for docetaxel and prednisone alone."

Does the fact that the headline makes a distinction between median overall survival of the OGX-011 arm and overall survial for the control arm imply perhaps that all events have occurred in the control arm?

There's also this language later in the PR: "Results currently indicate that patients in the OGX-011 arm have a death rate approximately 40% lower than patients in the control arm. The current results are based on study data with a median follow-up of approximately 30 months for both arms. Additional survival updates are needed before a mature median survival for the OGX-011 arm can be reported. Based on the current results, OncoGenex has calculated that the final median survival for patients in the OGX-011 arm can not be lower than 22.7 months." By not also referring to the contol arm in these last few sentences, this almost implies to me that the company is saying they have a final overall survival number for the control group and that is 16.9 months.

My point is that if we know 16.9 months is the final overall survival number for the control group and 22.7 months is the minimum overall survival number for the OGX-011 arm (again, assuming all events have occurred in the control arm but we must wait on the remainder of events in the OGX-011 arm), then this data should likely be viewed in very good light even if a p value wasn't quoted.