Say Nay to Northfield

This is the point at which diehard Northfield bulls are yelling at me for not going into a detailed explanation of the non-inferior statistical test and so-called confidence intervals, which they believe ultimately will vindicate Polyheme and convince the FDA to approve it.

Nonsense. There was a 38% relative difference in the death rates of the two arms of this study, going against Polyheme. You can throw whatever statistical test you want to against that, but intuitively, there is no way that the FDA, or anyone else, would deem Polyheme non-inferior to saline and blood.

Think about it another way (and I apologize for the gruesome example): Your mother is bleeding to death and an EMT coming to her aid asks you a question: He can give your mother Polyheme or saline in the ambulance on the way to hospital. In a clinical trial, the former caused more deaths than the latter by a relative difference of 38%. Which would you pick?

Even Northfield isn't so irrational to make the case for Polyheme's approval based solely on this data. Instead, the company's parsing the results. Apparently, there were a lot of mistakes made while conducting the trial, to the point at which 126 patients, 18% of the total, were deemed to be "protocol violators," meaning they shouldn't have been enrolled in the study, or their data should have been kicked out for various reasons.