HGEN's falsehood Something I've been meaning to

HGEN's falsehood

Something I've been meaning to take a look at for a while.

Let's reverse engineer HGEN's reverse engineered statement of 37% improvement for lenzilumab. How did HGEN come up with 37%? They divided the initial recovery target for patients of 257 by the new target of 402. That is faulty methodology. The trial was not going to hit stat sig at 257 so an increase greater than actual efficacy was needed to try and hit stat sig.

We know 29% improvement was needed to hit the minimum statistical value to not shut down the trial due to futility. That was most likely set at higher than p= .05 with an eye towards it achieving p= .05 with the increased number of patients at full enrollment. If lenzilumab had hit that p= threshold than no additional patients would be needed. That 37% additional patients were needed to achieve stat sig shows efficacy must be below 29%.

If you believed HGEN's explanation in their PR then you've been had.