Hi Benelong the results did appear to strongly favour the faith side. But too much faith is required.
The problem stems from the efficacy table where the Ns are incorrect; which HT first spotted.
At the time I believed that it was a PR person type stuff up; and said so. Unfortunately from the conference call this appears not to be case.
SI believes that including the detail about the correct Ns that were used calculate those percentages is too much information for an announcement.
Lets accept this.
But then what is the excuse for the 20 page presentation that contained half a dozen completely redundant slides of sheep that we have all seen dozens of time before.
What SI is doing, as a later analyst twigged to, is not wanting to reveal the data of patient flow and in particular the number of subjects rescued out of their allocated groups. In other words subjects are switching treatments.
This is equivalent to me trying to sell you a second hand car but refusing to let you look under the hood.
The bottom line is that without this information it is impossible to assess the validity of the efficacy results.
The problem of patient switching occurs because it is unethical to keep subjects in allocated groups if they are not responding positively and there are treatments available that could help them.
Naively just excluding subjects who switch is well known to produce biased results. And the strong suspicion is that is exactly what MSB have done in that table. The extent to which is actually a problem (as opposed to a theoretical problem) needs to be tested through sensitivity analyses where different modelling assumptions are made and a range of plausible effects calculated.
Two explanations.
One is this information is withheld to deliberately mislead. The other is MSB cannot understand how ridiculous it makes them look to present a table to the market where the numbers don’t add up. On the idea of dummying things down for the market.
Neither explanation is really that great.
Having watched a few trial announcements now I’m amazed how MSB finds ways of shooting themselves in the foot each and every time. The process through which MSB is producing trial result market announcements is fundamentally flawed.
Having said all this my gut tells me the efficacy signal is still encouraging and the sleight of hand that has gone on is not fatal. I also think the surrogate endpoint idea in an adaptive P2b-P3 trial is potentially a real winner with the Cures Act.
I suppose what is also grating is that you read the Bell Potter report about the trial and they produce the same dumb version just with different words. The concepts and complexities are way too hard for them.
But its Friday. So its past the time for serious stuff. So hence a funny video about the problems of patient switching which is at the heart of the problem here.
(20min delay)