Mmm having had a chance to look at these results more carefully. This morning glass half full – tonight extremely negative.
Forget the absolute legion count and percent legion count change. Simple legion counts do not reflect the variable expression of acne and analyses of these are very sensitive to outliers. A percentage change statistic is very sensitive to baseline differences between groups in severity.
Focus instead on the Investigator Global Assessment (shown below) which was rated at the end of the trial.
The IGA provides a direct FDA recommended method for establishing clinically significant change. The FDA want to see a two-point IGA reduction. See figure.
View attachment 1794257What do we get here:
Aus: 14% meeting the criteria from clinical improvement (combined treatment) versus 7% clinical improvement (combined placebo)
US: 15% meeting the criteria for clinical improvement (combined treatment) versus 17% clinical improvement (combined placebo).
Best case (Australian data) we have net benefit of clinically significant change for less than 10% of subjects. Worst case (US data) it is zero.
Would you sign up for a treatment that helps (produces clinically significant change) for less than 10% of people. Not if you had any other options.
What exactly is the typical placebo IGA placebo response? Google a bit - you will find IGA placebo responses of between 10-20%.
https://www.ncbi.nlm.nih.gov/pubmed/30235387https://www.ncbi.nlm.nih.gov/pubmed/30681791https://www.sciencedirect.com/science/article/pii/S0190962219303354The US placebo response (17%) is around the mark. It is the Australian placebo response that is the anomaly.
Why do you get placebo responses in the first place? Better cleaning regimes that are part of the trial protocol. It appears our Aussie teenagers might have let the side down here in the cleaning compliance department.
Why should anyone believe a word of this?
Is it any less believable than the idea that BOT accidentally created a highly effective active in their vehicle through a manufacturing stuff-up. If anyone can find in the history of clinical trials someone stuffing up the manufacture of a placebo and accidentally producing a highly effective new drug treatment … post it here.
But second … there is proof (or rebuttal) very easily made.
The power calculation for this study will tell you the placebo response expected as well as the treatment effect they hoped to find. I think you will find the placebo response assumed in the powering was 10%+ and the treatment effect net of placebo for IGA translated to 15%+.
BOT can easily provide that power calculation (it is not by any stretch of the imagination commercially confidential). In absence of this proof I wouldn’t believe the US placebo manufacturing problem for a second. To me it just looks like one big red herring; a very successful exercise in misdirection … brilliantly executed by the company.
From here .. this trial will quietly disappear. No-one is going to spend money on a P3 acne trial with P2 trial results that show in the best case (Australian results) less than 10% of people show clinically significant improvement. With the distinct possibility the US results are true and it is of zero benefit.
Not what anyone wants to hear I appreciate ....
(20min delay)