Absolutely! That's the reason why I try to do some 'sensitivity' tests.
I calculate, if possible, at least
1. Some statistical test (Fisher's exact, Log-Rank etc.) to check for type 1 error (even if values are given, I try to double check). In this case, 6/18 or 6/19 vs 4/23, Fisher's two-sided exact test gives p = 0.289 or p = 0.468. Now I know that statistically speaking there is about 30 to 50 % chance that the prurisol phase 2a results are a lucky coincidence and 50 to 70 % chance that there is at least the difference of 31.6 to 33.3% vs 17.4 % in 2+ point IGA drops.
2. Check historical values for like drugs i.e. oral psoriasis drugs. The only ones that I found with near comparable results available are Otezla Esteem plus Liberate trials. I repeat results with a worst case correction I did NOT DO BEFORE (my bad).
Otezla vs placebo, sPGA/IGA response: Final IGA = 0 or 1 and 2+ point drop in IGA at week 16.
Esteem 1: 21.7 to 30.4 % vs 3.7 to 5.4 %
Esteem 2: 20.4 to 28.1 % vs 4.4 to 6.9 %
Liberate: 21.7 to 27.3 % vs 3.6 to 5.0 %
About that worst case correction and some other things. Otezla numbers include IGA 4 subjects, prurisols do not. Worst case assumption: Only IGA 3 subjects in Otezla trials satisfied sPGA/IGA response. That gives you the upper limits you see above. Note that All results are for week 16, prurisol results are for week 12. Eyeballing from available graphs Otzela scores improved slightly between weeks 12 and 16. Hence what you see are slight overestimates for Otezla when compared to prurisol.
So, I learned something, again.
A. Otezlas result are very consistent.
B. With worst case correction prurisol percentage is barely above Otezlas upper limits, but it is above.
C. The percentages for placebo in prurisol trial is well above that seen in Otezla trials. Maybe my assumption that all 2+ IGA drops in placebo happened in IGA 3 group is too harsh. That has consequences in Fisher's exact test. 6/18 to 3/23 gives p = 0.147 or p = 0.257. In this case there is about 15 to 25 % probability that the result is just a coincidence and 75 to 85 % change that the difference is at least 31.6 to 33 % vs 13 %.
The difference between Infinity and me is not that we do not agree about the above results. I think we do. The difference is that we see them differently. For Infinity the above is still too high risk, to me it is a risk worth taking. Currently we are in quantum situation: neither of us is wrong, yet. That gets resolved with P3 trial or just possibly earlier.
About a month or so ago, when I was still mulling over whether to write an article about IPIX I came across this trial. Otezla Unveil trial. It happens have ONLY moderate subjects, IGA = 3. So here is UNVEIL:
Otezla (n = 148) vs placebo (n = 78), percentage of subjects having PGA/IGA score 0 or 1 at week 16
30.4 % vs 9.6 %
The numbers for 2+ point IGA drop are also the worst case scenario for prurisol trial considering IGA 0 or 1 score. So let's repeat:
prurisol (n = 18 or 19) vs placebo (n = 23), percentage of subjects having PGA/IGA score 0 or 1 at week 12
31.6 to 33.3 % vs 13.0 to 17.4 %
So far, prurisol is, at least, tracking Otezla. And if the situation stays the same it is more than enough because prurisol has one major trump card. Otezla is immunosuppressant, prurisol is not. Prurisol is based on abacavir which is an antiviral. A slight exaggeration:
With Otezla psoriasis goes and tuberculosis comes.
With prurisol psoriasis goes and so does influenza.
