Replies to post #66619 on Amarin Corp Plc (AMRN)

[ggwpq]

Thanks for keeping up with the R-I interim reminder update. It's helpful to be reminded at least on a weekly basis.

[Sam81]

Projected Timeline for Interim Analysis

Raf,
Can you please explain the numbers you got?


As per H.C. Wainwright Update (Nov-2015) they mentioned a timeline projection which goes as follows;

4.68% Annual Placebo Rate: 10% Reduction, 967th event in Apr-16 / 15% May-16 / 20% Jun-16 / 53% Nov-16

5.20% Annual Placebo Rate: 10% Reduction, 967th event in Jan-16 / 15% in Feb-16 / 20% in Mar-16 / 30% Apr-16 / 53% Aug-16

5.72% Annual Placebo Rate: 10% Reduction, 967th event in Dec-15 / 15% in Jan-16 / 20% in Feb-16 / 30% Mar-16 / 53% Jun-16


So taking the worst case of annual placebo rate of 4.68% as per HCW, we can assume that once May-16 passes, V most probably will be stopped at Interim

Considering other placebo rate scenarios, it is safe to assume once Feb-2016 passes, also most probably V would be stopped due to good results

[Biobillionair]

CVE rates for diabetics where >10% in IMPROVE-IT... IMO we are looking at a higher than expected efficacy >40%.

BB

[chas1232123]

Raf – I ran a few numbers and got results very similar to what you posted. I used a simple model with constant placebo and RR (risk reduction) rates (these are approximations). I used the following estimated quarterly enrollment numbers (in reverse chronological order starting with Q4 of 2015):
8000
7900
7800
7450
7350
7150
7000
6800
6500
6075
5535
4451
3043
1756
575
108

I used these estimates to estimate the current number of subject-years at about 20,500.
This leads to the following minimum estimated risk reduction if event 967 has not yet occurred, for various assumed placebo rates:


Placebo Rate (%) Risk Reduction (%)

5.0 , 11
5.2 , 18
5.4 , 25
5.6 , 28
5.8 , 31
5.9 , 34
6.0 , 37
6.2 , 40

There are various ways to do this type of estimate based on incomplete information. The estimates are approximate and it is natural for different people to get slightly different results.