Replies to post #31173 on Amarin Corp Plc (AMRN)

[Whalatane]

Hi Redlyon

Biwatch has done the most work on statistical insight re Reduce it .
He wrote an article that you can find in a search of Seeking Alpha ..as well as posts on this board. I can try and find the post #'s if you can't find them

HDGators as well as JL and Ziploc ,I think have also posted on it .

Going from memory I believe the guess that makes the most sense to me is a 20-25% benefit over the placebo ( statin only ) by late 2015 /early 2016 with 90% confidence ( in my lay mans terms --meaning enough patients for long enough that what you see is what you get )
In Adcom I believe Amarin stated their had been over 200 events by that date ..late 2013 , did not disclose ( and may not know ) how many of those were statin only patients .

Anyway ..I would enjoy a more rigorous P value discussion ..but full disclosure ..my bio stat education is limited to a Johns Hopkins online course I took .
I'm interested in your perspective . Some are hoping for an early 2015 date for the trial to be stopped ..personally I think thats very optimistic given that this is a CV trial and DMC's usually want at least 500 events and up to 25% separation of event lines before advising co

Kiwi

[Whalatane]

Redlyon


Quick follow up
Biwatch post # 27162 and JL post # 29020 as a start re Reduce It trial
HDGator has some calcs also

I'm more conservative in my expectations .

I'd be interested in your pt of view /calcs

Kiwi

[HDGabor]

Reduce-IT

With eff. 25% and 5,0% events rate (I decreased from 5,2% since the March 13 modification - TG min. 200, instead of 150 - based on lower than expected events) app. YTD data:
- Enrollment: 7.250
- Events: 439 (188 Vascepa - 250 placebo)

I could not find any guidelines or standard for DMC driven early stop, but using Kiwi's determination "DMC's usually want at least 500 events and up to 25% separation of event lines before advising co" the 500th events will be around beginning of Oct 2014. (I could not calculate the exact power, but based on Biwatch calculation for different number of events - used in #28220 - it will be more than 90%)

We also know that the DMC meeting occurs quarterly (let's assume they will meet just before the 500th events, so they will recognize it at the next meeting).

So if the 500 events and 25% is enough for early stop DMC will advise it by end of 2014 / beginning of 2015.