Yes.. This is the point...
Statistics say it takes more numbers to prove a point if the event you are measuring is not very probable to begin with, and you are trying to prove is your treatment will make it even more unlikely.
In the JELIS trials only "Sudden Death" and Fatal MIs are listed. The cohorts are 9000 treated and 9000 placebo...The incidences of Sudden Death was 0.2% in both control and Placebo...
The incidence of Fatal MI was 0.2% in control and 0.1% in the EPA treated group....Translated into numbers this means there was only nine deaths in the EPA group due the fatal MIs, 18 in the control group and a total of 63 deaths in 18,000 due to these two important events...
We know from Biomaven's post on stroke...These numbers are simply not big enough to offer signicance when the incidence is this low...The reason why the CV "EVENTs" were significant is because totaled they represent a signficantly higher incidence.
Dr. Rodriguez should be reminded that not guilty does not mean innocent. It is altogether logical that if a treatment results in stat sig lower indices of a number of events which a knowledgible person would recognise as dangerous or life threatening, and does not increase the likelyhood of other conditions which could be fatal ...Then the burden of proof should be on those who say it does not improve survival...In this case Dr. Rodriguez was intrigued by the fact EPA decreased most of the CV events, but questioned the importance of these because she felt there was little effect on the mortality...She should have wondered why the mortality rate was not lower since most of these events are risk factors.
We see this effect in a lot of cardiology trials...The LP-PLA2 test Dew mentions...LP-PLA2 increased levels have been show to be stat sig predictive of all cause mortality (a classic large group) yet because the CV mortality group is smaller and does not show stat sig the test is not recommended as a predictive test..Even though experts agree LP-PLA2 is the very heart of AS inflamation...
Nice work Biomaven...
":>) JL
Market Data 
Markets 
AI
