Unfortunately considering their great importance to medicine, science, patients, companies, and investors alike, clinical trials are not an exact science. Several aspects of trial design that can impact perceptions of outcome are controversial. I don’t expect many of you to read it, but the excellent article shown below by the English clinical trials statistician Stuart Pocock illustrates the point. The use of composite endpoints is one of the controversial practices, and we have seen with BETonMACE why this is so. Take out stroke and replace it with CHF readmissions and you have changed a nonsignificant result to a positive one.

There was a time when composite endpoints were forbidden as potentially misleading. Then everything changed. Why? Merely to make clinical trials smaller, shorter, and cheaper. I personally do not believe stroke should ever be lumped together with heart disease, as it can be due to blood clotting or bleeding.  Nevertheless, it is a very common practice and we are stuck with it. But the FDA and BP are very familiar with these issues, and no doubt will take them into account. So in that sense it’s not such a big issue.

http://www.onlinejacc.org/content/accj/66/23/2648.full.pdf