Posted by Jane D on August 24, 2001, at 23:05:49
In reply to Re: martin jensen book - , posted by Lorraine on August 24, 2001, at 22:27:02
>
> Right. The problem with sequential long term trials is that if you are treatment resistent, it could be 25 years before you hit the right combo. If you look at Preskorn's column "Do you feel lucky" (www.preskorn.com/columns/9801.html), he does the math using 600 possible drugs in combinations of 2, 3, 4 and 5. The odds of hitting the right combo is 1 in 179,000; 1 in 35,820,200; 1 in 5,346,164,850 and 1 in 637,262,850,120. My thought when I read this was, well why is he using a universe of 600 drugs? I mean I have eliminated certain classes of drugs (SSRI's) by this time and why not just look at the big players. So I took my remaining drugs to try short list drawing from the categories of stimulants, MAOs and Anticonvulsants. I came up with a list of say 18 drugs to try in various combos (hey, this is already much better than 600, right?). Then I eliminated those drugs on the list that I had already tried unsuccessfully. Now my list is down to 10 drugs to try in various combos--not too many right? Ok, but when you crank out the math, my odds using a two drug combo are 1 in 45 and using a three drug combo are 1 in 240. Well, you get the drift, we could be going through trials a long long long long time if we do "adequate" trials. There is a fallacy in this thinking, but I think it is small. The fallacy is that once you eliminate a drug b/c of side effects, you don't need to include it in combos and Preskorn's formula (which is pure statistics) does not take this into account. However, my use of just 10 drugs is surely too small as well so it all evens out in the wash.
....Lorraine,
There is a fallacy in the original calculation too and I think it's a big one. It's the assumption that there is only one combination that will work. Probably there are many. There is so much overlap in what these drugs do.Jane
poster:Jane D
thread:76220
URL: http://www.dr-bob.org/babble/20010822/msgs/76307.html