> >> > What are the problems with applying the scientific 
> method to human 
> >> > health issue?
> 
> >> The basic problem seems to be in the assumption that test subjects 
> >> are nearly identical, ...
> 
> > That's ok.  Just increase your test base (ie more people) 
> and assume 
> > it averages out.  The more the merrier (bar the outliers, 
> they may be 
> > unhappy).
> 
> That's the exact issue which prior posters are commenting on.
> "Statistical significance"is generally taken as starting 
> around 2 standard deviations or about 2% in each tail 
> (although you can draw some good conclusions at lower 
> confidence levels than that (and the latest sweep all before 
> it Global Warming report is based on a 10% level !!!!) .
> 
> With a sample of 1 million people and 2% in one "tail" (and 
> assuming that the 2% in the other tail are uproariously happy 
> at how well THEIR system responds to a given drug, that 
> eating excess carbs make them thinner and that smoking 
> decreases their cancer risk, then that's 10,000 very unhappy 
> people that the results are unrepresentative for. 
> In the whole of the US that's about 3 million very unhappy 
> people. And in the world that's about 200 million very unhappy people.
> 
> As you fix the "easy" problems and turn to the hard ones you 
> meet the unrepresentative people, and if that happens to 
> include 'you and yours' you are not going to be at all happy 
> with a statistical treatment.


And the alternative?  Everyone unhappy?

As you initially pointed out, some bits of biology don't really make sense.
Until they do, we're stuck with rounding up as many guinea pigs as possible
to see what happens.

I've had diseases for which there isn't a cure (hepatitis-A) besides 'go
home for a month' and other conditions (epilepsy & migraines) where 'this
may or may not work'.  

I have found a migraine drug that 'sorta works'; it take the edge off the
migraine but has side effects in my case.  One was described as 'nasal
irritation', which I guess really means 'you'll bleed like a stuck pig in
your sleep'.  In that case, I'll wear the side effects.  If anyone out there
gets migraines as well, you'll know waking up covering in blood is pretty
trivial in comparison.

Amusingly, one drug had listed 'may cause headaches & nausea'.

Tony

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist