Posted by Larry Hoover on May 10, 2006, at 8:31:37
In reply to Re: Statistical question on SSRIs - Psychobabble says » Squiggles, posted by Larry Hoover on May 9, 2006, at 17:21:21
> > > "What was found was that there was a "significant change in slope" (a reduction) of the suicide rate, following the introduction of SSRI meds. A change in slope can only be caused by a change in the independant variables"
> > >
> >
> > Sorry, is this a quote from the Healy paper?
>
> What you've got there is a me quoting the Sweden population study authors.
>
> > And when you describe a slope ( i guess that
> > is on a statistical graph ) as possible only
> > by independent variables - what would those
> > be? Are they typical of this clinical study
> > alone?
> >
> > Squiggles
>
> A plot of "all cause suicide deaths" against "total population" for consecutive time periods would yield a graph where the first derivative, the slope, is equal to the rate.
>
> The rate was stable over two periods of time, but different, one from the other. One time period of stable rate preceded the stable rate seen around the time of the SSRIs. Inferences were drawn.
>
> LarSorry for the interuptions, Squiggles. I had a medical test done Saturday and it's still messing me up.
My quick summary of the methodology for gaining the measure of the slope, i.e. the rate, was incomplete. I skipped one step. I do it all the time in my brain, but I forgot how to describe it.
The revised version:
A plot of "all cause suicide deaths" against "total population" for consecutive time periods would yield a graph where the first derivative of the best fit line, the slope, is equal to the rate. In this case, plots from two periods eighteen years apart yielded different best fit lines (i.e. different rates). Statistical testing showed that the difference between the two rates was significant. The rate during a period of high sales of antidepressants (including SSRIs) was significantly lower than a period when antidepressants were prescribed less often.
Then I went into blather, which is why my brain called a halt to the proceedings. I have a weird brain.
I know very well that there are alternate explanations for the significant inverse correlation (i.e. I'm not trying to say SSRIs brought it down). The fact that it is inverse, and significantly so, places an upper bound on the effect size for your hypothetical SSRI-mediated suicide induction process.
As it happens, on reloading the pages I was working from last night, I found another more recent Scandinavian report.
Therein, it is stated that, "In all four countries decreases in suicide rates appeared to precede the widespread use of SSRIs."
As my intent here has been to show that SSRI-induction hypothesis has an upper bound, this new evidence is also consistent with my earlier conclusions.
If I had evidence to support your hypothesis, Squiggles, I'd show you. I haven't found any, yet.
Lar
poster:Larry Hoover
thread:640557
URL: http://www.dr-bob.org/babble/20060510/msgs/642064.html