First off, I'm not a mathematician. Maths, or if you're American, math, is another language to me, in fact, for me at least, it would be easier to learn Chinese than it would to grasp maths. I was the kid at the back of the classroom who still counted on his fingers underneath the table, out of view of those who were better at numbers than I.
I was going through some of the archives and recent media articles regarding the original 329 study.
There were 275 subjects in the 8 week trial, 12 of whom had suicidal thoughts (in the paroxetine arm)
12 suicidal thoughts out of 275 subjects = 4.363636363636364%
Okay, out of the 275 subjects, 90 were on paroxetine. The table below is based on efficacy results for 90 patients. The actual figure of adverse events was based on 93 patients but I'm keeping the figure at 90 for ease.
12 suicidal thoughts out of 90 subjects = 133.33333333333334 (Roughly 13%)
Given there were approximately 2 millions prescriptions for paroxetine written for adolescents one year after the study was published in the JAACAP this means, if I have my maths correct, that...
133.33333333333334 of two million = 266,666
In 2003, GSK's then Head of Psychiatry, Alastair Benbow (pictured above) went on national television and claimed that, "less than a small class size would have these suicidal thoughts." (video below - 1 minute 21 seconds, Alastair Benbow
Benbow's 'small class size' is a classroom, therefore, that can fit 266,666. pupils. What school did Benbow attend?
Okay, so two million prescriptions does not necessarily mean two million adolescents. The two million prescriptions may have been repeat prescriptions to individuals. So, lets say the number of individuals written a prescription is between 200,000 and 2,000,000
Whatever way you slice it, Alastair Benbow must have gone to an awfully big school!
To put this into some sort of perspective, America de Cali, a Columbian football team, play their homes games at the Pascual Guerrero stadium which holds some 33,000 spectators. So, if 250,000 individuals were given prescriptions for paroxetine one year after the JAACAP review by Keller et al had been published, Benbow's small classroom would have looked something like this...
Okay, let's give Benbow the benefit of the doubt here and say that just 175,000 individuals took paroxetine one year after the JAACAP review.
That would give us (excuse the Devil numbers) a class size of 16,666 pupils.
In other words, Benbow's small sized class could be compared to the Nilson Nelson Gymnasium in Brazil (pic below)
Furthermore, and once again cutting Benbow some slack, let's say just 87,500 took paroxetine one year after the JAACAP review - that would still give us a pretty large classroom, - 8,333 pupils.
An averaged sized class (not small sized class) in the US was between 19 and 26 pupils in 2011/12. (Source)
Remember, Benbow claimed that, "less than a small class size would have these suicidal thoughts."
What exactly is less than small? English language and usage may be worth looking at to understand what Benbow was actually saying. (Smaller vs. less vs. lesser) Maybe it was yet more coding from GSK (see http://study329.org/) for explanation of "emotional lability"
Maybe Benbow was talking about the 12 patients who had suicidal thoughts that the restoration team found... or maybe he was talking about the 6 that Keller et al found?**
Either way, it was an extremely poor choice of words to use. Perhaps Benbow should have used the same criteria I've used here, although going on national television and saying something like, "We believe that between 8,000 and 266,000 would have had these suicidal thoughts one year after our study was reviewed and published" , wouldn't really have gone down too well with GSK executives, or, indeed, their shareholders.
Truth is over two million prescriptions were written for children/adolescents one year after the Keller study was published. It is unknown how many children/adolescents were individually prescribed paroxetine, it's believed to be, as I mentioned, between 200,000 and 2,000,000. Even if the figures are less we still see a potential for thousands of adolescents and children that may have been at risk of suicidal thinking. One also has to take into account that these figures are based on US prescriptions only. Add to it prescriptions written for adolescents and children in other countries and we have stadiums bursting at the rafters!
Maybe a more striking headline regarding the recent restoration study findings should have been 'Between 8,000 and a quarter of a million kids put at risk one year after flawed paroxetine data published.'
The interview with Benbow (below) is from 2003, some two years after the Keller paper was published.
Some patients did go on to take part in a 6 month continuation phase and these were followed up at various points, but this data has never been published.
If any statistician stumbles on this post of mine they may be able to clear things up for me?
**Keller et al only classed 5 of 6 as serious adverse events (suicidal ideation/gestures)