1 Boring Old Man

Nihil sapientiae odiosius acumine nimio
Nothing is more hateful to wisdom than excessive cleverness

^{This epigraph to The Purloined Letter attributed by Poe to Seneca was not found in Seneca’s known work. It is from Petrarch’s treatise "De Remediis utriusque Fortunae". [ref]}

I thought I had my ducks in a row, so I submitted my letter. It moved from place to place until Friday [the 13th]…

… when I got this:

I was at work at the clinic and it came on my phone. At first I thought it was an acceptance, having only a moment for a glance, but at lunch I read it and realized that wasn’t the case. It’s a rejection "for publication" and an acceptance for some online forum only available to subscribers. I’d been warned about this and knew it was the cantina at the end of the galaxy – out of sight and mind. My point in sending it in the first place was to have Gibbons’ articles flagged in PubMed. So I decided I’d rather just post it here than have it becalmed in the doldrums on their site – a purloined letter. And as I thought about it, I had so sanitized the letter that it didn’t really say all I wanted to say anyway. I mentioned the obvious – inappropriate data selection, faulty arithmetic, and a peculiarly inflated outcome. But what I didn’t talk about was the deceitfulness in the presentation.

The study was a meta-analysis, yet with the initial article  we weren’t even told what trials were being analyzed. As my focus was on the child data, I searched high and low for the included Lilly Trials in vain. And during that month, Dr. Gibbons was all over the media making broad, clinical pronouncements. On the day the first article was published, he was quoted by the LA Times as saying, "I hope that the warnings will not prevent depressed children and adults from getting treatment for depression. The greatest cause of suicide is untreated or undiagnosed depression. It’s very important that this condition be recognized and appropriately treated and not discarded because doctors are afraid to be sued." On the day after the first article, he was on NPR saying, "I worried that what we might end up with was a real epidemic of suicide. And the data suggests that that is exactly what happened. Rather than the black-box warning leading to decreases in child suicide rates, they were followed by some of the largest increases in child suicide rates both here in America and around the world." That’s a remarkably distorted comment, verging on an outright lie. It’s true that he and a number of other groups jumped on a minor up-tick in suicides in 2004 that promptly disappeared [see the apogee…, pretty loud coi…, tortured numbers…, peaks and valleys…, watchful waiting…]. Later in the month, he was on Medscape talking about the original black box warning. "Dr. Gibbons, who was on the FDA’s advisory committee that voted in favor of the black box warning, said he was very concerned about the validity of the data that prompted the affirmative vote. ‘The adverse event reports for suicidal thoughts and behavior showed a fairly small signal in children, but the prospective measurements showed no effect of treatment whatsoever. As a statistician, I put more weight in prospective data than these retrospective reports,’ he said." That’s a surprising comment, since he’s talking about his own meta-analysis of retrospective reports.

When the second article came out a month later, it did have a table with the trials, but the ringer, LYAQ, was never described and was not referenced  – either the on-line version on the Lilly site or the published paper. Had there been any mention of the fact that this was an ADHD study of Strattera, that less than half were diagnosed with MDD, and a third didn’t meet their criteria for depression on the CDRS-R [< 36], no one would have accepted the inclusion of this study. And his analytic statistical methods were written in a way that no General Psychiatrist could possibly follow – more tangled web than description.