The Great Glaxo Transparency Swindle

For those who believe in the fairy-tale that Glaxo will ever manifest any real sense of transparency -or  give uncensored/uncontrolled access to their private trove of clinical trial data..    

..get real.

329 2

GSK 329

Glaxo ‘Falls Short Of Open Data Disclosure:’ Jureidini Explains

Over the past few months, a group of researchers has been haggling with GlaxoSmithKline over access to detailed data for an infamous 2001 study of its Paxil antidepressant called 329 that tested the pill for treating depression in adolescents. The researchers, who are led by Jon Jureidini, a clinical professor of psychiatry at the University of Adelaide in Australia, want a 1998 clinical study report that they hope to reanalyze and republish. The original results reported that Paxil was effective, but the trial actually missed its endpoints and figured in a ghostwriting controversy (here is the study). As a result, Glaxo signed a consent order with the New York State Attorney General to publicly disclose trial data and has since vowed to do so. But the drugmaker says that Jureidini, who also led an unsuccessful quest to have the study retracted (see this), must follow its new system for submitting proposals (back story and more here). We spoke with Jureidini about his efforts. This is an excerpt of our conversation…

Pharmalot: Why did you make this request?

Jureidni: I’ve been occupied with trying to get some truth about that study for a long time. And just because it had such an impact in child psychiatry in Australia at the time it came out. It was used to try to shift the approach to (treating) childhood depression and (Paxil) became first line (therapy). I looked at it the first time I read it and thought perhaps medication is the answer. But then, I reread it and spotted some problems. I then had others look at it, and it seemed flaw and then deliberately so. This was in 2002 and 2003.

Pharmalot: Why does this matter now, though?

Jureidini: We have an editorial failure by a major medical journal. We have a failure of the peer review process. We have misconduct by a major pharmaceutical company. Most worrying for me is that we have a careless, at least, and probably more than careless, work by senior academics. And we have the whole issue of ghostwriting. Most of all, we have harm potentially done to children. Those sound like pretty significant issues to me.

Pharmalot: So what exactly are you seeking?

Jureidini: The primary goal is to get the data… We want to use the data to check what they published… Our guess is that the representation of the efficacy data in the clinical study report is accurate, but we have real concerns about the accuracy with which the adverse event data that has been transferred from individual patient records to the clinical study report, which is why we wanted individual patient records. The fact that it serves as a test of GSK resolve (to disclose trial data) is welcome to us, but not our primary interest.

Pharmalot: Glaxo has said publicly, though, that it has procedures for releasing data now and will review researcher requests.

Jureidini: The procedure doesn’t meet our requirements. That’s the problem. One of the positive things about us testing their process is that we’ve already illustrated we’re carrying out a project with GSK data, but the system doesn’t meet the needs of the research group.

Pharmalot: Which is….?

Jureidini: They won’t provide individual patient records. They’ve defined what they will make available and make it sound that’s all anybody what would want in order to behave ethically. We’re concerned that this (approach) gives them a level of control over the data, which might facilitate them withholding inconvenient truths. We haven’t been given any data yet, but you’ll see from the letters from Shannon (James Shannon, Glaxo’s chief medical officer) that the process doesn’t allow them to provide the kind of data we’re requesting.

Pharmalot: And what does he say?

Juredini: In the last letter, he says he’d like to have a phone call about that process and how it might change. I said I‘d like to do this by email, but haven’t heard back. But they have made clear it’s not the kind of information they’ll make available, even though they’re advertising it as open disclosure. It falls short of that.

I’ve done what they’ve requested, which is to put the application on the (new) web site (for data disclosure)…  And we’ll do what they request, which is to apply their analytical plan to the data, but we want to go a step further and ensure what’s in the data taggles collected by GSK reflects what was reported by patients to the (Paxil) researchers.

Pharmalot: When did this last exchange occur?

Jureidini: On November 5, he suggested a phone call. On November 8, I wrote back saying thanking him and didn’t think a phone conversation was best way to proceed.

Pharmalot: Why insist on e-mail and not a phone call?

Jureidini: Well, for one thing, I’m representing a team.  I’m not an individual functioning alone here. And it’s also partly because part of this process is to document interaction with GSK. Another reason is that I expect that it’s likely Shannon is a more skilled communicator than I am, so it’s more neutral, I think, to communicate in writing.

Unless they can come up with some plausible reason why there are genuine ethical concerns about providing the data, which may satisfy an independent appraisal, then you’d have to suggest there are other reasons why they don’t want us to see it. But we should be clear that GSK hasn’t said no to this request and I don’t want them backed into a corner of being acused of not providing information. I‘m not accusing them of that at all.

But the ability is there. It’s really easy. It’s a small expense to render documents confidential. It’s a question of whether there are more bad things happening that needn’t be hidden. It would be they had something to hide.

Pharmalot: Beyond this one study, what do you expect to accomplish?

Jureidini: There’s some good that can be done here. We can bring to account different parties that failed the community in the process of the publication of the (original) article – the company, the key opinion leaders… And we’re hoping that other people will be interested to make (the same sort of request). It’s an onerous task, but if you form a group, it can be done in the margins of your ordinary working life.