Neuron Culture

[Note: An update from 25 Sept 09 is at bottom]

Here’s a sad mess. It seems a
potentially important finding — that getting a seasonal flu shot might
increase risk of contracting the swine flu — is being sat on by a
journal, with the authors forbidden from talking about it, until they
get through the slo-mo publishing process.

The finding may or
may not be accurate. But as it regards an important issue, it needs to
be vetted and discussed openly, with the data at hand, as soon as
possible. But it’s not.

This is a tricky situation, to be
sure. There are, at least theoretically, both good and bad reasons to
withhold this information. But it seems to me the bad outweigh the good
here, and both science and the public interest would benefit from a
full airing of this study asap. At minimum, the journal holding the paper (which journal’s id we don’t even know) should state its reasons for doing so. I hope they’re good ones.

The good argument for
withholding this information is that the journal’s peer reviewers need
time to finish their review. Okay as far as it goes — only in this case, I don’t think it goes very far. For many subjects, three peer reviewers is about all you’re going to muster to vet a paper. But in a case this
urgent and high-profile, you’d get a much more thorough airing if you released the thing, for  scores of epidemiologists — whole armies of them, actually, all over the world — would be eager to help out. Unless there’s something we’re missing here, it woulud seem much more helpful to let these
findings get vetted out in the open, by the entire epidemiological
community, by publishing the study and the underlying data online asap so that
a little open science can be done. Let the hivemind at it.

Instead, we get rumors of an unsettling finding — and a process that just flat looks bad.

And this gets us to the bad reasons for sitting on a paper like this: the proprietary embargo disguised as careful review.  As a
journalist, I run into this all the time. Findings — non-peer-reviewed so far,
yes, but still of interest and often based on substantial work, and
very often already aired in public at conferences and meeting — can’t be
discussed “on the record” because … well, because the journal
wants to be the one to break the story. Particularly
upsetting is that the authors aren’t allowed to talk to the press about
their findings, having been forbidden to do so by the journals. This gag practice is
not so much in the interest of ensuring accuracy as it is of ensuring a
splash when the journal publishes the paper.

I find it hard to fault
the scientists in these cases; they know well that if they break a
journal’s gag order, their chances of publishing again at that journal
are quite slim. This is questionable enough in any case, and especially
given that science is ultimately a collaborative endeavor. But when
issues of public health are clearly and directly at stake, it’s
particularly hard to defend.

Here’s the thread; see what you make of it yourself.

Here’s the main news kernal, from the ever-alert ace flu reporter Helen Branswell:

September 24, 2009

The Canadian Press
Unpublished Canadian data are raising concerns about whether it’s a good idea to get a seasonal flu shot this season.

Drawn
from a series of studies from British Columbia, Quebec and Ontario, the
data appear to suggest that people who got a seasonal flu shot last
year are about twice as likely to catch swine flu as people who didn’t.


A
scientific paper has been submitted to a journal and the lead authors -
Dr. Danuta Skowronski of the British Columbia Centre for Disease
Control and Dr. Gaston De Serres of Laval University – won’t speak to
the media. Journals bar would-be authors from discussing their results
publicly before they go through peer review.


While
few people appear to have actually seen or read the study, the puzzling
findings have been a poorly kept secret and many in the public health
community in Canada have heard about them.


In
fact, concern about the unconfirmed findings is playing into calls from
Quebec and possibly other jurisdictions to delay or even cancel this
year’s seasonal flu shot campaigns across the country.


The
findings are causing consternation abroad as well, with officials at
public health agencies and even at the World Health Organization
worried the alleged link will deter people from getting vaccinated in a
fall when many people are being asked to get both seasonal and pandemic
flu shots.


The
Public Health Agency of Canada knows of the findings and has been
seeking help here and internationally to try to figure out if the
effect is real or if the studies are flawed.


“An
arms-length review of the various methods is currently underway to
assess the validity of the studies relative to that observation,” Dr.
David Butler-Jones, Canada’s chief public health officer, said by email.


“We
are also examining other data that will help to understand what if any
association there is. We look forward to the results of the review and
other data to inform our recommendations as we go forward.”


The
U.S. Centers for Disease Control also knows of the work. It said it has
looked for similar evidence in the United States but sees none.


“It
is difficult to speak about a study that has yet to be published,
however, as this is an important issue involving the subject of
seasonal influenza and the fast moving global pandemic of 2009 H1N1
influenza it is important to note the scientists at the Centers for
Disease Control and Prevention have not seen this effect in systems we
have reviewed in the United States,” spokesperson Joe Quimby said by
email.


A number of influenza and infectious diseases
experts know of but are unwilling to speak publicly about the paper.
But several were quick to note that British and Australian researchers
haven’t seen the phenomenon either. The lack of corroboration in other
jurisdictions is “a red flag,” said one expert, who does not believe
the findings are true.

This
shouldn’t be something that flu experts feel compelled to discuss sotto
voce. If the journal has good reasons to sit on the paper for now, it
should declare them. If not, it should get the paper out in the open so
the data and findings can be examined and vetted openly.

Update 9/25/09, 9:25 am EST: Two things:

  1. As per my comment below, I was glad to read this morning a thread at FluTrackers
    that the CDC and Health Canada will be discussing this whole flap this
    afternoon — apparently at 1:30 pm EST by the CDC and 2:00pm EST by
    Health Canada (though I can’t find those at the CDC and Health Canda
    sites!). Not clear whether they’ll make the paper public or whether the journal will step up.

  2. Helen Branswell reports this morning that the (rumored/unconfirmed) small f findings have led Canadian officials to decide to delay seasonal flu shots for most people until after the H1N1 shots become available in November, so that the latter, being given earlier, can work without any possibility of, ah, influence from the former, now to be given later.

Nobody said this would be simple. I’ll say this: I’m glad I’m writing about this and not calling the shots.

Comments

  1. #1 red rabbit
    September 24, 2009

    This was all over the news this morning, and I was hoping to get my hands on the actual paper before getting bombarded with questions. My office is a nightmare.

    Gah. Thanks a whole freaking bunch, CBC.

  2. #2 daedalus2u
    September 24, 2009

    This is the problem with for-profit journals putting their profits ahead of the public good. The journal better have a damn good reason for delaying publication or I suggest that scientists boycott the publisher and migrate to open access journals.

  3. #3 Alex
    September 24, 2009

    I am not a doctor, so assuming this was true, could somone explain to me why would the seasonal flu vaccine increase the risk of getting swine flu?

  4. #4 ERV
    September 24, 2009

    *GUESSING*

    This might have something to do with antigenic sin. So, say you got the seasonal flu vaccine last year. Maybe theres a little bit on it that your antibodies recognize– when antibodies recognize this same little bit on a seasonal flu virus, the viruses are neutralized, and YAY, you dont get the flu!

    Maybe swine flu has a bit on it that looks like the little bit on the seasonal flu vaccine, but antibodies to it are non-neutralizing. So your body is primed to make an antibody that doesnt work, and wont make antibodies that do work.

    This is a potential problem with all flu vaccines… and just getting the flu… Seasonal or swine.

  5. #5 Douglas McClean
    September 24, 2009

    Couldn’t such a correlation result from the right people (those at risk) having taken the vaccine last year?

    At risk for flu in 2008 -> Take 2008 flu vaccine -> (Still) at risk for flu in 2009 -> Get flu (with the novel H1N1 strain now predominating) in 2009

    Of course, I agree with the OP wholeheartedly that having access to the actual paper would be extremely helpful in sorting it out.

  6. #6 Theodore
    September 24, 2009

    This sounds closer to poor science and not just journalistic profit hoarding. It could also be the case of a few scientists taking advantage of the swine flu hype to convert otherwise poor epidemiology into a paper. Maybe those people who received the flu vaccine were more likely to get tested for swine flu when they had flu like symptoms? Or maybe those individuals who wanted to get a vaccine were more likely to approach the medical establishment when they had flu like symptoms? Maybe it is just bad luck. If there are people in the know holding this till they can do a few serum assays or a little bit of tissue culture work that would be a great call. And you couldn’t really announce that you were doing that without inadvertently giving a media sound-bite.

    If the press release said the following I would be a lot more worried.

    “Drawn from a series of studies from British Columbia, Quebec and Ontario, the data appear to suggest that people who got a seasonal flu shot last year have serum that cannot neutralize swine flu as effectively people who didn’t.”

    or

    “Drawn from a series of studies from British Columbia, Quebec and Ontario and controlling for a wide range of variables, the data appear to suggest that people who got a seasonal flu shot last year are about twice as likely to catch swine flu as people who didn’t.”

    Either way this gets read even despite the writers abundant qualifiers as flu vaccine causes swine flu.

  7. #7 bsci
    September 24, 2009

    The population that gets the seasonal flew vaccine is skewed towards vulnerable populations because public health policy and doctors push those people harder to get the vaccines. That population would obviously be more vulnerable to getting H1N1 too. This isn’t my research area, but, if I were a reviewer of that manuscript, I’d definitely be trying to confirm that the authors addressed the skewed population issue in some manner.

    I definitely think this should have been kept quieter until reviewers has a chance to comment and it was grossly irresponsible for the CBC to publicize this. If it turns out to be bad science, this will still be taking a life of it’s own now. That all said, those reviewers should consider evaluating this article as their top priority and the journal should release the article immediately if it deems it quality science.

  8. #8 Greg Laden
    September 24, 2009

    The article should have been sent to PLoS ONE to begin with.

  9. #9 Coturnix
    September 24, 2009

    Or even faster, to PLoS Currents – Influenza.

  10. #10 Mark B
    September 25, 2009

    Maybe people who get flu vaccines are then less fastidious about the hand-washing and staying away from people who are sick?

  11. #11 David Dobbs
    September 25, 2009

    Thanks to everyone for the extremely interesting comments. Many aspects, dimensions, crags, and crannies to this story. As some of you noted, it might be the science is just shoddy; then again, maybe not. And the findings may be just picking up selection biases of the sort mentioned by ERV or bsci above.

    But light is needed at this point, and I was glad to read this morning a thread at FluTrackers that the CDC and Health Canada will be discussing this whole flap this afternoon — apparently at 1:30 pm EST by the CDC and 2:00pm EST by Health Canada (though I can’t find those at the CDC and Health Canda sites!).

  12. #12 Rev Matt
    September 25, 2009

    The CDC webcast is at flu.gov/live. He’s in the Q&A segment as I type this and hasn’t addressed this issue at all yet.

  13. #13 Rev Matt
    September 25, 2009

    He did finally address “The Canada issue” twice, both times by saying “there’s no published studies indicating any problem so la la la la la.”

  14. #14 acı cehre
    September 25, 2009

    I definitely think this should have been kept quieter until reviewers has a chance to comment and it was grossly irresponsible for the CBC to publicize this.

  15. #15 acı cehre
    September 25, 2009

    I definitely think this should have been kept quieter until reviewers has a chance to comment and it was grossly irresponsible for the CBC to publicize this.

  16. #16 vpills
    September 25, 2009

    This sounds closer to poor science and not just journalistic profit hoarding.

  17. #17 Orac
    October 4, 2009

    God, I hate science by press release. It’s vile.

    The study hasn’t been published yet, and there is a way of vetting it. It’s called peer review. Moreover, it’s pretty standard practice at journals to tell authors not to discuss with the press a paper under peer review.

    In any case, this sounds like a load of crap. Whenever you see this sort of story, it’s always a good idea to be very, very skeptical of whether the science is any good or not. Usually it’s not. It was hugely irresponsible of the CBC to publicize this before peer review had had it’s chance to run its course.

  18. #18 Orac
    October 4, 2009

    Actually, I meant to say “science by press leak,” not science by press release. Which is way, way worse than science by press release.

  19. #19 David Dobbs
    October 5, 2009

    I’m not sure I’d fault the CBC here — or think it’s early to do so, at any rate. For one thing, to do so is to assume Helen Branswell is being successfully worked; possible, as we all have our off days, but unlikely.

    The bigger problem in this case seems to be the closed system, in which research on a subject important to public health is bottled up and, far as anyone can tell, not discussed openly even among the immunology community.

    That said, this is indeed complicated. As I noted before, there are good and bad reasons for holding onto this paper. The good is to have it evaluated quietly out of the spotlight. (Though as I’ve noted before, peer review is no guarantee this will happen.) That’s no longer possible. If they haven’t done so already, the journal editors should have this paper and its data being vetted by a wider community, including teams at CDC and Health Canada, or offer good reasons for not doing so. (Conceivably they’ve done so while I was tending the kids this weekend; if so … good.)