Tuesday, July 13th, 2010

When Science Stories Go Crazy: "Desperate Addicts Inject Others’ Blood"!

THEY BLINDED US WITH SCIENCEIt is very easy to write a bad science story. It is tantalizingly easy.¹ Often that is because writing a nuanced story requires a level of scientific explanation that will far exceed a reader's patience and a newspaper's wordcount. It's also easy because reporters let those folks who are stuck churning out releases for journals and universities do the hard work. There are whole sites devoted to these releases and they are superb. There are unintentionally hilarious headlines ("Double-teaming a whole-genome hunt") and more quantum-physics-related superlatives than you can shake a stick at. They are also full of puns ("Engineering could give reconstructive surgery a facelift"), so there's that too. It's the job of the press release writer to suck the reporter in-and it's the job of the reporter to be mildly entertained by the puns (and impressed by the ability of the writer to summarize an excessively complex story) and then factcheck it, gather some reaction quotes and send it off to the boss. Problem is, it's tricky.

Press releases are misleading-you all know that. Studies are half-baked, or preliminary, or contradictory. Findings are not quite so pat when you start looking at the numbers involved.

What newspaper wants a story titled: "An Alkaloid, That Happens To Also Be In Wine, And Also In Some Less Interesting Substances, Has Been Found To Slow Tumors in Mice When Injected In Extremely-High Dosages? No one wants that story. I almost died of boredom just writing that sentence. We want "Wine Cures Cancer" or, at least, a nice coverline like "Weasels Ripped My Flesh."

And then there's the cruel mistress of needing to get it out before xyz.com does and none of the paper's 50 gazillion authors will get back to you (why are researchers so hard to get on the phone!) and, oh god, alexa.com! So yes, the job of the science writer can be awful.

But sometimes you get something worse than a bad science story. Sometimes you get a terrible, blood-curdlingly wrong science story, a whirling vortex pulling matter in and spewing blackness out.

Gosh, do I hate these.

SARS was a golden era of bad science. Research puts Sars mortality rate at 20%! And: Can SARS Be Stopped? Experts Differ, but Fear a Third-World Epidemic!

How do you like this one? "Pets at risk as bird flu kills cat"! Yes: one cat. "How to tell if your cat is infected"! That was because scientists infected some cats with bird flu to see how they'd react. Don't worry, they didn't lay their hands on your cat.

HIV provides an incredible opportunity for bad science stories. We'd all like to pretend that the "new" "killer" HIV strain media panic of 2005 never happened-but it did.

And those stories go on. Today, there's this, in the Times: Desperate Addicts Inject Others' Blood. Alarming, right?

There are some simple elisions performed in the name of nice-sounding-copy. "Desperate heroin users in a few African cities have begun engaging in a practice that is so dangerous it is almost unthinkable…" has a nicer ring than: "A handful of desperate heroin users in one city and one semi-autonomous region in Tanzania were recorded five years ago, and then three years ago, engaging in a practice that is so dangerous it is almost unthinkable…" Don't get me wrong! I appreciate language, and wanting to write beautiful sentences. I really do. At some point, however, the desire to convey information with some impact to readers becomes tabloid and incorrect.

That also raises a question. If this is really such a shocking deal, why wasn't it covered it in 2005 when it could possibly be said that "begun" might have applied (as in, had begun to be noticed by researchers)?

Then there's a lack of respect for the readers' ability to discern the holes. With this headline atop it, it's a deliciously scandalous article; down low, they can do what they like.

Lets just say, as a newly made up rule-of-life, that if you have to say more than twice in a story shorter than one thousand words that "it isn't common," you can't pretend said subject is a the Big Problem of the day. Since this story was written prior to my made-up rule, we'll let it slide. But, even without the rule, it is simply not cool to take such lengths to imply otherwise.

This article quotes two small studies done in two cities; each study worked with fewer than 200 people. "About 9 percent of the 200 drug-injectors interviewed practiced" injecting each other with blood. Here's a point for admitting the pool was 200 people, which will now be deducted because-18 people, seriously?

Meanwhile, the main study that our trusty reporter is relying on–isn't associated with any numbers, which makes me wary. That study is, as it turns out, more focused on the demographics of these "flashblooders" (studying what percent were abused as children or young drug users) and doesn't mention a ratio of flashblooders to heroin users (at least, not in the abstract, which is all I, unwilling to shell out $52 to buy the paper, have access to).

And then: two successive paragraphs starting with "there have been reports [related to flashblooding]" ending with "not confirmed by medical researchers."

The expert quote and kicker of the story is thanks to a "horrified" blood bank president-elect who has never heard of the idea.

Most of the story, at least the bulk of the story which is backed up by facts and figures (how dull), is actually not at all about flashblood but about the rates of HIV/AIDS among heroin users. Unmentioned: are "flashblooders" people who always or regularly share needles already… making this a somewhat nonstory? Personally, I think "Increasing use of heroin in parts of Africa has the potential to magnify the AIDS epidemic" might stand on its own as a story.

What really tears me up about this type of article is that in the 24 hours or so between when the article was posted and now, the number #flashblood tweets jumped from two to about 80. Meanwhile, maybe you've already seen the link on peoples' Facebook pages and, since this appears in this today's paper, the other dailies will latch on, CNN will do a take, the evening news might try and localize it, and on, and on.

Meanwhile, in Dar es Salaam and in Zanzibar (also, not to put too fine a point on it, but maybe in Baltimore and Panama City and London!) drug users will still be contracting HIV through shared needles and sex workers through unprotected sex. Maybe this article will inspire some much needed funding. Probably not. To be honest, I suspect that by focusing on such a tiny-but-sensational vector of disease transmission, this kind of article puts Africans and drug users in a box labeled "cannot be helped."

Anyway, remind me to tell you about the Cambodian cholera brouhaha of February 2010 someday. Happy flashbloodin' y'all.

¹ And it is just plain tantalizing to run a bad science story because you can run them under fantastic headlines which will be picked up by myriad websites sending person after person to yours and they click and click and click on those related stories links and on those ads running alongside and. Excuse me. I seem to have soiled myself.

Abby Seiff is perhaps too easily angered.

Photo by James Vaughan from Flickr.

19 Comments / Post A Comment

deepomega (#1,720)

Way to bury the Weasel-related lede.

(My favorite remains "Chewed to Bits by Giant Turtles")

Mine is "I Battled a Giant Otter"

KarenUhOh (#19)

"Famed Psychic's Head Explodes!"–
Weekly World News

Mount_Prion (#290)

So that's where Frank Zappa got that song title from.

And yes, I read the rest of the post after.

MollyculeTheory (#4,519)

Tip: Scientists respond to emails about 87 times more often than phone calls. The corresponding author gets 9 billion more important emails a day, the extent of the 48 gazillion other authors' knowledge is where in the freezer one thing is. Stalk the first author; if the paper is not in a Top 10 journal he/she will be so gobsmacked that anyone actually read the goddamned thing you will trick him/her into responding helpfully.

Multiphasic (#411)

Or stalk the second author if you want the person who actually did the work and understands the implications of the study, just saying.

Or if you really want an attentive interviewee, go for the back-of-the-list folks without a doctorate. They're either grad students, in which case desperate for any sort of career validation, or else they're the statistician, who is undoubtedly desperate for any sort of human contact at all.

MollyculeTheory (#4,519)

Mm, does this depend on whether it's an epidemiological study vs. bench science? In my experience (with the latter), ability to explain the work decays exponentially with authorship.

Multiphasic (#411)

Oh, I'm just thinking about getting any goddam body at all to talk to you, whether or not they're actually useful.

I don't know if bench scientists tend to be less sanctimonious nitwits, but in much of the medical research I've covered, the address-all-questions-to author was almost always the second one. The lead author was generally the one with the shiny brass plaque who lent everyone else a laptop and three rabbits.

Multiphasic (#411)

Can I maybe not have said "sanctimonius nitwits" in an internet trail that can eventually lead to my real name? No?

Well crap. Intended to be a satirical generalization, not an actual assessment of the true function of lead author as figurehead/imprimatur, nor a judgment passed on any actual individuals, except the ones who were mean to me when I was trying to do some reporting.

MollyculeTheory (#4,519)

I doubt it's a matter of fewer s—– n–s; just that in bench science, the beplaqued rabbit distributor is generally the *last* author.

(Ugh that insult redaction looks so much worse than it is.)

Multiphasic (#411)

What's wrong with a splendid ninja?

Anyway, I've found first authors useless more often than not, though you're right that last authors are likewise often on there based on impressive overall beard growth more than anything. Our cite/source list actually tracked first, corresponding, middle, and last authors in separate columns.

Sean Jordan (#734)

Milk-blood to keep from running out.

Art Yucko (#1,321)


MikeBarthel (#1,884)

Thanks for writing this. Yeah, it's an interesting phenomenon, and I can see why she wanted to get the finding out there – and I can also see why she would have a hard time getting IRB approval to do a wider study – but its existence is no indication of how widespread it is. The sample size is too small to generalize from, there was no random assignment, etc. It's sorta anecdotal evidence but with charts. Again, no harshing on the researcher, who was the one who actually talked with a bunch of drug users in Tanzania, but I'm not entirely sure we can blame this one entirely on the NYT. The article itself, which is titled "Flashblood: blood sharing among female injecting drug users in Tanzania," is a little sensationalistic too.

Multiphasic (#411)

I remember reading an entire paper on HCV transmission in Lagos, examining cases of dentists contracting it from being bitten during examinations. Six pages in, it acknowledged the possibility that all of these transmission tales were anecdotal.

It's amazing at how bad the press releases can be sometimes given that the University staff typically work directly with the researchers to craft it. More typically it's a giant game of telephone – the press release summarizes & simplifies but generally gets things right, then a harried reporter rewords without fact checking the new copy, editor makes further changes and adds click-bait headline, hilarity ensues.

La Cieca (#1,110)

I, for one, am eagerly awaiting the ripped-from-the-headlines summer blockbuster film about a girl who's a Pittsburgh steel-mill welder by day, and vampire by night.

gregorg (#30)

I long ago gave up trying to keep up with the outrageous press releases for parenting and kid-related studies, and put them all into one weekly blogpost, the Friday Freakout.

Which turns out to be completely out of sync with the Science PR Industrial Complex, which sets all their embargoes to expire on Tuesday. Ahem.

Annie K. (#3,563)

All this — the exactly-right post and comments — is why I don't write about biomed stuff. Why do those biomeds do these things? All I have to deal with is NASA's tendencies toward premature press releases about alien planets, plus the occasional physicist with a cloak of invisibility, so not life-threatening.

Post a Comment