Tracking down a source.

Maybe you saw the story in the New York Times about new research that may show that ingesting too much caffeine while pregnant increases the chances of miscarriage. And, if you're like me, one of the first things you did was try to track down the actual research paper discussed in the newspaper article.

If so, I hope you've had better luck than I have.

The New York Times article (dated January 20 -- that was a Sunday) describes the research as "to be published Monday in the American Journal of Obstetrics & Gynecology" and identifies "the lead author of the study" as De-Kun Li. Writing about the same study (and quoting Dr. De-Kun Li), MedPage Today gives the citation for the research as Weng X, et al. "Maternal caffeine consumption during pregnancy and the risk of miscarriage: a prospective cohort study" Am J Obstet Gynecol 2008; DOI: 10.1016/j.ajog.2007.10.803. (Presumably that means that the "lead" author is not, in this instance, the first author.)

But here's where things get really interesting. The January issue of the American Journal of Obstetrics & Gynecology seems not to include this article -- or, indeed, any article about caffeine and miscarriage. And no such article appears in that journal's listing of articles "in press" (i.e., articles that have been accepted for publication by the journal).

So ... does this article that the New York Times and MedPage Today cited actually exist? Is it actually published, or at least accepted for publication (and if so, where)? And how exactly, given this particular trail of breadcrumbs, is the curious reader supposed to get her hands on the peer reviewed research to see what it actually says?

Because if the science writer can't get straight on the details of when and where the research is published, why should I trust the science writer's analysis of what the research shows?

More like this

by Elizabeth Grossman Nurses face many hazards on the job, and one that clearly demands more detailed analysis than it's received to date is the effect of occupational chemical exposures on nurses' reproductive health. A recent study by researchers at the National Institute of Occupational Safety…
Photo credit: Robert Hanashiro, USA TODAY It is the norm today to discuss race as a social construct. Less fashionable is it to explore race as a biological concept. When there's no up or downside and the discussion is abstract I think most people can get away with benign neglect in regards to…
Antivaccinationists just love the Vaccine Adverse Event Reporting (VAERS) database. As in love it to death. As in "can't get enough of it." The reason, of course, is that VAERS is a lot like an unmoderated discussion forum or, at best, a minimally moderated forum. Anyone can say anything they like…
This CNN Health article provides some useful recommendations on how to lower one's risk of miscarriage. I have summarized them below. 1. Do a little detective work When you're planning to get pregnant, your first move should be a careful pre-pregnancy checkup to reveal potential risk factors like…

Could the reporter have taken someone's word for it that the paper would be published?

I saw this story on the Today Show on Monday, I think. I had some doubts about it based purely on the facts that it is one study (if, indeed, it is actually one study), and that it was being reported in the general news media before publication in a peer-reviewed journal.

You're totally right, it's always an eyebrow raiser when you can't find evidence of the article on the journal's web site.

If it helps, here's the press release that announced the find. If you're extra curious, you can call or email the press officer to figure out exactly which issue the purported paper is meant to appear in.

I agree with you, this is a tough one to track down, but an awful lot of main stream news orgs are reporting it.

The only potentially useful piece of information I did uncover is that it came out of a research study done by Kaiser Pemanente Div. of Research in Oakland, CA. The website of that org (contact us) gives a phone number (510) 891-3400 and has an web form to generate email from comments. I did not follow further to try to determine the actual recipient email address.

You're on.

By Super Sally (not verified) on 23 Jan 2008 #permalink

Dr Li's research web page includes studies of pregnancy, but a search for "caffeine" gave me nothing.
His home page:

None of his selected papers listed there have an "X. Wang" as a coauthor. Perhaps a shy grad student? A copy of the press release would help...
That is here:…
There are severla quotes in the press release, and I can't tell if the coverage simply lifeted those or if the reporters actually spoke to Dr. Li.
Xiaoping Weng is listed as being at the same institute in the press release, but searching for his names does not reveal a web page for him.

That's all I got.

I heard it reviewed on NPR earlier this week. They made the point that this was a retrospective study and that many of the women interviewed had already miscarried (and thus the experts they were interviewing were skeptical of the result). Nonetheless, I also read that the March of Dimes is changing its recommendation about allowable caffeine consumption based on this report.

None of their citations are newer than 04 though.

I found this reference to the study at the tail end of (UK) NHS news report on same:

Weng X, Odouli R, Li D-K. Maternal caffeine consumption during pregnancy and the risk of miscarriage: a prospective cohort study. Am J Obstet Gynecol 2008; Jan 21 [Epub ahead of print]

Best case scenario: the NYT's reporter, Denise Grady incorrectly stated the study would be published Monday. Still can't find it online, though. So am not sure what Epub ahead of print actually means.

This makes me quite uncomfortable. I don't like the idea of the media being able to pick up a story (and this one is all over the place this week) before the research is properly published and peer reviewed. I also intensely dislike the notion that such research will never get a fair and critical appraisal before recommendations are changed. If it is a retrospective study of women who have already miscarried then it is even more troubling.
I think the bottom line is publicity for publications should occur after formal publication, not before.

I haven't come up with anything yet. Pubmed comes up empty, although it finds several articles from De-Kun Li on pregnancy issues over the past couple of years, so it is at least plausible that this article exists. Based on the DOI number I initially thought this was an old article that was resurfacing somehow; obviously that is not the case. Judging from my search results it does not appear that Am. J. Obstet. Gynecol. has ever published an article on this subject, or that it has ever published an article by De-Kun Li.

However, I did find that an article on this subject was published in Epidemiology on 19 Jan 2008. Curiously, this study appears to have come to the opposite conclusion from whatever Li may or may not have published. This will hopefully link you to that article.

I will try to remember to look for this tomorrow at work when I have more resources available. The most likely explanation is that there is a mix-up of some kind: the phrase "appearing in the online edition" suggests that the article was supposed to appear in the advance publication page on Monday and may not have for some reason... perhaps because of the publicity?

I wouldn't be surprised if this story were connected to someone playing the coffee commodity market or shorting Starbucks stock.

This sort of thing happened to me a few months ago with a physics story. Wasn't on ArXiv, wasn't on the journal page as forthcoming or whatever, wasn't on the researcher's page... so I e-mailed the journalist and he actually e-mailed me a copy of the article (a surprise). In that case the article was accepted but it didn't even get to the forthcoming papers for a month after the story and I don't know when it ever actually was in print. I sure hope medicine is faster than information science where it could be "accepted" for 9months to a year prior to "in press"!

It is fairly common for reporters to get access to articles after they're accepted but before they're published, with the agreement that no stories get printed until the article publication date (also known as the embargo date). Science, Nature, and PNAS all do this; it's possible AJOG does as well. The idea is to give journalists time to report and write a balanced, accurate story rather than racing to be first. Whether this works in practice is frequently a matter of debate among science journalists.

Also, many journals take a very hard line on researchers talking to the press while an article is being reviewed and sometimes even at the accepted-but-not-published stage unless there's an embargo agreement in place. Scientists run the risk of the journal pulling the article from publication. Anything said at a conference in a talk or poster is fair game, although scientists with a submitted or in press paper will still sometimes be skittish about answering questions.

My best guess here is that AJOG editors thought the paper would be up on the web on Monday but for some reason it was delayed. I had that happen with a story last summer about an ophthalmology paper--the journal editorial office thought the paper would be out "that week" but the publisher (Elsevier, same as AJOG) didn't get it up on the web for another week or two.

A number of medical journals publish online well ahead of printing. I had the impression that "Epub ahead of print" was fairly common, but perhaps not. Those items should be available online to journal subscribers -- and often to library patrons if their library subscribes.

I don't see any mention of "embargo" in the comments so far. Is that another quirk of medical publishing? (Let's see if I get this link past the spam filter:

Note that the MSM's lack of concern about accuracy is not restricted to matters scientific. Once you look into most any report you'll find errors, mis-attributions, and deliberate falsehoods galore. It's why a lot of people started blogging. Mainstream media is run by and for hardcore bigots convinced you and I are worthless scum.

Yes, but Epub ahead of print means it's, er, *been* published. Online. At this very moment there's a link to the epub on the front page of Am JOG (email me if you want but can't get a copy). It now says it will come out in the March issue. Also it still doesn't exist on PubMed, which is as good as not existing to the rest of the biomedical community really.

Personally, I am much more inclined to belive Motherisk's data; they have citations and everything. Of course the methodology of the studies they review could still be flawed, but that's another problem altogether.

I wouldn't be surprised if this story were connected to someone playing the coffee commodity market or shorting Starbucks stock.

Posted by: PengieP | January 23, 2008 9:14 PM

As skeptical as it makes me sound I wouldn't be surprised if you are right! LOL! There is no article and as soon as they close the deal on Wall Street the spin masters get everyone's attention on something else right away!
Dave Briggs :~)

Is this sinister, or just a screw up?

By Jeff Chamberlain (not verified) on 24 Jan 2008 #permalink

They just jumped the gun. It's there now. The DOI doesn't resolve but it's there under "in press" as a corrected proof:
Maternal caffeine consumption during pregnancy and the risk of miscarriage: a prospective cohort study
Presented at the 19th Annual Meeting of the Society for Pediatric and Perinatal Epidemiology, Seattle, WA, June 20-21, 2006.
Xiaoping Weng PhD, Roxana Odouli MSPH and De-Kun Li MD, PhD
Received 2 April 2007; revised 20 July 2007; accepted 13 October 2007. Available online 25 January 2008.
(gee, look at how long it took to get online)

Since Jenny F has pointed to the Motherisk's "Data" ( with "citations and everything") I believe it is worth noting a few things.

Dr. Gideon Koren is the founder and director and spokesperson in chief for the Motherisk Program.

He has a history of research misconduct, bullying of colleagues and telling of lies. See for example

The Scientific Misconduct Blog



Several other aspects of misconduct remain unexamined (apparently) by the University of Toronto.

Koren is also Chair of the Publication Committee of the Canadian Society for Clinical Pharmacology which Publishes the Canadian Journal of Clinical Pharmacology.

Despite this scientific position, he feels able to coauthor multiple byte sized review articles in his own Journal (See this month's issue for several) with "Naturopath" coauthors who also write this sort of gobbledegook bilge about "adaptogens" (

Apparently without batting an eyelid.

You can't be too careful when thinking about the origins of "data".