Respectful Insolence

Blogging on Peer-Reviewed ResearchThe other village quack of the Chicago Tribune has decided to enter the breast cancer fray again.

No, I’m not talking about the main village quack of the Chicago Tribune. That would be Julie Deardorff. Rather, I’m talking about the Chicago Tribune‘s newly minted breast cancer crank, Dennis Byrne. We’ve met him before, parroting credulously an incredibly bad study claiming that it had found a slam-dunk association between abortion and breast cancer. How bad was the study? Well, it was so bad that it was published in that bastion of politically-motivated pseudoscience, the Journal of American Physicians and Surgeons. I had a fine old time dismantling the study that Byrne found so convincing, as did Mark Chu-Carroll, Mark Hoofnagle, and the Arch Pundit. Clearly, that Byrne found that study so convincing tells me right off the bat that he has no clue when it comes to evaluating scientific studies, and he proceeds to prove me right in a spectacular fashion, as he cherry-picks a study that purports to show that oral contraceptive pills increase the risk of premenopausal breast cancer. But first, a little playing the martyr always comes in handy:

Not wanting to become known as the town quack, I am reluctant to write another politically incorrect column about breast cancer.

Four weeks ago, when I reported a study that found a statistical link between abortion and breast cancer, the hate e-mail poured in, denouncing me for being an ignorant, stupid, anti-science, anti-choice and anti-woman lunatic.

Based on his previous article, I agree that Byrne was ignorant, stupid, and anti-science. I have no idea if he is anti-choice (although his falling for such a crappy study that claims to show that abortion causes breast cancer is highly suggestive that he is is likely to be) or anti-woman. I also doubt that he’s a lunatic, but he is definitely in over his head when it comes to evaluating any sort of scientific study. Of course, that doesn’t stop him for more than a second. Perhaps chastened a bit but still defiant, Byrne marches right back into the fray:

Yes, I know, this debate has been going on for years, if not decades, and judging by the last studies given wide exposure a few years ago by the media, the issue seems settled: Oral contraception does not significantly increase the risk of breast cancer.

There’s just one problem. According to an analysis in one of the most credible peer-reviewed journals in the country, the Mayo Clinic Proceedings, the risk is real. The study employed an often-used medical research technique called “meta-analysis” that allows researchers to combine data from other studies on the risk to get a larger picture. The result: Premenopausal women who used oral contraceptives prior to having their first child have a 44 percent higher chance of getting cancer than women who didn’t use the pill. If they used the pill for more than four years prior to their first full-term pregnancy, the risk increased 52 percent. Chris Kahlenborn, an internist at the Altoona (Pa.) Hospital and the study’s lead author, suggests one additional woman in 200 could get breast cancer. Extrapolated throughout the population, that could mean thousands more cases every year. I’d say that’s an important story.

Of course, I can’t resist pointing out that the Mayo Clinic Proceedings is hardly “one of the most credible peer-reviewed journals in the country.” It’s an OK journal, but certainly nowhere near top-tier. In fact, it’s probably at best in the middle third and at worst in the lower third of journals. Certainly its impact factor is nothing to be that impressed about. We’ll take a look at the study itself in a moment, but first let’s hear Byrne whine about the conspiracy of silence the mainstream media supposedly has about this story:

The reaction? Nearly total silence. Since it was published more than a year ago, I couldn’t find a single reference to it in the archives of the New York Times, Washington Post, Los Angeles Times or this paper. The Associated Press appears not to have covered it. I couldn’t find a single mainstream media article about it in a Google search. But stories about other breast cancer risks were plentiful, including one about how sleeping with a night light on can increase your chances of getting breast cancer. The National Institute of Cancer doesn’t mention the study on its Web site, but it did detail a 5-year-old study claiming to find no higher risk to pill use. The American Cancer Society also doesn’t mention the study and concedes only that “it is still not clear what part” the pill plays in breast cancer. Such guidance, if not deceptive, is certainly incomplete.

Oh, no! Those dastardly journalists! In their leftist, ideologically driven desire to abort fetuses and encourage promiscuous sex, the better to enslave women in their horrific anti-life and anti-Jesus ideology, once again they’re ignoring evidence of how deadly the pill supposedly is to women, as they did a couple of months ago with the dreaded abortion-breast cancer link! Or so Byrne would have you believe. It couldn’t, perhaps, be that this study didn’t cause much of a stir because it’s rather unimpressive, could it? Of course not. It must be a ideological blindness.

So let’s look at the study itself, entitled, boringly enough, Oral contraceptive use as a risk factor for premenopausal breast cancer: a meta-analysis. Yes, it’s a meta-analysis. To me that makes me rather suspicious right off the bat. Over the years, I’ve become more and more suspicious of meta-analyses. The main reason is that they are totally dependent upon the studies chosen to analyze and are an excellent example of “Garbage In, Garbage Out.” Not that the studies that were chosen for this meta-analysis were “garbage,” but perusing the list of them does reveal a couple of peculiar things. First, Kahlenborn includes studies dating back to 1980, which is curious. Women likely to have been developing breast cancer in the 1980s were also likely to have taken a version of the pill in the 1960s and 1970s that contained considerably more hormone in it than today’s versions of the pill. Another peculiarity is that there were few studies after 2000 included. Kahlenborn states that this was chosen to allow comparison between earlier studies, where few women took the pill before having their first child to later studies, where more and more women take the pill for a period of time before having their first children. In this, he reveals a major potential confounding factor that wasn’t even discussed in the study, namely controlling for number of children delivered and age at first childbirth. It has been well established that giving birth to more children decreases the risk of breast cancer, as does having a child at a younger age. Hmmm. You don’t think that maybe, just maybe, women who take the pill at a younger age, before having children, may be delaying having children, do you? Even Kahlenborn has to concede: “Hence,we could not adjust for potential confounders, such as age at menarche and age at first birth.” He goes on to handwave a bit about why he thinks this isn’t a problem, but he’s not convincing.

In other words, this study is nothing new and, like Byrne, not that convincing. And don’t even get me started about how publication bias alone might have accounted for such a small increased risk in a meta-analysis.

Of course, if Byrne were truly interested in “telling women the truth” about the pill, he could have done a literature search. If he did, he would have found that there is a good evidence that the pill decreases the risk of ovarian and endometrial cancer and growing evidence that it may decrease the risk of colorectal cancer. These benefits have to be balanced against the known low risk of thrombotic events and, possibly, a slightly increased risk of breast cancer, if the risk even exists. In terms of cancer alone, even if the Kahlenborn study is 100% correct, one would have to balance that 1 extra case of breast cancer in 200 women with how many fewer women would get other cancers. But a careful discussion of the relative risks and benefits of the pill is not what Byrne is about. He’s about scaring women with the breast cancer bogeyman.

More amusing is that Byrne apparently didn’t even bother to look into Kahlenborn’s history at all. From Completely Unnecessary, I learn that Dr. Kahlenborn’s work is very much favored by One More Soul, a “non-profit organization dedicated to spreading the truth about the blessings of children and the harms of contraception,” where it is claimed that contraceptives destroy marriages, degrade women, and “treat children like a disease.” Indeed, Kahlenborn has even written a pamphlet How the Pill and Other Contraceptives Work for One More Soul. In fact, he’s even released a “revised and updated” edition after the Mayo Clinic Proceedings study, and the blurb states:

Newly revised and updated in September 2007. How the Pill increases the risk of breast cancer, and effective strategies for prevention. Women who take contraceptive pills before their first full pregnancy are 44% more likely to develop breast cancer before menopause. This is a real eye opener for people who thought the Pill was harmless.

You’d think that if Kahlenborn wasn’t a crank, he’d try to distance himself from these organizations or at the very least wouldn’t use his research to push an agenda, but instead he is using the legitimacy of the journal in which he managed to get his meta-analysis published to push his “the pill causes breast cancer” fear-mongering.

Most hilarious is this comment by Byrne near the end of his article:

As for me, I am not opposed to contraception, oral or otherwise. I am not plotting to get the pill banned. I am not writing this column for hidden religious reasons. I am not saying that the Kahlenborn study is the last word; I’m not a scientist, so I can’t vouch for its methodology or conclusions. Just like the abortion/breast cancer study, I’m writing about it because people have a right to know about the existence of health information, even if it is contradictory to the given wisdom.

Touchy, touchy, Dennis. Settle down. Take a deep breath.

I love the way he’s taking the “I’m just sayin’, you know” defense so beloved of the crank journalist, cloaking his agenda in “the public’s right to know.” Never mind that he’s openly advertising his ignorance of science, as was made mind-numbingly obvious by his credulous acceptance of the JPANDS abortion-breast cancer “study” about five or six weeks ago. Brave iconoclast that he thinks himself to be, he’s getting that information out there, inspiring ignorant and credulous bloggers to link to him and spread the misinformation. He does women no favors and certainly does not promote a sober and science-based discussion of the real risks and benefits of the pill; instead he prefers to scare them with likely unfounded fears of breast cancer and defend “chastity balls” against “vile displays of hatred against Christians.”

Contrary to his apparent vision of himself as the brave warrior against “conventional wisdom,” Dennis Byrne is merely parroting an unremarkable meta-analysis that probably overestimates the risk of breast cancer due to the pill (if there even is a risk) that was published in a lower tier journal by an ideologue with an agenda regarding the topic as evidence of “suppression” of the science allegedly showing a link. All the while, he’s ignoring the full, scientifically supported risk-benefit profile of the pill and soft-pedaling the ideological bias of the study’s lead author. I suppose Byrne’s stepping up in the world, though. At least this new “ignored” study that he is citing in his quest to convince women that abortion and birth control cause breast cancer appeared in a peer-reviewed journal, rather than a politically motivated propaganda rag like JPANDS. Too bad his understanding of science remains stuck on stupid. Indeed, his supposedly self-deprecatingly labeling himself the “village quack” is an insult to quacks everywhere.

REFERENCE:

Kahlenborn C, Modugno F, Potter DM, Severs WB. Oral contraceptive use as a risk factor for premenopausal breast cancer: a meta-analysis. Mayo Clin Proc. 2006 Oct;81(10):1290-302.

Comments

  1. #1 J-Dog
    December 5, 2007

    Thank you for taking the time to research and explain this.

  2. #2 Ed
    December 5, 2007

    Even if there is a significant link between the Pill and breast cancer, let’s not forget that

    a) the protective effect against ovarian and endometrial cancer is large – about 40-50%. That’s far more substantial than the small potential increased risks for breast cancer.

    b) most women take the Pill at an age where their absolute risk of breast cancer is very low (it’s a disease of old age after all).

    c) studies which find that the Pill affects the risk of breast cancer also show that the risk falls away to normal levels once women stop taking it. Which the vast majority of them do well before their absolute risk of breast cancer starts to climb.

    d) this large study concluded that “oral contraception was not associated with an overall increased risk of cancer; indeed it may even produce a net public health gain”

  3. #3 PalMD
    December 5, 2007

    You may wish to add a peer-review icon!
    Great analysis. Sounds like the guy is developing a whopping Galileo complex.

  4. #4 Janine
    December 5, 2007

    Dennis Byrne has been writing these type of columns for many years. And he is anti-choice. He was on the editorial board of the Chicago Sun-Times for many years before moving on to the Chicago Tribune.

    Nothing new for him, he has written about the supposed abortions linked to breast cancer for years.

  5. #5 Rev. BigDumbChimp
    December 5, 2007

    Our wonderful President who always makes great choices in his appointments, appointed someone who worked for a “women’s center” (a complete misnomer) who in their literature used the abortion – breast cancer “link” to scare women into certain choices to head the Dept. of Health and Human Services family planning program.

    Brilliant

  6. #6 Rev. BigDumbChimp
    December 5, 2007

    Yikes. Where are the grammar police when you need them?

  7. #7 Bob O'H
    December 5, 2007

    I’ve just drawn a funnel plot of the data. Someone fancy writing an article about publication bias? :-)

    Bob

  8. #8 ArchPundit
    December 5, 2007

    Another aspect of Meta-Analysis (which I’m not a fan of, but I’ll save that) is that it largely is premised on the case matches have to included relevant characteristics. If not, then you don’t have controls over other factors such as smoking. In the studies chosen, the only consistent thing about the case matching in each study was age. So smoking or other risk factors are not really controlled.

    But essentially this boils down to Dennis Byrne is a moron. Something we already knew.

  9. #9 vlad
    December 5, 2007

    “But essentially this boils down to Dennis Byrne is a moron.” Nope he’s a man with an agenda, nothing more and nothing less. When someone with no accountability has an agenda there are no limits.

    To Dennis:
    “How long will the industry’s political flunkies, who receive millions in campaign funds from this special interest, be allowed to turn a blind eye to a danger that kills thousands of women every year?”
    (http://www.right-to-life.org/news%20archives.htm)
    As opposed to the fundies who pay for your own idiots in congress? I can list plenty of government officials who are on the pro-life bank role, here’s just one with real numbers. http://classof2007.pbwiki.com/
    Senator+Allen’s+Campaign+Contributions

  10. #10 Bob O'H
    December 5, 2007

    Sorry, I’m looking at this in too much detail. But can anyone tell me how the authors of the meta-analysis got their OR for the under 35s part of the Wingo study (Fig. 1, second line)? They give an OR of 1.04, I get 0.11. Almost the same.

  11. #11 Monado
    December 5, 2007

    Abortion might have a small effect by preventing breast-feeding, which does give some protection against breast cancer. But not so much that you’d want a kid just for that reason.

    For the other side of “the harms of contraception,” take a look at “Confessions of an abortionist (exerpts),” the online biography of a doctor who did abortions in the early 1900s. Somewhere along the way he estimates that there are five abortions for every live birth and he describes what people use to try to end pregnancies that they can’t afford: bicycle spokes, umbrellas, bicycle pumps, knitting needles…

  12. #12 Monado
    December 5, 2007

    I also wish to point out an article on Daily Kos called, “Why I Provide Abortions.”

    … When told that she – already unable to adequately feed and clothe her family – was again pregnant, she looked up at me and the resident. There we stood, two white males, well clothed, well feed young men with superior educations. We were, in her eyes, stunningly blessed and obviously going places in the world. She began to weep silently. She must have assumed, for good reason, that there was no way that we would understand her problems; she knew also that there was nothing that we could or would do to relieve her lacerating misery.

    “Oh God, doctor,” she said quietly, “I was hoping it was cancer.”

  13. #13 jen_m
    December 5, 2007

    Bob – I get the same answer, but the OR in the third line is consistent. I think there must be a typo in their raw numbers, maybe? If you use “1117″ instead of “117″ for the case OC users, it comes out right, and the order of magnitude for the under-35 cases overall is more consistent with the reporting of Wingo’s under-35s in the next few tables.

    It’s really pretty decent, as meta-analyses go – I don’t know why they didn’t publish in a better journal.

  14. #14 Jesse
    December 5, 2007

    Thank you for the excellent fisking of Byrne’s woeful pile of dreck he so eagerly touts as an exposé of some merit.

  15. #15 Brie
    December 14, 2007

    Hey there,

    Your trackback got hit by my spamblocker, so I didn’t get to read this until just now. I’m not sure when you wrote this excellent analysis, but you may have noticed that Byrne came back at me, asking me to address the ‘substance’ of the study.

    I’m not a scientist, so I looked through his reference list, as it appears did you. Few studies after 2000, etc., etc.

    It really angers me that Byrne gets to trumpet this study as ‘revolutionary’ and put it out there for the public. Sigh, I guess I should write that letter to the Trib I was thinking about.

    Cheers,
    Brie (Completely Unnecessary).

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.