Talk:Harvard abortion study

Rarely have a seen a more blatant example of junk science. You can't exclude people who die of breast cancer, from a study on breast cancer! --Ed Poor 13:06, 24 April 2007 (EDT)

I'm sorry but wait just one cotton-picking horn-swaggling carn-sarnint (civility filter on) minute here, it doesnt say that the people who were excluded from followup as a result from death had died of breast cancer. I have a hard time believing they'd just ignore those deaths if they did occur, where on earth did you get tie impression that they did? Further, where the entry reads that it excludes women who got cancer afterwards, its excluding women who did and who didnt have abortions who got cancer afterwards because in those circumstances they're outside of the control of the experiment. You're misreporting out of context factoids that otherwise would show how stirctly they adhered to the scientific method in a pathetic attempt to show how the study ignored the most relevant data it was examining, as though the entire scientific community wouldnt have jumped on their necks already. No wonder why the CPedia front page gloats that "Only on Conservapedia" will use of the scientific method be deemed a sin.--RexMundane 13:45, 24 April 2007 (EDT)

1. It doesn't say that. It says Only on Conservapedia: this study debunked.
2. Thanks for turning on your civility filter: I read you loud and clear.

--Ed Poor 13:58, 24 April 2007 (EDT)

It seems that footnote 6 to the article argues that there is not a link between cancer and abortion.--1048247 13:52, 24 April 2007 (EDT)1048247

How about a link to the actual study so we can draw our own conclusions instead of reading Andy's slanted opinions? Jrssr5 13:08, 24 April 2007 (EDT)

Here's the NY Times's spin on it: [1] --Ed Poor 13:11, 24 April 2007 (EDT)

And here's an older NY Times story: [2] --Ed Poor 13:13, 24 April 2007 (EDT)

Can I point out, that although I think I have a scientific background, that I have no idea what:

"Among parous women [the overwhelming majority of the study], the [hazard ratios] HRs were 1.58 (95% CI, 1.13-2.20) for PR- breast cancer and 0.80 (95% CI, 0.60-1.05) for PR+ breast cancer (P for heterogeneity = 0.002) among women with induced a.."

actually means in real terms for women. If this is meant to be an informative article, than it fails because it makes the mistake of in attempting to be scientifically accurate, it actually presents scientific gobbledegook. The HR is presumably a hazard ratio. Does that mean that the increased risk represents a 1.58%, or 58% increase, and an increase on what? If the chance of developing breast cancer is 50%, then is the increase to 51.8% (hardly a significant increase) or to 76% (a pretty devastating significant). On the other hand if the risk is 0.05%, then and increase to 1.63% would be pretty significant, whereas one to 0.076% would not really be significant at all.

Please consider your audience. --CatWatcher 13:37, 24 April 2007 (EDT)

That part should probably be rewritten. Basically it means that the women in the study who had children (parous) were 58% more likely to get progesterone negative breast cancer if they ever had an abortion, and the same women were somewhat less likely to get progesterone positive breast cancer. If I did the math right there were 442 breast cancer cases among women who ever had an abortion. 59 (13%) were progresterone negative and 92 (21%) were progesterone positive. Murray 13:57, 24 April 2007 (EDT)

And according to [3] the prevalence among US women of breast cancer in general is about 1 in 8. Though to extrapolate the risk for PR- from the study (ie, what % of women who had abortions will get it by these figures) we'd need to know the rate of PR- breast cancer. Murray 13:59, 24 April 2007 (EDT)

That makes more sense. However, the point still remains (and it's not just here, it's generally in reporting of medical research). Whenever anyone says there is an incresed risk of X, if X is very low probability, then the incresed risk is negligible. It only matters if the original risk is significant, and the increase is significant. --CatWatcher 14:03, 24 April 2007 (EDT)

Thanks, Murray, your comments are enlightening. Catwatcher, realize that I'm quoting from the study and it is the fault of the authors for downplaying that fact and not explaining it better. But as Murray shows, these are significant percentages, not the unimportant ones that you suggest might be the case.--Aschlafly 14:05, 24 April 2007 (EDT)

In this case, I'm not playing devils advocate or anything. I am genuinely confused by this article. When I read the article as it stands at the moment, I cannot actually understand what it is saying, and whether there is an increased or a decreased risk of cancer if a woman has had an abortion. Perhaps I should come back and readi it tomorrow when all the argy-bargy has finished!--CatWatcher 14:10, 24 April 2007 (EDT)

Here's a summary of the findings: breast cancer in general (including all types) was not more likely in women who reported ever having an abortion. Women who reported miscarriages (spontaneous abortions) were somewhat less likely to get breast cancer, though the effect was fairly small and did not quite meet the criterion for statistical significance. Still speaking of all types of breast cancer, the findings were the same regardless of whether the women ever gave birth, or whether an abortion occurred before or after they gave birth. The hormones estrogen and progesterone are implicated in some breast cancers, so they also looked at these specific cancer types. Reported abortion had no effect on the likelihood of getting estrogen receptor positive or estrogen receptor negative (ER+ or ER-) breast cancer. Progresterone receptor positive (PR+) breast cancer was not related to abortion in women who never had kids, and was a little less likely in women who had kids and reported an abortion compared with those with kids and no abortion (that effect also was not quite statistically significant). PR- negative cancer, though, was 58% more likely in women who reported having an abortion if they also had children at some point. About 80% of the women in the study had children. Murray 15:40, 24 April 2007 (EDT)

Great analysis, but I do quibble with your statement about women who did not have any children. The ranges are so wide on the 95% CI that no conclusions can be drawn about them. The numbers are statistically insignificant, another sign of how the 100,000+ women in this study were simply too young for studying breast cancer.--Aschlafly 15:59, 24 April 2007 (EDT)

Andy - It was statistically not significant, though barely - p = .06 - if I understand correctly what you were referring to. The criterion for statistical significance is arbitrary and doesn't always reflect whether something is meaningful. On the other hand, statistical significance is dependent on the sample size, and with such a huge sample any effect that doesn't reach significance is bound to be small and likely not meaningful. So I can't argue with your quibble. I already have a copy of the article but I appreciate the offer. Murray 16:06, 24 April 2007 (EDT)

Opinions in article

this age disparity is analogous to an attempt to draw conclusions about heart disease by studying teenagers.

This statement is not a fact, however, it leads readers to believe it is. So I changed the sentenced and attributed the sentence to the person that first stated the sentence to make it a fact.

The analogy is factual. Don't make your change a third time. It's inappropriate to insert someone's user id. into an entry anyway. The page history has that info.--Aschlafly 14:02, 24 April 2007 (EDT)

Would like to start by pointing out one clear error for starters. At the start of the study, the women where between ages of 29-46 [4] So, how can one get the avrage age of the womens in the study to be 42 now 14 years later when the yongest women in the study should be atleast 43? Timppeli 14:03, 24 April 2007 (EDT)

By the way, I have a copy of the paper and can email to a few open-minded readers here who want to analyze its or my claims further.--Aschlafly 14:12, 24 April 2007 (EDT)

I would be interested in reading the study to draw my own opinions ... my email link should work. no guarantees as to how quick i'll get to reading the study though. Jrssr5 14:15, 24 April 2007 (EDT)

As no one seams to have an answer to the question i raised, im going to remove the mention about the average age of 42. Timppeli 15:07, 24 April 2007 (EDT)

I see my edit got reverted, but could i have some explanation why please, as it sounds kinda weird to claim the average age of the women in the reasearch to be 42 when youngest is 43. Timppeli 15:22, 24 April 2007 (EDT)

I thought I had explained this here, but now I can't see it. So maybe what I wrote was not saved. Your point is a good one.

The study ended in 2003, when the youngest would have been 39. The average of 42 applied to those who had an abortion, which means they were on average 12 years old when Roe v. Wade legalized abortion nationwide. It makes sense that the younger side of the population was more likely to have had an abortion, so an average of 42 is plausible.

Regardless, that is the age given by the study says and I've offered to email it to folks who want to analyze it. One took me up on my offer here.--Aschlafly 15:26, 24 April 2007 (EDT)

Thanks for the clarification. Timppeli 15:38, 24 April 2007 (EDT)

They did a pretty poor job of reporting the age - presumably the 42 reported is at the end of the study, but if they clarified that anywhere I can't find it. Murray 15:42, 24 April 2007 (EDT)

That's part of the spin, of course. If they reported that of 42 up front, then many informed readers would immediately react, "that's too young for most women to have developed breast cancer!" FYI, I found the age 42 in table 1. Murray, let me know if you need a copy. Your comments are good and I appreciate what you have to say.--Aschlafly 15:56, 24 April 2007 (EDT)

For the sake of comparison, the Reuters report on this (I haven't seen the original article so am not sure if this is accurate) said: "Michels' team noted that the studies that had seemed to show abortion caused breast cancer also mostly looked at younger women who had not reached menopause."

By the way Andy, thanks for your above comment. Much more data would probably be required for either of us to change our mind about the issue so that we agreed, but it's nice to discuss it calmly and rationally Murray 21:19, 24 April 2007 (EDT)

Full study text

Anyone wanting the study text, contact me. ColinR^talk 14:18, 24 April 2007 (EDT)

Hey, Colin could you please send me the full study text. --Staple 18:52, 24 April 2007 (EDT)

I fail to see why it would have been necessary to include African American or Hispanic women: their reproductive organs do not differ from that of white women. The only difference could thus come form lifestyle and environmental differences, but that was not what this study was about.

May I remind everyone here that an earlier study that seemed to show a correlation, was not ripped apart and scrutinized, in fact it was supported and cited from.

Middle Man

Ethnicity is a valid issue. Reproductive organs don't differ but genetics sometimes do. Given that breast cancer has a genetic component, it could be different for different ethnic groups (I have no idea whether it is). Murray 21:15, 24 April 2007 (EDT)

You would have a good point, if this was a study about inherited vulnerability to breast cancer. The existence of a special gene carried by African American or Hispanic women only, that would somehow increase vulnerability to breast cancer, only after an abortion, and not even in the case of a completed pregnancy, seems highly unlikely to me, although I do not have a Ph.D in genetics. In any case I do not believe this is sufficient to "debunk" this study.

Middle Man

I don't have a PhD in genetics either but that's not quite what I meant. It's possible that, for example, that the prevalence of genetic polymorphisms that are associated with breast cancer occur at different rates in different ethnic groups. At least some polymorphic genes are related to hormone function, and abortion alters hormone function relative to giving birth. So it's conceivable there could be a difference. Having said that, I wouldn't claim that this issue means the study is meaningless, and my reading of the research is that there is no link between abortion and breast cancer - my talk page has a list of studies that have found no link. Murray 11:04, 25 April 2007 (EDT)

The authors should have disclosed this weakness in their study. Breast cancer prevalence does have a strong ethnic component. Asians, for example, have breast cancer much less than whites. I bet a significant percentage (e.g., 5%) of the population in the Harvard study were Asian, but the authors concealed that.

Murray, I looked at your list on your talk page that purports to present studies deny an abortion-breast cancer link. How many of those "studies" have you read? Zero? Less than 10%? Don't you think you should disclose that to the reader???? One that I recognize I have read (Lancet), and it is neither a study nor a denial that abortion increases breast cancer.--Aschlafly 11:15, 25 April 2007 (EDT)

Hi Andy. I've read the methods and results sections of all of them at some point, if not the intros and discussions. I'm not sure what your issue with the Lancet article is. It is essentially a meta-analysis, which is a useful and novel contribution even if no new data were collected. For the sake of discussion I will post more details on each study when I have time - have already done some of that. Murray 12:03, 25 April 2007 (EDT)

Murray, you claim that all 24 articles are "Studies showing no abortion-breast cancer link." I have read 1 out 24, and your claim is completely false with respect to that article (Lancet). The Lancet article is not a study and it does not deny that abortion increases breast cancer risk. Your claim may also be false for your other 23 articles, which you cannot say you've read.

Are you taking an Oprah Winfrey approach here, as in "so what" if the book was false? It's a good book anyway? Or so what if the Wikipedia editor is a fake, his edits were likable any way? Your statement about the Lancet study on your talk page is false in at least two independent ways, and you can't vouch for the other articles either. Enough said. I know I'd immediately correct such a falsehood, but I leave that up to you.--Aschlafly 20:24, 25 April 2007 (EDT)

Andy, I have read the studies. I find it ironic that you falsely claim that I haven't, then dismiss them while proclaiming that you haven't either. If you'll bother to look at the list again, I've listed some details of the methods and the results of many of them, and I will add more as I have time. Perhaps you could clarify what exactly your issue is with the Lancet article. As I said above, it combined data from a number of previous studies and analyzed it. I will quote from it: "...neither the results from the individual studies nor their weighted average suggest any adverse effect on the subsequent risk of breast cancer for women with prospective records of having had one or more pregnancies that ended as an induced abortion, compared with women having no record of such a pregnancy. Furthermore, among the studies with prospective records of induced abortion, no significant variation in the results was found between those with objective and those with self reported information" p. 1010. If you want to claim that doesn't deny that abortion increases breast cancer go right ahead, but your logic is difficult to follow. I guess the collaborative attitude disappears when the data make you unhappy. You don't have to believe the data are accurate, but to deny them without bothering to look is ridiculous. Murray 21:56, 25 April 2007 (EDT)

Murray, you have a falsehood on your talk page about the Lancet article, and you're apparently sticking to it. I'm not going to spend all night arguing with you any more than I would spend all night arguing with Oprah Winfrey about her defense of the false book, or Wikipedia's defense of the fake editor. Your quote from Lancet does not deny that abortion increases breast cancer. It compares the effect of an abortion to a mythical lack of pregnancy. It does not deny that the pregnant woman increases her risk of breast cancer by having the abortion. If you don't grasp the distinction, then please read, and reread, and reread, until you do. Also, Lancet was not a study.

Earlier you said you "read the methods and results sections of all of them at some point, if not the intros and discussions." Now you say something quite different, that you "read the studies." Which is it? Regardless, if you read the Lancet article and still insist on calling it a "study" that denies that abortion increases breast cancer, then your persistence in this falsehood would be enough cause for concern.--Aschlafly 22:17, 25 April 2007 (EDT)

I guess that's the long version of saying that you're not going to bother to read or try to rebut the other studies, or indicate what your issue is with the Lancet article. Yes, I was careless with language in the last post. I read the methods and results of each. Obviously part of the issue here is that we have different definitions of risk. Here's the bottom line: a woman who has an abortion has the same risk of breast cancer as if she'd never been pregnant, or if she had a miscarriage, which is far more common than induced abortion. Pregnancy confers protection, but it does not follow that ending pregnancy increases a woman's risk, because her risk remains the same as it was before she got pregnant. It's pointless to discuss this further, but I would appreciate it if you would refrain from further implications that I'm lying. Others should read the studies themselves and draw their own conclusions. Murray 22:35, 25 April 2007 (EDT)

My opinion hasn't changed that the listing of the Lancet article on your talk page as an example of "Studies showing no abortion-breast cancer link" is false. Lancet is the only study in that list that I've read. Maybe some or all of the other articles listed are also unjustified. I see from your list that several of the other articles had very few subjects, and may thus be misleading in their claims.

But I'm not going to spend any more time repeating myself here. I'll refer this to the Conservapedia panel and they will make their own decision. Due to other pressing issues they may not get to this for a while.--Aschlafly 00:23, 26 April 2007 (EDT)

Ok, I read the study and have a few comments about the study article ...

1. "The research study excluded women who had abortions and then died from breast cancer!" This quote is misleading. The researchers censored women who had forms of carcinoma, which is cancer that can originate and spread elsewhere through the lymphnodes and is not breast specific. This makes sense because you wouldn't want to include cancer that originated in the lung in a breast cancer study.
2. "The research study treated women who left the specific question about past induced abortions blank, perhaps due to embarrassment, as though they did not have an abortion" They addressed this and said they also tried ommitting those who missed questions and did not see a change in results.
3. "The research report concealed how almost none of the subjects of the study were African American or Hispanic" while not covered specifically in the article, my assumption would be that with so many other variables included (age, body mass index, family history, birth control usage, etc) you would want to limit the study to a more similiar group of people. Keeping them of the same/similar race would keep potential conditions that vary by race out.
4. "Over 25% of the respondents to the questionnaire exhibited confusion about the wording, where the unfamiliar term "spontaneous abortion" was used to mean "miscarriage"." I don't see the point here, how could there be a difference between inducing and miscarriage? In both cases the fetus is lost prematurely. Also, maybe I missed it, but I don't see where you got that 25% from.

And I also want to point out that of the 105,716 participants: 2916 got breast cancer, and only 535 of them reported having abortions. If abortions did contribute to the risk of breast cancer you would expect that number to be way higher. Jrssr5 14:05, 25 April 2007 (EDT)

Age of Participants

The section on the participants being too young is just wrong. The study indicated that the subjects were ages 29 to 42 when they were enrolled in the study. Since this is longitudinal research (i.e. an over-time study) they have aged since then to 39 to 52. Given the sample size and the longitudinal component, it's unlikely that the sample itself is biased by age.

Do you also think we could make meaningful claims denying causation of heart disease based on studying teenagers? How about claims deny that cigarette smoking causes cancer based on 20-year-olds?--Aschlafly 14:41, 25 April 2007 (EDT)

Mr. Schlafly, you're sufficiently educated to understand the distinction here. The study is longitudinal, meaning that it enrolled participants and then studied them over a period of time- in this case ten years. So, it did not base its conclusions on a pool of 29-42 year olds but, rather, on a pool of participants who were in that range of ages when the study began but are between 39 and 52 now. While I would certainly like to see the longitudinal component continued for another 10 years or more, the data already available can be used to analyze the question at hand. If abortion does cause breast cancer then we would expect to see a statistically higher rate of cases of breast cancer emerge over those ten years among those who have had abortions than among those who have not. What we see is that women who have had abortions are no more likely to develop breast cancer than women who have never been pregnant and that a completed pregnancy reduces the likelihood of breast cancer. Thus, pregnancy is a protective factor but abortion is not carcinogenic. I don't expect you to recognize that distinction but, at the very least, be honest about the ages of the study's subjects. Your snide comments about studying cancer from smoking in 20 year olds misrepresent my point entirely. A crossectional study wouldn't convince me, but a study that enrolled subjects between the ages of 14 and 26 (median age 20) and then studied them for ten years until they were ages 24 to 36? Well, that's certainly a good start. If smoking does cause cancer, we would expect to see higher cancer rates among smokers over that ten year period, wouldn't we?

Miscarriages

Why does the article claim that miscarriages would not increase cancer rates? I have trouble imagining any biological mechanism that would cause abortion to increase cancer rates that would not apply to miscarriages as well. Do we have any basis for this claim? JoshuaZ 13:33, 26 April 2007 (EDT)