(or other health
By Joyce Arthur, Abortion Rights Coalition of Canada
March 1, 2010
In trying to determine whether there is an increased risk of breast cancer after an abortion, knowledge of medical research is important, but not all medical research studies are credible or useful. Unfortunately, the abortion issue has attracted more than its fair share of poor methodology.
There is a large body of complex and conflicting research on the alleged abortion/breast cancer link going back over 40 years, but many studies are flawed and essentially worthless, especially older ones. The most common errors are recall bias, small or non-random samples, non-matching control groups, and wrong assumptions that correlation equals causation.
A comprehensive 2003 review of the evidence by the National Cancer Institute (NCI) and over 100 of the world’s leading experts on pregnancy and breast cancer risk concluded that having an abortion or miscarriage does not increase a woman’s subsequent risk of breast cancer. Reputable health agencies all continue to support this conclusion, including the World Health Organization, American Cancer Society, and Canadian Cancer Society.
The anti-choice movement has been touting a 2009 study (Dolle at al, Journal of Cancer Epidemiology, Biomarkers and Prevention) that concluded that abortion raises the risk of triple-negative breast cancer (TNBC) by 40%. However, they never mention that this is a rare type of cancer that typically strikes women under 40 years of age. Because TNBC makes up such a small subset of all breast cancers, the overall conclusion that abortion does not lead to breast cancer has not changed. As Kathi Malone, one of the study authors, stated: “The weight of scientific evidence to date strongly indicates that abortion doesn’t increase the risk of breast cancer.”
Even the reported 40% increased risk for TNBC needs to be treated with caution. The statistical adjustments the researchers used to arrive at that figure may simply be an error caused by not accounting for all possible factors that might be contributing to the increased risk. The abortion data was apparently not adjusted for important factors like income or education. Further, to put things in perspective, a 40% increase in risk is not considered substantial in the field of epidemiology — a woman’s risk of getting breast cancer increases by 200-300 per cent if her mother had breast cancer.
The abortion data used in the Dolle et al paper is not even new, as it was taken from older studies in the 1980s and 1990s. This points to a further serious problem with the data. Most older studies (and even many recent ones) obtained women’s abortion history by interviewing them. But when women self-report their abortion, it has been shown that they under-report past abortions by 20 to 80 per cent, depending on the interview circumstances. This is because women with diseases that may be linked to past behaviours are much more likely to admit to those past behaviours, since they’re looking for answers. Women acting as controls are not as inclined to admit to strangers that they’ve done something in the past that might be disapproved of, especially when they have no stake in the outcome. Conclusions from such flawed methodology are not justified.
Anti-choicers often mention other studies that have allegedly found increased risks of other negative effects after abortion, including subsequent premature birth, low birth weight, and "post-abortion syndrome", which includes mental health problems, drug and alcohol abuse, suicide, and other negative psychological after-effects. By and large, these studies suffer from the flaws listed above, especially the common problem of confusing correlation with causation. Correlation means some other factor could be causing the negative effect, not abortion. For example, women with planned pregnancies tend to be in better physical and/or mental health compared to women who have abortions, and tend to have more stable lives. This can lead to significant differences in outcome that should not be attributed to the birth or abortion itself. The American Psychological Association (in its 2008 report Mental Health and Abortion) has concluded that mental health problems are not a direct result of choosing to have an abortion.
Related to this,
many studies have not used an appropriate
control group for comparison purposes (for example, mistakenly
comparing women with planned pregnancies to women with unplanned
pregnancies, even though the two groups have dissimilar
characteristics). The best control group would be women
carried to term after being denied an abortion. Finding such a control
is not impossible in western countries, because women are still
frequently denied abortion
various access barriers, particularly in the U.S. Such studies have
say they want patients to be given all
the information necessary to make an informed decision about abortion.
medical profession, including abortion clinics, already do this, but
they supply accurate
based on the best scientific evidence. It would be irresponsible and
to frighten women with false or unsubstantiated claims of medical risk.
Links to Further Information:
National Abortion Federation: Abortion and Breast Cancer Fact Sheet. http://www.prochoice.org/about_abortion/facts/abortion_breast_cancer.html
Obstetricians and Gynecologists of
Joyce Arthur. Abortion
and Breast Cancer — A Forged Link. The Humanist, March/April 2002,
Vol 62, No. 2, pp. 7-9. http://www.prochoiceactionnetwork-canada.org/articles/abclink.shtml
Federation: Post-Abortion Syndrome. http://www.prochoice.org/about_abortion/myths/post_abortion_syndrome.html
Joyce Arthur. Beware of Meaningless Studies by Anti-Choice Researchers. Pro-Choice Press, Autumn/Winter 2003. http://www.prochoiceactionnetwork-canada.org/articles/cmaj-study.shtml