Harvard abortion study
A group of researchers at the Harvard School of Public Health and Harvard Medical School, operating under a government research grant, published a report on research they conducted concluding that a group of relatively young nurses show no increased incidence of breast cancer from abortion.
The average age of diagnosis of breast cancer in America is 61 years old. But the average age of participants in this study who had an abortion was only 42 years old, too young for the average person to develop breast cancer. This age disparity is analogous to an attempt to draw conclusions about heart disease by studying teenagers. Furthermore, based on the definition of "average", 50% of breast cancer patients would have been diagnosed at an age older than 61. When added to the patients diagnosed between ages 42-61, a very large portion of breast cancer patients are not included.
However, the authors of the Harvard study asserted that earlier studies that reported that abortion increases the risk of breast cancer had been carried out "largely" on premenopausal women. This study is ongoing or longitudinal rather than retrospective, because young persons are monitored before they develop health problems rather than studying them after they develop health problems. There are advantages and disadvantages to both types of studies.
The research report concealed how almost none of the subjects of the study were African American or Hispanic, Because the researchers concealed this information, the press did not report the lack of African Americans and Hispanics in the study.
The research data show that abortion causes a sharp increase in the deadly progesterone receptor-negative (PR-) breast cancer. "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 abortion." In non-statistical language, this 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 25% less likely to get progesterone positive breast cancer. Among the study group there were 442 breast cancer cases among women who ever had an abortion. 59 (13%) were progesterone negative and 92 (21%) were progesterone positive.
However, the press reports concealed this alarming result, making it look like abortion is safe. The report's abstract concealed this important result also.
The research study excluded women who had abortions and then died from breast cancer! The study also excluded women who developed breast cancer early in the follow-up. "During follow-up, we censored women who were newly diagnosed as having breast cancer, had died, or were lost to follow-up." So those who were hurt most by their abortions were excluded, and this skewed the results towards a claim that abortion is safe.
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. Many would draw the opposite conclusion. Re-categorizing women from "had an abortion" to "did not have an abortion" inflates the numbers of breast cancers by women who ostensibly did not have an abortion. It would then falsely appear that abortion did not cause breast cancer.
Over 25% of the respondents to the questionnaire exhibited confusion about the wording, where the unfamiliar term "spontaneous abortion" was used to mean "miscarriage".  Miscarriages do not increase breast cancer risk, and this deceptive word choice would cause many respondents to mark their prior abortions as "spontaneous abortion" or miscarriage. That would reduce any genuine differences between induced abortion and miscarriages in the observed results.
Specific Examples Possibly Indicative of Fraud
Several statements, and omission of statements, in the study appear to some to be fraudulent:
- The study included only relatively few African American and Hispanics. But the researchers conceal this, perhaps knowing that disclosure would detract from the significance of the results for Americans, where a relatively high number of abortions and breast cancers are in these ethnic groups. A close examination of the researchers' statements suggests an artful deception. The researchers disclose merely that the subjects were 92% non-Hispanic white, falsely implying that most of the remaining 8% were African American and Hispanic. The researchers state that the study subjects "may not be representative of the social classes most likely to use abortion services." In fact, the researchers likely know that the term "may" is misleading and that the study subjects "are not representative of the social classes most likely to use abortion services."
- The researchers say that "99.1%" of the respondents "answered the question on spontaneous and induced abortion." But that statement is false. There was not one question on spontaneous and induced abortion. There were two questions. Later in the report the researchers say that 28,392 respondents did not answer the question about induced abortion.
- The abstract for the study, and the resultant publicity, concealed how the data showed that abortion was correlated with an increase in the most dangerous type of breast cancer. Instead, the researchers publicize the misleading "conclusion" that induced abortion was not "associated with the incidence of breast cancer." Causing a more dangerous type of a disease is "associated with" its incidence and should have been included in the abstract and publicity. It was misleading to state otherwise.
- Other issues remain unresolved and warrant explanation by the researchers. Why did they exclude from their analysis people who had an induced abortion and were newly diagnosed as having breast cancer? How would the results change if they were included? Also, how would the results change if those who declined to answer the question about induced abortion were treated as if that were a "yes" answer (i.e., the non-answer was due to privacy) rather than a "no" answer? Finally, do the researchers make their data available for independent review?
- None of the researchers hold positions in oncology (the branch of medicine that studies cancer), and one of the researchers is a nutritionist.
- The full text of the study is available for a $15 charge at http://archinte.ama-assn.org/cgi/content/full/167/8/814
- Michels K. Induced and Spontaneous Abortion and Incidence of Breast Cancer Among Young Women. Arch. Inter. Med. 2007;167:816 (Table 1).
- Id. p. 820.
- The study stated that the subjects were 92% non-Hispanic white; many of the remaining 8% must have been Filipino and other Asian groups, but the racial make-up of this group was not provided. For example, 15 percent of the registered nurses in San Diego are Asian or Filipino or other Pacific Islanders. More importantly, the researchers conceal the lack of African Americans and Hispanics, merely admitting that "The nurses in our cohort may not be representative of the social classes most likely to use abortion services." (p. 820).
- Study, p. 817.
- The misleading study abstract is here.
- Study, p.815.
- "Of the 105,716 women included in this analysis, 28,392 did not answer the question on induced abortion but answered the question on spontaneous abortion .... We assumed that the women who answered only half of the question did not answer the other question because of an oversight or because they felt that the question did not apply to them; we thus coded the missing response as 'no induced abortion.'" (p. 815)
- Study, p. 815.
- Study, p. 815
- Study p. 814
- Study p. 815