Post by MO on Jun 6, 2004 19:18:42 GMT -5
Expert on Cancer's Link to Abortion Calls Recent Study Flawed 6/4/2004
By Austin Ruse
A widely touted study denying a link between abortion and breast cancer failed to include research that contradicted its findings.
A widely touted study denying a link between abortion and breast cancer in the British medical journal, The Lancet, failed to include research that contradicted its findings according to one prominent researcher. Dr. Joel Brind said the Lancet report incorrectly compares rates of breast cancer among women who have had abortions with women who have never been pregnant rather than with women who have given birth, a methodology that ignores the fact that a pregnant woman who chooses abortion is "at a higher risk of breast cancer than ... had she chosen to carry the pregnancy to term."
In the Lancet article, "Breast cancer and abortion: collaborative reanalysis of data from 53 epidemiological studies, including 83,000 women with breast cancer from 16 countries," the authors claims to analyze "(d)ata on individual women from 53 studies undertaken in 16 countries with liberal abortion laws," according to a summary of the study's methodology found on the journal's website (www.thelancet.com). In a recently written rebuttal to that report, Brind, president and co-founder of the Breast Cancer Prevention Institute, says those who oversaw the study engaged in a "highly biased selection process" that ignored 11 "studies showing valid evidence of the ABC (abortion/breast cancer) link" and included "invalid studies whose flaws had been documented in the scientific literature." Lind said the researchers also fail to even mention or acknowledge four valid studies that support the ABC link.
The study readily admits that, "(p)regnancies that result in a birth are known to reduce a woman's long-term risk of developing breast cancer" yet the authors do not include women who give birth in their statistics, a method of measurement which Brind calls "very artificial." "Obviously, a woman considering abortion is already pregnant, and does not have the option of 'never having had that pregnancy.'"
The Lancet report also makes a distinction between studies that make use of retrospective data and prospective data. The data in retrospective studies relies on women reporting whether or not they have had abortions after having been diagnosed with breast cancer; prospective data is collected prior to diagnosis. Results from retrospective studies show a clear ABC link while the prospective data has not shown the same type of connection. The Lancet report interprets this to mean that retrospective data is unreliable calling the discrepancy proof that cancer patience are more likely to admit to past abortions than healthy women. Brind says this explanation, called a "response bias hypothesis," is invalid in part because "it is a violation of epidemiological methodological principles." Further Brind says the study that The Lancet report cites as proof of "response bias" "has been shown to be invalid" with the original authors retracting a key piece of evidence, a significant fact ignored by the Lancet study. Brind says that four studies done on the topic of reporting bias have "convincingly ruled out" response bias as an explanation for the differing results in retrospective and prospective studies.
In the coming days Brind will publish on his website his full response to the Lancet study.
www.cwfa.org/articles/5761/CWA/life/index.htm
By Austin Ruse
A widely touted study denying a link between abortion and breast cancer failed to include research that contradicted its findings.
A widely touted study denying a link between abortion and breast cancer in the British medical journal, The Lancet, failed to include research that contradicted its findings according to one prominent researcher. Dr. Joel Brind said the Lancet report incorrectly compares rates of breast cancer among women who have had abortions with women who have never been pregnant rather than with women who have given birth, a methodology that ignores the fact that a pregnant woman who chooses abortion is "at a higher risk of breast cancer than ... had she chosen to carry the pregnancy to term."
In the Lancet article, "Breast cancer and abortion: collaborative reanalysis of data from 53 epidemiological studies, including 83,000 women with breast cancer from 16 countries," the authors claims to analyze "(d)ata on individual women from 53 studies undertaken in 16 countries with liberal abortion laws," according to a summary of the study's methodology found on the journal's website (www.thelancet.com). In a recently written rebuttal to that report, Brind, president and co-founder of the Breast Cancer Prevention Institute, says those who oversaw the study engaged in a "highly biased selection process" that ignored 11 "studies showing valid evidence of the ABC (abortion/breast cancer) link" and included "invalid studies whose flaws had been documented in the scientific literature." Lind said the researchers also fail to even mention or acknowledge four valid studies that support the ABC link.
The study readily admits that, "(p)regnancies that result in a birth are known to reduce a woman's long-term risk of developing breast cancer" yet the authors do not include women who give birth in their statistics, a method of measurement which Brind calls "very artificial." "Obviously, a woman considering abortion is already pregnant, and does not have the option of 'never having had that pregnancy.'"
The Lancet report also makes a distinction between studies that make use of retrospective data and prospective data. The data in retrospective studies relies on women reporting whether or not they have had abortions after having been diagnosed with breast cancer; prospective data is collected prior to diagnosis. Results from retrospective studies show a clear ABC link while the prospective data has not shown the same type of connection. The Lancet report interprets this to mean that retrospective data is unreliable calling the discrepancy proof that cancer patience are more likely to admit to past abortions than healthy women. Brind says this explanation, called a "response bias hypothesis," is invalid in part because "it is a violation of epidemiological methodological principles." Further Brind says the study that The Lancet report cites as proof of "response bias" "has been shown to be invalid" with the original authors retracting a key piece of evidence, a significant fact ignored by the Lancet study. Brind says that four studies done on the topic of reporting bias have "convincingly ruled out" response bias as an explanation for the differing results in retrospective and prospective studies.
In the coming days Brind will publish on his website his full response to the Lancet study.
www.cwfa.org/articles/5761/CWA/life/index.htm