Abortion-Depression Study Flawed and Misleading
Abortion-Depression Study Flawed and Misleading Experts Dispute Interpretations of Federal Study
Springfield, IL (Nov. 2, 2005) — In 2002, an analysis of a nationally representative study of women funded by the U.S. government found that among women with an unintended first pregnancy, those who had abortions had significantly higher depression scores. But now a new analysis of the same data set, published last week in the British Medical Journal, has captured international headlines with the unqualified assertion that “abortion does not raise depression risk.”
Using new definitions for case selection and data coding, the authors of the new study, Sarah Schmiege of the University of Colorado and Nancy Felipe Russo of the University of Arizona, found no statistically significant difference in the depression scores of the two groups of women they studied: women who aborted a first unwanted pregnancy, and women who carried a first unwanted pregnancy to term.
In a lengthy rebuttal posted on the British Medical Journal web site, the lead author of the original study takes issue with the methods employed by Schmiege and Russo. David Reardon, the director of the Elliot Institute, based in Springfield, Illinois, charges that their results have been presented in a fashion that was calculated to mislead the public. He argues that while the authors’ report tacitly alludes to the fact that his earlier findings were correctly reported, they go to great lengths to imply that his findings have been supplanted by their own analysis, when a careful reading actually shows no contradiction in the two study’s results.
Reardon notes that his findings clearly showed that elevated rates of depression were most significant among women who were married at the time the depression scores were measured, an average of eight years after their unintended pregnancies. For these women, the risk of depression doubled, but the results among single women were less clear, probably because single women who carried to term were facing the struggles of being single mothers. But most notably, he said, the single women who had abortions were not faring better than the single moms.
“Schmiege and Russo have actually avoided replicating our actual analysis,” Reardon said. “Their study does not even attempt to look at the married among whom we found the greatest differences. They only look at mixed groupings where we already knew the data is too weak to support statistically significant findings.”
Indeed, Reardon says, it is very easy to find statistically insignificant results with this particular data set because women in the sample only report 40 percent of the expected number of abortions compared to national averages. In other words, 60 percent of the abortions were concealed from interviewers. Reardon believes that women who hide their abortions are also most likely to experience shame, grief, and depression following an abortion.
His complaint isn’t just that Schmiege and Russo have failed to show results for married women. He also charges that they chose new selection rules that were designed to bias their results and “muddy the waters.”
First, they omitted women who had abortions and who, up to ten years later, reported that the pregnancy they had aborted had been wanted. Since aborting women who have mixed feelings about their pregnancies are most likely to have depression, Reardon says, this selection rule biased their sample by throwing out depression cases among the abortion group simply because they were predictably likely to have depression. Schmiege and Russo, on the other hand, insist this exclusion was appropriate because women who abort wanted pregnancies are not the same as women who abort unwanted pregnancies.
Second, to limit their study to the effects of a first unwanted pregnancy, Schmiege and Russo argue that women who have subsequent abortions should be included in the control group. Reardon objects to this selection rule saying that the women aborting a first pregnancy should be compared only to women who do not report any abortions.
“What they are doing is comparing women who have depression following abortion of a first pregnancy with a group that includes women who are having depression after abortions of second or third pregnancies,” he said. “This only serves to muddy the statistical analysis. Of course there will be less significant differences between the two groups, precisely because you are adding women who have had abortions to the wrong side of the equation. Our original results were already diluted once because the high concealment rate caused many women who actually had abortions to be classified as not having had abortions. But now these researchers are deliberately putting women more women who have had abortions into the control group.”
Schmiege and Russo also misled readers, Reardon says, in the portions of their discussion which would lead uninformed readers to believe that Reardon’s previous study is the only study linking abortion to depression. In his rebuttal, Reardon cites over a dozen studies linking abortion to higher rates of depression. One, a study of 2,525 women showing significantly higher rates of depression and suicidal thoughts among women who had abortions, was actually written by Nancy Felipe Russo, co-author of the current study.
Reardon believes the body of research linking abortion to depression is so definitive that there is no longer any question about whether or not the two are truly associated. While more research is needed to understand what other factors may help explain the connection to abortion, he writes, there is no debate among nonpartisan researchers about the existence of the association.
Russo and Schmiege decision to avoid any mention of this body of research in their current study, Reardon worries, effectively put headlines before science. Leaving these citations out, he suggests, cleared the way for them to assert in their conclusion that “well designed studies have not found that abortion contributes to an increased risk of depression.” It was that claim which many reporters translated into an unqualified assurances that “abortion does not cause depression.”
Reardon concludes his rebuttal with an appeal to Schmiege and Russo to join him in a call for federal funding of a major longitudinal study examining the associations between women’s reproductive history and their physical and emotional well-being. “Former Surgeon General C. Everett Koop recommended such a study over sixteen years ago, but it has never been done,” said Reardon. “With the proper data, we could finally get conclusive answers to these important questions.”
# # #
Schmiege and Russo’s study can be viewed at http://bmj.bmjjournals.com/cgi/content/abstract/bmj.38623.532384.55v1
Reardon’s rebuttal is also posted on the BMJ web site at http://bmj.bmjjournals.com/cgi/eletters/bmj.38623.532384.55v1#1204555
