Public health researcher discusses new push for abortion study retractions

A University of Maryland public health researcher joined scholars from institutions worldwide to call for the retraction of four studies purporting to show damaging psychological effects of abortion, including one submitted as evidence in a decision overturning Roe v. Wade and now involved in a current Supreme Court case over availability of the abortion drug mifepristone.

The studies published from 2002–11 use fundamentally flawed evidence and analysis to come to invalid conclusions, said Julia R. Steinberg, an associate professor of family science who conducts research in the area of abortion and mental health. She previously published rebuttals of three of the four studies, calling for retraction of two of them upon publication.

Steinberg and co-authors from 16 other institutions laid out their case in a commentary published this week in the BMJ, which published one of the studies in question. The articles, which have been used in court cases and policy decisions for decades, include authors who have publicly declared themselves abortion opponents.

"This has previously been portrayed as a scientific battle, but it's not," Steinberg said. "The science is not divided on whether abortion causes mental health problems; existing research from various methodologically rigorous studies show that abortion does not cause mental health problems.

Their commentary came on the heels of the retraction of three articles about the dangers of chemically induced abortion by the publisher Sage Journals, drawing critical reactions and allegations of political targeting from some of the authors. Those studies are at the heart of the current case concerning the drug mifepristone, the most common means of abortion in the United States, as all aspects of the procedure come under increased scrutiny with the Supreme Court's striking down of the right to abortion.

Steinberg spoke with Maryland Today about the effect of bad science on society and her own research on abortion and mental health.

What are these studies about, and what are the problems?

They papers all essentially say abortion increases the risk of certain mental health issues, whether it's depression, anxiety, PTSD or substance abuse. There are four publications we talk about in the piece we just published, but there are two in particular that we've been pursuing retraction of since they were published.

One is a meta-analysis (of abortion and mental health studies) in the British Journal of Psychiatry, and the other study in the Journal of Psychiatric Research looked at abortion and a range of mental health disorders in a nationally representative U.S. dataset. It had fundamental problems with the study design and methodology, and inaccurate or incorrect factual statements in their methods, material errors in the authors' analysis of the data, and misleading presentation of the data. Each of the four had at least one of those kinds of problems.

Is this mainly about bad science existing at all, or concerns over how it's being used?

First of all, we don't want scientific studies out there that are not legitimate— for the sake of accuracy and factuality. The other reason, as you mentioned, is we don't want the effect that bad science can have on society. The study that claimed vaccines lead to autism was published in 1998, and it took until 2010 for it to be retracted. That article is still influencing people. And these studies that claim abortion harms women's health are being cited in courts today. The judge who ruled against medical abortion in Texas last year specifically mentioned the harms of abortion, and that he was going by the Coleman meta-analysis paper we are trying to get retracted. This paper gets cited all the time.

What perspective does your own research provide?

In my own findings, I may find a correlation between abortion and poor mental health outcomes, but we know from any intro to statistics class that correlation and causation aren't the same. If you simply look for the correlation without trying to control for variables, you find it, but it doesn't take into account that it may be driven by pre-existing mental health problems or other adverse experiences that could help lead to an unintended pregnancy that ends in abortion. When I do control for things like prior mental health problems or intimate partner violence, than that correlation goes away.

Was this a problem for the studies in question?

Yes, the issue with one of the pieces is that the authors said they looked at mental health outcomes at the time of the procedure to ensure that the abortion preceded the mental health effects that the authors attribute to abortion. But that's not what they actually did. Instead they used lifetime mental health outcomes, which could include those that occurred long before the abortion. That is exactly, 100% the wrong way to do it, and why that study needs to be retracted. And in analyses I have conducted with these data and presented to the editor, I have found that among those that had an abortion and mental health problems in this data, 72.4% had a first mental health problem at an age before their first abortion.

Some of the authors of these papers have complained that the previous retractions and your current effort are more political than scientific. How do you respond to that?

By saying you can have your beliefs about whether abortion should be legal or illegal and still carry out science in good faith or according to principles laid out by professional or ethical organizations. C. Everett Koop, President Reagan's surgeon general, was known as an opponent of abortion. He was commissioned by President Reagan to write a report on the health effects of abortion. Rather than conclude that abortion led to mental health problems as President Reagan wished, he told President Reagan that the scientific evidence was inadequate to support any scientific findings on abortion leading to women's mental health problems .

So this is not Julia Steinberg saying these four studies we discuss in the BMJ piece don't measure up to my standards. They don't measure up to the standards of professional, scientific, or ethical organizations such as COPE (the Committee on Publication Ethics) or the American Psychological Association, and according to COPE's guidelines should be retracted.