Last week, I wrote about how most of the studies examining the impact of abortion on mental health – or lack thereof – can’t tell us much. The crux of the issue is that women do not get abortions according to some random criterion – there are non-random differences between women who have abortions and women who do not, and these non-random differences may drive any differences in outcomes observed between women who had abortions and those who did not. However, I did mention one study – the Turnaway Study – that attempted to use a natural experiment to circumvent the non-randomness issue.

The Turnaway Study recruited a large number of women over a three year period who were seeking abortions across the United States. The researchers worked with staff at abortion-provision facilities to recruit study participants. The women were told that if they participated in the study, they would be interviewed by telephone once every six months over five years. About 40% of the women agreed to hear more about the study, and of them, about 85% ultimately participated in the study.

The natural experiment that the Turnaway Study sought to exploit was variation in gestational limits. The researchers argued that a woman who is seeking an abortion just below a State’s gestational limit is probably pretty similar to a woman seeking an abortion just above the State’s gestational limit in all the ways we care about. Therefore, by comparing women who were just below the gestational limit, and received an abortion, to women just above the gestational limit, who did not, we can get an idea of the impact of getting a wanted abortion versus being “turned away” from a wanted abortion – hence the name of the study.

As natural experiments go, this isn’t a bad one. A woman who looks for an abortion at 11 weeks and 6 days gestation is indeed probably similar in pretty much all the ways we care about to a woman looking for an abortion at 12 weeks and 0 days gestation. Given the comprehensive longitudinal nature of the study (a large sample recruited, and studied for a long time), it’s no surprise that the data from the study have spawned quite a number of studies.

However, the groups the study uses are a bit wider than a one day cut-off. Instead, the study compares women who were two weeks or less under the gestational limit (the “near limit” group) to women who were three weeks or less over the gestational limit (the “turn-away” group). In other words, if a particular State allows abortion up to 20 weeks, the study compared women who sought abortions and were between 18 and 20 weeks pregnant to women who sought abortions and were between 20 and 23 weeks pregnant.

While some States included in the sample had gestational limits as low as ten weeks, they are outliers. When this study was conducted (2008 – 2010), most States had abortion access for the first two trimesters. This is reflected in the data: the average gestation of women in the “near limit” group was 19.7 weeks, while the average gestation of women in the “turn-away” group was 23.1 weeks. Straight away, there’s a problem: this study might tell us lots about the impact of having an abortion, or not, in the second trimester, but can it tell us much about the impact of having an abortion in the first trimester, which is when the vast majority of abortions happen? I’m not so sure! So I think the results should be interpreted in this light.

Which brings us to: the results! Many studies have been written using the Turnaway dataset, but the main study to examine mental health impacts was published in 2017 in the Psychiatry Journal of the American Medical Association. The results are cited, confusingly, by pro-choice and pro-life groups alike in support of their respective positions, perhaps because, like many public health results, they are nuanced!

The study found that after one week, women in the “turn away” group were more likely to report anxiety symptoms, lower self-esteem and lower life-satisfaction compared to those in the “near limit” group. There was no difference in the levels of depression between the two groups. Pro-choice advocates point to this paper to show that being denied abortion has adverse impacts on women’s mental health in the short term.

However, all these differences between the “near limit” group and the “turn away” group had disappeared within a year, and some had disappeared by the first follow-up, at six months. In fact, there was no statistically-significant difference on any mental health outcomes between the two groups over the five years of the study. Pro-choice advocates use these results to argue that there is no evidence that abortion has adverse impacts for women’s mental health: to quote the JAMA study directly, “These findings do not support policies that restrict women’s access to abortion on the basis that abortion harms women’s mental health”. Pro-life advocates,on the other hand, use these results to argue that any negative impacts for mental health from being denied an abortion are short-lived.

So, everyone’s right! No matter what your view on abortion and mental health, the Turnaway Study will support it! As a pro-life person who’s had her fair share of pregnancy-related stress, it’s hard for me to see how someone in the midst of a crisis pregnancy would not experience anxiety, lower self-esteem or lower life-satisfaction. I am, however, heartened by how short-term these impacts seem to be. Pro-life people aren’t crazy when we say that real-life support is so important for women in crisis pregnancy. Getting through the short term shock, anxiety and stress may be just what women need, and the Turnaway Study can help reassure everyone that there is no reason to think that any of these mental health impacts will last for the long term.

But the real message from the Turnaway Study is that the question, as ever, boils down to whether abortion is right or wrong. If abortion is an ethically acceptable choice, then pro-choice people are correct to point out that there is no evidence that we should restrict abortion in order to protect women’s mental health. On the other hand, if abortion is not an ethically acceptable choice, then pro-life people are correct to point out that real, non-judgemental support is what women need, in order to enable them to navigate the short-term stresses and anxieties that accompany unplanned pregnancy. Either way, in the long run, women’s mental health does not appear to be impacted.