Abortion and mental health research not as clear-cut as reported; no surprises there

In a previous post I questioned recent research which was widely reported as “proving” that there’s no positive mental health benefit associated with abortion – thus basically “disproving” the idea that abortions are being legitimately permitted on mental health grounds in NZ.

Via some helpful pixies, I was able to obtain a copy of the full article, and … yeah.  No surprises here.

The fact is, it’s a literature review, which revisits the results of previous studies which had pretty inconclusive results regarding the abortion-mental health link, usually because:

  • they didn’t distinguish between unwanted and unintended pregnancy
  • they didn’t compare people granted abortion against people denied abortion (it’s a lot easier to come to terms with things when you have no other option)
  • some of the studies were carried out by people with explicitly antichoice views

So … yeah, pretty much what we already knew.

Here’s the rub, though:

It may also be suggested that the studies reviewed contain multiple problems research design, analysis and interpretation that prevent any clear conclusions from being drawn. In comparison to the ideal of testing the mental benefits of abortion using a randomized controlled trial, it is clear that existing observational studies provide only limited and potentially flawed evidence on the mental health consequences of abortion. However, this observation does not impugn the validity of the conclusion that: at the present time there is no credible scientific evidence demonstrating that abortion has mental health benefits.

So sure, you might say that some/many/all of the studies we looked at were flawed/biased/unscientific, but that doesn’t change the fact that there are no studies which prove the opposite.

Oh, except that:

In addition, it could be suggested that the comparisons made in the study between those having abortion and those having unwanted or unintended pregnancy do not provide an appropriate test of the mental health effects of abortion. A better comparison would be between those having abortion and those refused abortion.

In addressing the research question, we have taken the approach used by the majority of the reviews of the mental health consequences of abortion (Bradshaw and Slade, 2003; Charles et al., 2008; American Psychological Association, 2008; National Collaborating Centre for Mental Health, 2011) by comparing those having abortion with those coming to term with unwanted or unintended pregnancy. Further, to our knowledge, the only study that has compared those having abortion with those refused abortion is the re-analysis of Gilchrist et al. (1995), conducted by the AMRC review.

This re-analysis found that, for a number of outcomes (psychotic illness, non-psychotic illness, self harm), those refused abortion fared worse than those provided with abortion, with this difference being statistically significant (p<0.01) for psychotic illness. This evidence suggests the possibility that further studies making such comparisons could demonstrate positive benefits for abortion.

However, at the present time the evidence is far too limited to conclude that abortion reduces any mental health risks of unwanted or unintended pregnancy.

[Emphasis and paragraph breaks mine.]

Sure, you might raise the totally valid point that unplanned =/= unwanted and this might muddy the results, but fuck you, that’s what everyone else does.  And anyway, there’s a small amount of research which does actually suggest that if we compared apples with apples we’d get different [more accurate] results than when comparing apples and tractors, but fuck you, because there’s not much of that so go away.

And here’s the Bonus Rub Cookie:

A NEW ZEALAND professor whose work has been used by pro-life groups to contend that abortion holds no mental health benefits for pregnant women has said that his research is too limited to make any definitive conclusions.

… are you fucking kidding me.

Look, apparently Dr David Fergusson considers himself prochoice.  Which is great.  And no one wants to be the big scary feminist meany-head who scares off the gentle, placid, well-meaning allies.

But what the fuck is with a supposedly pro-choice researcher putting out research which is too limited to make any definitive conclusions when anyone with half a fucking ounce of awareness would understand exactly how said limited research will be twisted to fuck with the lives of pregnant people?

Is there some kind of “papers published” quota researchers have to meet?  With no other important KPIs like “papers must be actual good research” or anything?

Why the fuck didn’t I go into academia?

Another issue with the original research:  here’s the five “mental health outcomes” measured, which were then equated with a general picture of “mental health”:

anxiety, depression, alcohol misuse, illicit drug use/
misuse, and suicidal behaviour.

… all of which are pretty complex things.  I mean, are we really going to equate a once–pregnant person who smokes the occasional joint among friends with a once-pregnant person who starts huffing paint thinner to get through the day?  Is post-partum depression – common even in pregnant people who are happy and enthusiastic about having a baby – being included and thus compared with other types of depression which might exist regardless of birth status?  Do we ignore the fact that suicidality might be affected by the gigantic social pressure on new parents not to “abandon” their infants?

Sure, some of these questions may merely highlight my own lack of clinical psychological training, but come on.  Even the dude who wrote the damn article thinks it doesn’t pass muster.


H/T Alison McCulloch and the magic pixies.


  1. Chris Miller

    “Is there some kind of “papers published” quota researchers have to meet? With no other important KPIs like “papers must be actual good research” or anything?”

    This is actually sort of iffy, because while I definitely agree that he should have known what would happen, there is a big push in a lot of scientific circles to publish the results of ALL studies, not just the mostly status quo of only publishing if you get the result you were looking for. (A lot of people actually want databases where you register a study before you perform it and then if you don’t publish the results people can tell, but obviously there’s quite some pushback against that idea by marketing research folk.) It’s especially useful for studies with inconclusive results because if you get a whole bunch of them on the same sort of topic and group them together you can see patterns that you can’t see by looking at just one. Whether that should apply to studies that are inherently flawed is up for debate too though, because on the one hand people without the right training may have trouble realising that, but on the other some of the data might actually turn out to be useful for other reasons.

    • QoT

      This is a very good point, but as you say – Fergusson should have known what would happen, being someone with a lot of expertise in this field. It kind of bugged me how the report kept saying “no one’s looked at these studies from THIS specific angle before” … but [unless, as is likely, I'm missing some awesome stats context] then the result is just “oh yeah. Still can’t really say one way or the other.”

      • Daniel Copeland

        I don’t know. There’s only so much you can do to avoid being deliberately quoted out of context by people with an agenda. I’d put the responsibility on the people doing the quote-mining.

        • QoT

          Except this isn’t “quote-mining”, Daniel – quote-mining is pulling things completely out of context, e.g. the Margaret Sanger quotes I recently posted on.

          You don’t have to pull things out of context when they clearly state “there isn’t enough scientific evidence to establish that abortions are beneficial to mental health” and explicitly link this to questioning the legality of abortions performed on mental health grounds. That IS the context.

      • Chris Miller

        Yes the quoted segments you provide are pretty heavily weighted towards “we don’t have enough information” and without really discussing the implications of the information we do have.

        • QoT

          Yeah, the “implications” are pretty much summed up as “there’s probably no benefit. But we don’t have enough information. And the one study which compared abortions vs refused-abortion showed there might be a benefit, only we don’t have enough information.”

  2. AlisonM

    I found this paper just totally perplexing for so many reasons including the ones you raise. We’ve been dealing with Fergusson’s abortion-related work drawn from the longitudinal Christchurch Health and Development Study for many years – and now this leap into meta analysis of studies of (mostly) known iffy quality on the same topic. I’m really curious about the motivation. I understand a commentary on the paper for the same journal may be in the works, and am looking forward to that, so won’t try to unpick this further. Just to observe what a neat little closed circle anti-choice peeps have going: (i) Make women feel bad about abortion by telling them it’s murder/death/kill/sin etc. (ii) study them to see if they feel bad about abortion (iii) conclude they feel bad about abortion. [Then, as an additional bonus: (iv) argue because they feel bad about abortion they shouldn't be able to get one.] Nifty. Imagine what you might do if you REALLY cared about the mental health and wellbeing of those who are pregnant (whether they’re considering abortion or plan to continue the pregnancy).

  3. Jo

    “Is there some kind of “papers published” quota researchers have to meet?”

    Yes. The current model of university funding is principally based on research output, ie how many papers are published. The more papers published, the more money the university makes. So there’s that.

  4. Pingback: From a prochoice position, changing our abortion laws DOES MATTER | Ideologically Impure