In defence of David Fergusson

Tom Chivers is something of a rare beast; a Daily Telegraph journalist who is both scientifically literate and honest enough to offer a critical commentary on articles appearing in his own newspaper. As such, I was pleased to see Tom weighing in on the subject of Nadine Dorries’ latest anti-abortion campaign and even more pleased to see him point out Dorries’ misleading use of David Fergusson’s 2008 paper on abortion and mental health disorders [1].

But what caught my eye was a claim on the Right to Know website that “women who have an abortion are 30 per cent more likely to develop mental health problems”.

Now. I’m hugely in favour of evidence-based policymaking. But for evidence-based policymaking to work, you have to be open about what evidence you’re basing your policymaking on. There is no link from their website to any research paper or other source, so I have no way of knowing what their claims are.

However, some resourceful people on Twitter suggested that it might have been this paper by Fergusson et al, published in the British Journal of Psychiatry. It’s from 2008, and it does, indeed, say that women who have undergone abortions are 30 per cent more likely to develop mental illnesses. So that’s that, right?

Well: no. Not at all. For two, very important, reasons.

Of Tom’s two very important reasons, the first is more or less correct.  Correlation does not equal causation and, as he rightly points out, Dorries has cherry-picked Fergusson’s research for a statistic to support her argument, while ignoring entirely Fergusson’s actual estimate of the attributable risk of exposure to abortion, which amounted to 1.5% to 5% of the overall rate of mental disorders. Tom is, however, slightly wide of the mark in his interpretation of these results:

To my eye, that suggests that women with mental health issues are more likely to have abortions than those without. That in itself is very sad, of course, but it’s a very different thing.

Being picky, it’s not that women with mental health issues are more likely to have abortions but rather women who have abortions are more likely to present with multiple risk factors for mental health disorders; e.g. low socio-economic background in childhood, family dysfunction/breakdown, history of domestic violence, etc. That said, Fergusson’s discussion of the evidence for causality reveals that he introduced statistical controls into his study for over 30 confounding variables, including all of the above, and still found an elevated risk of subsequent mental health problems in women who had had an abortion, a finding that cannot reasonable be dismissed without further explanation.

Tom’s second reason is, however, somewhat wide of the mark – largely I suspect because he’s skim read Fergusson’s paper and hasn’t had the time to either digest his findings or read around his work and take into account other relevant publications (a flaw, yes, but an eminently forgivable one):

Second, that study is now rather out of date. The BJPsych itself has revisited the topic, with a 2009 systematic review. They suggest that while the Fergusson paper overcame some of the “methodological problems of previous studies”, it was still flawed. In their review of the literature, they found that “Only four studies fell into the authors’ ‘good evidence and low risk of bias’ category. All four studies showed a neutral effect of abortion on mental health, indicating no significant differences between the study comparison groups.” The 2008 Fergusson paper was the only “good quality” study that found a negative effect, and as described, it was weaker than Dorries and Field seem to claim.

Tom’s link leads to correspondence published in BJPsych [2] that is, itself, now out of date inasmuch as it does not address the findings of Fergusson’s follow-up paper on abortion [3], which was published in the same issue. This second paper by Fergusson et al. refines the findings of his previous paper in response to criticism which pointed out an important confounding factor that he had not previously controlled for; the degree to which women may have wanted to continue with their pregnancy even as they took the decision to have an abortion.

After reanalysed his findings, Fergusson found that:

…increasing reports of abortion-related distress were associated with increasing risks of mental disorders: women who reported at least one negative reaction to the abortion had rates of mental health problems that were approximately 1.4–1.8 times higher than women not exposed to abortion, and between 1.2 and 1.6 times higher than women who were exposed to abortion but did not report any adverse reactions to abortion. All of these findings are consistent with the conclusion that unwanted pregnancy terminated by abortion is an adverse life event that increases risks of mental health problems, with these increases in risk being proportional to the degree of distress associated with the abortion of an unwanted pregnancy.

These are anything but controversial findings and Fergusson is far from being the first researcher to identify the existence of a small sub-group of women who react badly to having an abortion and who are, consequently, subject to an elevated risk of subsequent mental health problems – see Major et al. (2000) [4] and Goodwin & Ogden (2007) [5].

Equally, it cannot be said that Fergusson’s findings are inconsistent with any of the published studies which failed to find a causal link between abortion and subsequent mental health disorders, including the two meta-analyses cited by Tom – Robinson et al. (2009) [6] and Munk-Olsen et al. (2011) [7]. We are looking here at a small sub-group of the population of women who have had an abortion and a modest increase in the risk of subsequent mental health problems, a ‘positive’ effect that is easily swamped by the much large ‘negative’ effect of a much larger group of women who experience no significant adverse reaction to abortion to the extent that their presence will have only a negligible impact on the findings of a large scale population study or meta-analysis.

Neither of these studies actually overturns Fergusson’s findings, rather they serve to validate the view expressed by both Fergusson and Brenda Major that vast majority of women who do have an abortion do not subsequently experience mental health problems without negating their identification of the existence of a small sub group of women who do. And it must also be acknowledged that neither of Fergusson’s papers exhibits the kind of serious methodological flaws that are rightly criticised by Robinson et al. Fergusson is categorically not a member of the coterie of politically-motivated anti-abortion researchers (David Reardon, Priscilla Coleman, Jesse Cougle, Vincent Rue, etc.) who have been justifiably criticised for trying to build ‘a literature to be used in efforts to restrict access to abortion’ [8] and cannot reasonably be subjected to the same kind of criticism as a consequence of his research being misused and misrepresented by anti-abortion campaigners.

For this, and other reasons, I would take issue with Rowlands and Guthrie’s assertion that:

Whether abortion causes harm to women’s mental health is a question that is not scientifically testable, as women with unwanted pregnancies cannot be randomly assigned to abortion v. abortion denied groups.

Although one cannot, for obvious ethical reasons, tackle this question by way of an RCT (randomly controlled trial) it does not automatically follow that this question is not scientifically testable. One could not, for the same reason, establish a causal link between cigarette smoking and lung cancer but few, other than perhaps the tobacco lobby, would see this as a convincing reason for dismissing the findings of Richard Doll’s British Doctors Study. Rowlands and Guthrie’s argument here strikes me as being political rather than scientific, hence their argument that:

We are supportive of their idea that abortion is not a psychiatric issue and that the Royal College of Psychiatrists should not develop a guideline on abortion. We would never want to go back to the psychiatric referral hurdle-jumping situation before and immediately after the Abortion Act came into force. The adverse effects of denied abortion must never be forgotten.

While a return to the days of ‘psychiatric referral hurdle-jumping’ would indeed be undesirable this is, in itself, no reason for the Royal College of Psychiatrists to eschew developing a guideline on abortion. Rather I would argue that the RCP has a clear ethical duty to develop clinical guidelines for the conduct of pre-abortion counselling and the identification of risk factors that may contribute to women finding themselves at an elevated risk of mental health problems following an abortion together with clear recommendations/guidelines on the provision of access to post-abortion counselling and aftercare. Guidelines of this kind would not, however, provide grounds for the refusal of an abortion unless there is evidence that a woman lacks the mental capacity to make an informed decision on whether to go ahead with the procedure, in which case it rightly becomes a matter for the courts to decide.

Even if one takes the view that it is not possible to establish a definitive causal link between abortion and subsequent mental health problems, it is certainly possible to identify a range of factors that, in conjunction with abortion, give rise to an elevated risk of mental health problems in some women and it would be unforgivably remiss of the RCP not to address these issues and provide guidance that ensures that women receive the best possible aftercare and that women who are thought to be at risk at the time they undergo pre-abortion counselling are made aware of the fact that they may need additional post-abortion support. In this I agree wholeheartedly with Fergusson’s observation that:

The finding that the extent of distress caused by the abortion is a predictor of subsequent mental health suggests the need for providers of abortion to: conduct thorough screening of abortion-related distress; to carry out adequate follow-up of those showing distress; and to counsel those showing distress about future mental health risks and the need for support.

If Fergusson’s papers contain anything at all that could be considered both contentious and open to accidental misinterpretation, as opposed to the deliberate misinterpretation and misrepresentation practiced by anti-abortion campaigners, then it his views of the potential legal implications of his work:

Collectively, this evidence raises important questions about the practice of justifying termination of pregnancy on the grounds that this procedure will reduce risks of mental health problems in women having an unwanted pregnancy. Currently there is no evidence to support the assumptions underlying this practice, and the findings of the present study suggest that abortion may, in fact, increase mental health risks among those women who find seeking and obtaining an abortion a distressing experience.

Fergusson is, I believe, commenting purely as a scientist here and stating, as a matter of scientifically supported fact, something that most of us already know; that the legal justification of abortion on mental heath grounds is nothing more than an outdated legislative fudge; a necessary concession made at the time that abortion was legalised in the UK in 1967, in order to secure parliamentary support for the Abortion Act, but one that has now outlived both its usefulness and credibility.

It is, perhaps unfortunate that Fergusson, in sticking firmly to the evidence, has advanced an argument that is open to interpretation and misinterpretation by both sides of the abortion debate. His comments could easily be taken, somewhat out of context, as supporting both the pro-choice view that abortion should, at least in the first trimester, be available on request and without the need to demonstrate any medical grounds to justify the procedure or as evidence for restricting abortion only to cases in which there is clear clinical evidence of a serious risk to the physical health and well-being of women.

As Fergusson states, unequivocally, that his research supports neither the strong pro-choice view nor the strong anti-abortion view, it seems reasonable to think that he is generally supportive of abortion rights, provided that appropriate provision is made to identify and support those women who are at risk of having adverse reaction to abortion, but would prefer greater legislative clarity on the matter, i.e. a clear acknowledgement in law of the legitimacy of ‘social abortions’. Such a position would certainly be consistent with the findings of some his other research, e.g.:

As a general rule, those becoming mothers prior to age 21 reported more mental health problems, had poorer educational achievement, lower income, and higher rates of welfare dependence. These associations also varied with age of first motherhood, with those becoming mothers prior to age 18 experiencing greater disadvantage than those giving birth between 18 to 21 years. These findings support and reinforce a large body of literature that has identified young parents as an “at risk” population for mental health problems, educational underachievement, and economic disadvantage (Caldwell & Antonucci, 1997; Corcoran, 1998; Osofsky et al., 1993; Phipps-Yonas, 1980; Simkins, 1984; Wellings et al., 1999; Zuckerman et al., 1987). [9]

Fergusson found, after controlling for confounding factor, that economic disadvantage was causally linked to early motherhood but that the increased prevalence of mental health problems and poor educational achievement were reflections of the influence of family, social and background factors that influence early motherhood.

Turning to the issue of abortion amongst young women and their subsequent life outcomes, Fergusson found that:

This analysis showed that, compared with those who became pregnant but did not have abortions, prior to adjustment for confounding factors, those having abortions had relatively advantaged outcomes on most measures of educational achievement, income, avoidance of welfare dependence, and partnership relationships. At first sight these findings clearly suggest multiple possible benefits of abortion. However, subsequent analyses suggested that most of the differences between the outcomes of pregnant young women having and not having abortion were explained by the fact that those seeking abortion were a more socially and educationally advantaged group prior to pregnancy. When due allowance was made for pre-pregnancy factors, most of the differences between pregnant women seeking and not seeking abortion became statistically non-significant. Nonetheless, even following control for pre-pregnancy factors, there was evidence of better educational attainment amongst those young women having abortions than those becoming pregnant but not having abortions. [10]

This, broadly speaking, confirms my own findings on the relationship between teenage pregnancy, abortion and socio-economic inequality in the UK, which show that although teenage conception rates are lower in middle class areas, abortion rates as a percentage of conceptions tend to be much higher. This is consistent with the view that it is, in fact, the middle classes that derive the greatest advantage from legal access to abortion, it being their offspring that are most likely to go on higher education and, consequently, most likely – if pregnant – to opt for an abortion so as not to compromise their educational and other life opportunities.

Lower down the socio-economic scale, the lack of viable economic opportunities available to young women not only makes it less likely that they will choose to have an abortion if they do fall pregnant but also ensures that those young women who do choose to have an abortion gain very little by way of socio-economic advantages over those who continue with their pregnancy.

Neither of these last two studies offer any substantive by way of evidential support for stringent restrictions on access to abortion, although this last paper does have profound implications for social and economic policy in the UK as it has operated since the 1980’s.

Education alone, the study suggests,  is not enough to tackle the decline in social mobility nor does it offer all but the very brightest of the working classes a route out of poverty let alone a means of avoiding welfare dependency. The same is also true of policies that seek to bear down on welfare costs by propelling people off benefits into low paid employment. The best that can be said for such policies is that they may serve to engineer a transfer of people from the group that the middle classes currently label ‘the undeserving poor’ to that which they label ‘the deserving poor’ but poor is still poor, irrespective of your source of income, hence the need to be deeply suspicious of any claims made by politician to the effect that simply getting people off benefits will automatically improve their socio-economic circumstances and/or improve the long-term outcomes of their children.

Hopefully, if I shown anything at all here it’s that David Fergusson and his team of researchers at the University of Otago are in no sense deserving of some of the more trenchant criticism that has been directed towards their work over the last 2-3 years as a consequence of its being adopted and wilfully misrepresented by anti-abortion campaigners.

Fergusson’s research is not ‘flawed’. It has limitations, which Fergusson freely and openly acknowledges in his papers and he cannot, therefore, be held to be at fault if others choose not to acknowledge those limitations when citing his work, even if some find his conclusions to be politically inconvenient and the matter-of-fact tone in which he presents his findings disconcerting for their lack of overt moralising.

That’s just how a top notch research scientist presents their findings.

References

[1] Fergusson, D. M., Horwood, L. J., & Boden, J. M. (2008). Abortion and mental health disorders: Evidence from a 30-year longitudinal study. British Journal of Psychiatry, 193, 444–451.

[2] Rowlands, S. & Guthrie, K. (2009). Abortion and Mental Health. British Journal of Psychiatry (2009) 195, 83

[3] Fergusson, D. M., Horwood, L. J., & Boden, J. M. (2009). Reactions to abortion and subsequent mental health.  British Journal of Psychiatry, 195, 420–426.

[4] Major B, Cozzarelli C, Cooper ML, Zubek J, Richards C, Wilhite M, et al. Psychological responses of women after first-trimester abortion. Arch Gen Psychiatry 2000; 57: 777 –84

[5] Goodwin, P. & Ogden, J. (2007). Women’s reflections upon their past abortions: An exploration of how and why emotional reactions change over time. Psychology and Health, 22,2. 231-248. (abstract only)

[6] Robinson, GE., Stotland, NL., Russo NF., Lang, JA., & Occhiogrosso, M. (2009). Is There an “Abortion Trauma Syndrome”? Critiquing the Evidence. Harvard Review of Psychiatry, 17,4, 268-290. (abstract only)

[7] Munk-Olsen, T,. Laursen, TM, Pedersen, CB., Lidegaard, O., & Mortensen, PB. (2011). Induced First-Trimester Abortion and Risk of Mental Disorder, New England Journal of Medicine, 364:332-339 (abstract only)

[8] Chamberlain, P. ( 2006). How Anti-Abortion Myths Feed the Christian Right Agenda. Public Eye Magazine.

[9] Boden J M, Fegusson DM, Horwood LJ. (2008) Early motherhood and subsequent life outcomes. The Journal of Child Psychology and Psychiatry, 49(2): 151-160

[10] Fergusson DM, Boden JM, Horwood LJ. (2007) Abortion among young women and subsequent life outcomes. Perspectives on Sexual and Reproductive Health, 39(1): 6-12.