Robust design

The prospective design, cross-checked against retrospective information, its resilience to confounding, and the use of structured interviews as distinct from clinical diagnosis on consultation ^{Reference Gilchrist, Hannaford, Frank and Kay3} make Fergusson et al's study arguably the most robust in its field, conferring on it significant authority. In addition, the attrition rate of 20% over 30 years was much lower than in other studies, ^{Reference Major, Cozzarelli, Cooper, Zubek, Richards, Wilhite and Gramzow4} and the accuracy of abortion ascertainment was much higher. ^{Reference Schmiege and Russo5} Based on the results, the authors conclude that abortion has a small causal link to subsequent mental health problems and that the impact on the reported prevalence of psychiatric disorders in women who have had an abortion compared with women who have not was up to 30%, while the attributable risk to the totality of psychiatric morbidity in the general population was modest at 1.5–5.5%.

Medico-legal and clinical implications

The results of this study reinforce the opinion that the Royal College of Psychiatrists was indeed wise in producing a nuanced position statement on abortion and mental health ⁶ and, although demurring from supporting a causal link, advised that the conflicting scientific information on this be provided to women seeking abortion in the interests of fully informed consent. Although Fergusson et al address legislative concerns in their discussion, there are more immediate medico-legal implications flowing from the findings; namely, the prospect of litigation against abortion providers for failing to provide women with information of a possible causal link between abortion and subsequent mental health problems.

The clinical relevance of this study is arguably more important, since these women are vulnerable to mental problems and as part of the wider population may present to their general practitioners or to the psychiatric services. Apart from issues of consent, there are other treatment implications. The first is the clear necessity for vigilance for the possible onset of adverse reactions post-abortion, as discussed by others, ^{Reference Gissler, Berg, Bouvier-Colle and Buekens7} although the high proportion of women who fail to attend for post-abortion assessment will often not enable this. Moreover, since the emergence of emotional problems and/or help-seeking may be delayed, general practitioners will have a key role in this. So, clear care pathways for those at risk of or suffering from psychiatric disorders in the context of abortion should be incorporated into guidelines developed by the various professional bodies, as recommended in the College's statement. ⁶ These should include guidance on the agencies to which such women should be referred.

There are some difficulties in this regard since it is the voluntary sector that offers much of the assistance to these women at present. Although it might be assumed that the obvious bodies to offer post-abortion interventions are those providing abortion services, there is evidence that women adversely affected by abortion do not return to the providers for this.

Developing interventions

Mental health professionals generally, with the possible exception of perinatal psychiatrists, have little experience in managing abortion-related disorders and, although these include a range of common psychiatric disorders, there are aspects of the symptoms that require specific expertise. These include managing guilt, anger, avoidance and dissonance concerning the status of the foetus. A further gap is the absence of scientific information on the preferred interventions, whether psychological, pharmacological or a combination of these. Moreover, some women, even those without specific religious beliefs, seek comfort from ministers of religion, and their role, along with that of voluntary organisations, in relation to that of doctors, will have to be clarified when developing guidelines.

This study also has implications for psychiatrists in training. The reticence to routinely inquire about induced abortion during history-taking is a noticeable deficiency among many trainees. Analogous to the growing awareness of sexual abuse during the 1990s, which led to changes in history-taking, an increasing awareness that for a minority of women abortion may be of aetiological significance should stimulate training is this aspect of psychiatric interviewing. The danger of either condemning the abortion decision or minimising the emotional impact is one that has been identified by women themselves, ^{Reference Goodwin and Ogden8} and this sensitivity must specifically be incorporated into training. Bearing in mind the increasing emphasis on service user involvement, a possible role for women who have had adverse reactions to abortion might assist in this aspect of training, along with qualitative studies.

The findings of this study will provoke controversy but they should not be clouded by ideology. Rather, the focus should be on identifying vulnerable groups of women and providing optimum treatment for them, whatever the aetiology of their mental health problems.

Are there vulnerable groups?

Despite these caveats, if abortion were associated with a modest increase in mental health problems this would not be surprising. Reproductive and gynaecological events have long been associated with psychological sequelae. Aetiological fashions change, as do research methodology, societal attitudes, reproductive epidemiology and technology. Studies of links between these events and adverse sequelae have produced variable results but consistently there are vulnerable subgroups who are at increased risk. Is this not also likely to be true for abortion?

Perhaps future research might help to better identify the vulnerable subgroups. It is unlikely that it will resolve whether or not abortion causes more mental health problems than continuing with a pregnancy. It will never be possible ethically to conduct randomised controlled trials. Women who choose to abort an unwanted pregnancy will be different from those who choose to continue; comparison will always be problematic.

The reasons for an abortion are personal and distressing. Some of these have changed since the Abortion Act 1967, when single motherhood was considered to be shameful. On the other hand, technology has changed and early-pregnancy diagnosis of foetal abnormality has given rise to new reasons for abortion. In the UK prior to the Abortion Act, and still throughout low- and middle-income countries, illegal abortion was a leading cause of maternal death. ^{Reference Lewis and Drife4} This, together with protecting doctors from prosecution, was a factor in the legalisation of abortion.

What are the implications of this study for the legal status and practice of abortion? The rights or wrongs of abortion are not primarily psychiatric or even scientific questions, but rather moral, ethical and legal issues. There are, however, two areas of controversy for which these findings have implications.

The Abortion Act: Clause C

First, very few women seeking an abortion have a mental illness and fewer will see a psychiatrist, yet 94% of abortions take place under Clause C in Certificate A (which the doctor approving the abortion must sign). ⁵ Interestingly, despite ‘the risk to mental health’ to which it refers this is often called ‘the social clause’. It is not possible to know whether a woman would have done better if she had proceeded with the pregnancy, kept her baby or given it up for adoption. Population-based studies showing a modest increase in mental health consequences are unlikely to help the individual clinician or woman. Because of these difficulties and the equivocal nature of the evidence, should society and legislators consider moving to a legal framework that acknowledges the ‘fig leaf’ of Clause C and the reality of almost unrestricted access to first-trimester abortion?

Counselling and informed consent

Second, some have argued for mandatory counselling and informed consent about risk to mental health for all women seeking abortion. ^{Reference Henderson6,7} All women should be asked by the professionals involved about their reasons and alternatives should be discussed. Most women will therefore receive this form of ‘counselling’. A wise clinician should spot the vulnerable subgroups. These could include women whose pregnancy was initially wanted but then they became terrified of some consequence, including the recurrence of a previous postnatal illness; the very young; those who have been put under undue pressure; those with previous abortions; and those whose ambivalence was evident. Such women might be referred for further counselling or psychiatric opinion. Late abortions, particularly for foetal abnormality, are associated with an increased risk of major depressive illness in the short and medium term. These women too might benefit from talking through their decisions and from the offer of support. However, this is a far cry from mandating counselling for all. A possible consequence of this could be delay, with more late terminations and an increase in psychiatric morbidity. It would be remarkable if abortion was not associated with a rise in distress and even episodes of anxiety and depression: all other gynaecological and reproductive events, and most surgical procedures, are, as indeed are life events. Informed consent for surgery does not include a warning of psychological hazard. We do not believe that the evidence is strong enough to support mandating such advice for abortion.

Perhaps the Royal College of Psychiatrists should not have a statement on abortion. Some medical Royal Colleges have this policy. There will never be a consensus among the College's members; indeed, there are a range of opinions among the authors of this commentary. We do, however, agree that abortion is not a psychiatric but a moral, ethical and legal issue, and that the views of College members will be as diverse as in the population at large.