The clear and thoughtful commentaries by Casey and Oates et al ^{Reference Casey, Oates, Jones and Cantwell1} raise a number of important issues about the implications of our research ^{Reference Fergusson, Horwood and Boden2} regarding the linkages between abortion and mental health.

The first of these is identifying vulnerable groups. Both commentaries raise concerns about the identification and treatment of vulnerable clients. These issues are most clearly stated by Oates et al, who record some disappointment that our paper did not identify the features of women who may be vulnerable to later mental health problems. An important reason for this was the length constraints imposed on our paper. Although the editors very kindly allowed us considerable latitude with the journal word limit, within the space we had the most we could do is report the main analysis of the data. However, we have a further paper under preparation that looks at the very issue raised by Oates et al. Without releasing our findings prematurely, we are able to comment that we have: (a) identified a group of women who are at increased risk of subsequent mental disorder following abortion; and (b) these women are distinguished by high levels of guilt and distress at the time of the abortion. We hope to be able to publish these findings within the next 6–12 months.

Second, the Abortion Act Clause C. It is our collective view that the most important implications of our findings relate to the current legal justification for abortion in the UK, New Zealand and a number of other jurisdictions in which abortion is authorised principally on medical grounds. ^{Reference Rahman, Katzive and Henshaw3,Reference Boland and Katzive4} In all of these jurisdictions, the great majority of abortions are authorised on mental health grounds. Our findings strongly challenge the use of mental health criteria as a routine justification for abortion. Our results suggest that the mental health risks of having an abortion may be greater and are certainly no less than the risks of coming to term with an unwanted pregnancy. Further, as far as we can tell, there is no evidence that suggests that the mental health risks of abortion are less than those of continuing with an unwanted pregnancy. To establish this would require a series of replicated studies showing that the mental health outcomes of those having an abortion are better than those of an equivalent series of women coming to term with an unwanted pregnancy. No such evidence exists. This situation creates a clear conflict between evidence on the one hand, and practice and the law on the other. Although Oates et al argue that population-based studies showing a modest increase in mental health consequences are unlikely to help individual women or clinicians, this evidence does provide an important context for a discussion of the therapeutic benefits or otherwise of abortion. What emerges most clearly from the accumulated body of evidence on abortion and mental health is: (a) the primary reasons that most women seek abortion are personal, social and economic rather than relating to mental health concerns; ^{Reference Finer, Frohwirth, Dauphinee, Singh and Moore5,Reference Broen, Moum, Bodtker and Ekeberg6} and (b) there is no body of evidence that would lead a reasonable person to conclude that the provision of abortion mitigates the mental health risks of abortion. Under these circumstances, there is a clear need to develop more comprehensive and realistic criteria for the provision of abortion with these criteria recognising the range of social, economic, personal and related factors that lead women to seek abortions, and (we conjecture) doctors to authorise these procedures.

Third, regarding counselling and support, both Casey and Oates et al pick up on the theme of the need for counselling, although from different perspectives. Whereas Casey emphasises the obligations our findings impose on clinicians and others to inform patients and treat risk, Oates et al are more cautious and emphasise the dangers of mandatory procedures, and argue that the evidence is not strong enough to mandate either advice or treatment. We are inclined to agree with Oates et al about this matter, and we think that it would be premature on the basis of the available evidence to present strong claims about the iatrogenic effects of abortion. At the same time, we believe that there is now a strong case for conducting randomised controlled trials of the extent to which various forms of advice, counselling and support mitigate any mental health risks of abortion. The introduction of good randomised controlled trials could do much to mitigate the generally parlous state of the literature on abortion and mental health.

Finally, we would like to thank the authors for their thoughtful comments, and we were very grateful for the fact that both sets of commentators avoided the tendency to rehearse the usual set of reasons why no useful conclusions can be drawn from observational studies of abortion and mental health.