On Marquis's future of value account, "what makes it wrong to kill those individuals we all believe it is wrong to kill, is that killing them deprives them of their future of value" (1,2). Recently Carson Strong (3,4), Don Marquis (1), and I (5,6) have been arguing about a set of supposed counterexamples to the future of value account proposed by Strong, involving either a terminally ill patient or an individual severely and permanently cognitively impaired. Strong argues that it would be wrong to kill those individuals despite their not having a future of value like ours. I have argued that there are some serious interpretative problems with both Marquis's concepts of "future like ours" and "future of value" and with Strong's counterexample (5); and that, on a charitable interpretation that sets aside those problems, Strong's counterexamples fail because they involve burdening and ultimately unacceptable moral claims in violation of basic principles of equality: they involve, to put it simply, the claim that the individuals in the counterexamples do not have a valuable future.

I have distinguished (6) between a narrow interpretation of "future of value", according to which those individuals do not have a future of value; and a liberal interpretation of "future of value", which allows us to regard the future of terminally ill patients and severely cognitively impaired individuals as valuable. And I have argued that we must not interpret "future of value" as narrowly as Strong proposes - so that we can avoid his violation of basic principles of equality; and that a more liberal interpretation has the consequence that it will enlarge the domain of wrongful killing: we should welcome this outcome.

Marquis has responded by also distinguishing between a narrow and broad interpretation of his future of value account; arguing that, on the narrow view, Strong's terminally ill counterexample does not work - because the terminally ill patient does have a future of value qualitatively like ours, just shorter. But Marquis concedes that Strong's second counterexample, involving a severely and permanently cognitively impaired individual, does succeed - because on the narrow view the cognitively impaired individual does not have a future of value like ours. This counterexample would not work against a broader view of future of value, concludes Marquis. "According to the broad view, one has a future of value just in case, if not killed, one's future will consist, on balance, of experiences one will value" (1).

While I argue (6) we should adopt the more liberal interpretation of "future of value", Marquis opts for the narrow view, because "the narrow version does not even suggest that killing rabbits or mosquitoes may be wrong" (1). Here I argue, contra Marquis, that we should instead pursue the broader more liberal interpretation of "future of value". There are overwhelming reasons for it: first, the broad view does not violate basic principles of equality; second, the broad view is not subject to Strong's severely and permanently cognitively impaired counterexample; third, the broad view, in widening the domain of wrongful killing, is ethically preferable.

Let me emphasize the sorts of problems involved in endorsing the narrow view: the narrow view cannot account - by Marquis's own admission - for the wrongness of killing severely and permanently cognitively impaired individuals; and it therefore cannot account for the wrongness of killing foetuses that will develop into severely and permanently cognitively impaired individuals. Those foetuses, according to the narrow view, do not have a future of value. So they might be killed - or, at least, their killing is not wrongful. That is going to represent a substantial exception to the general rejection of abortion that Marquis (2) put forward with his "future of value" account. It is wrong to kill some foetuses and it is not wrong to kill other foetuses. What's the difference? It is wrong to kill foetuses that have a "future of value"; while it is not wrong to kill foetuses that do not have a "future of value". This emphasizes in what sort of intractable moral territory the narrow view forces us: as I have argued before (6), we are forced into the nasty business of evaluating futures.

Now, it might be objected that this dirty work needs to be done. After all, recognising the differences between severely cognitively impaired individuals and healthy ones acknowledges the tragedy for those born with such impairments and their families. Those differences, the objection goes, must be emphasized. And that is painfully true. But the question is whether this difference should be emphasized in terms of moral value. That's where I think that the narrow view of "future of value" is in violation of basic principles of equality. We may be as brutal as to say that the lives of severely cognitively impaired individuals are qualitatively inferior to ours. And I agree that such brutality is necessary, as a form of respect towards the lives of the severely cognitively impaired. But that is still not the same as saying that those lives are morally less valuable than our lives. It is not the same as saying that it is wrong to kill some foetuses and not wrong to kill others. And it is in this respect that I think the broad view should be preferred to the narrow view.

A note of clarification: it might be thought that I have overstated my argument by emphasizing that the narrow view would imply that it is wrong to kill some foetuses and not wrong to kill other foetuses. It might be proposed that, rather, all that the narrow view implies is that killing some foetuses is morally worse than killing other foetuses; that there are differences in the various moral evaluations of the different killings is not the same as saying that some killings are permissible while others not. Those who might find it an unacceptable consequence of the narrow view that some foetuses may be killed while others may not, could on the other hand live with the weaker claim that the narrow view implies different moral evaluations for different killings. I agree that this latter proposal is importantly different and in this respect not as problematic, but I think that the narrow view implies the former claim, namely that killing some foetuses is wrong while killing others is not wrong. To see this, take Marquis's statement of the narrow view:

"Let us call our (ie, yours and mine, readers) futures of value `p- futures of value`. P-futures are the kind of future lives that can be characterised as the lives of persons. I have a p-future. The fetus I once was had a p-future. (Note that this claim is a simple consequence of the way 'future of value' was defined and well-known facts.) According to what I shall call `the narrow view` valuable futures are futures like ours as long as they are p-futures of value... The severely retarded human beings to which Strong refers do not have p-futures of value" (1)

Don has a p-future and it is therefore wrong to kill Don. The foetus that Don once was had a p-future and it was therefore wrong to kill that foetus. On the other hand John, a severely and permanently cognitively impaired 40-year-old, does not have a p-future. The foetus that John once was did not have a p-future either. Therefore it is not wrong to kill John now, and forty years ago it would not have been wrong to kill the foetus that John then was. So the narrow view does imply that it is wrong to kill some individuals while it is not wrong to kill others. And it does imply that it is wrong to kill some foetuses while it is not wrong to kill others.

So the narrow view allows for too much killing, I have argued. But Marquis is rather worried that the broad view does not allow for enough killing - that's his reason for sticking to the narrow view. I think, contra Marquis, that we should welcome the restrictions of the broad view. Here I should first of all caution about the difficulty in evaluating the future of value account against the worry that it does not allow for enough killing. Recall the statement of the future of value account with which we started: "what makes it wrong to kill those individuals we all believe it is wrong to kill, is that killing them deprives them of their future of value" (1). The account is a bit more specific than just the wrongness of any killing. The future of value account is, specifically, about the wrongness of killing a certain class of individuals: those individuals we all believe it is wrong to kill.

In assessing the worry that the broad view does not allow for enough killing, we need to know who those individuals we all believe it is wrong to kill are. I guess Don Marquis is one of them. I am probably one too, as are most if not all the readers that Marquis addresses. Already when we move to the foetuses that the aforementioned individuals once were, it becomes trickier. But that's ok, because after all that was exactly the point of Marquis's original argument (2): not to have to deal directly with the status of the foetus, but rather with the much less controversial status of the future individuals in which foetuses normally develop. But what about John? Do we all believe it is wrong to kill John? Amongst philosophers it is perhaps unsurprising that we don't all agree. (7,8) But it seems that in society at large there is also no consensus.

Why is this important? Individuating which individuals the future of value account of the wrongness of killing refers to is important because Marquis's reason to choose the narrow view against the broad view is that "the narrow version does not even suggest that killing rabbits or mosquitoes may be wrong" (1). If society cannot agree on John, it certainly cannot agree on the killing of rabbits. Are rabbits individuals we all believe it is wrong to kill? No, they are not - it's a simple statistical fact about human opinion that we don't all believe it is wrong to kill rabbits, nor do we all believe it is not wrong to kill rabbits. In this respect, there is a simple solution about Marquis's worry with rabbits, and therefore a simple rebuttal of his argument in favour of the narrow view: if we should choose the narrow view over the broad view because the narrow view, as opposed to the broad view, does not suggest that killing rabbits may be wrong, then that is no reason to choose the narrow view over the broad view simply because the future of value account is not about rabbits; and therefore neither the narrow view of the future of value account nor the broad view of the future of value account suggest that it may be wrong to kill rabbits simply because the future of value account does not talk about rabbits.

In one sense, then, we are already finished: we have given overwhelming reasons against the narrow view; and we have refuted the stated reason against the broad view. But it might be objected that we should not be so strict in the application of the future of value account; after all, the objection might go, the account is only interesting in so far as it deals with more individuals than just those about whom we all agree - there might not be many of those around, I am afraid. Therefore we should be charitable about the future of value account, and at least also evaluate it as a general account of the wrongness of killing, and not only as a particular account of the wrongness of killing those individuals that we all believe it is wrong to kill.

Let us pay our dues to the principle of charity and evaluate the future of value account also as a general account of the wrongness of killing. What should we then say about the worry that the broad view of the future of value account does not allow for enough killing because it suggests that killing rabbits might be wrong? I think we should just embrace this worry, and consider this consequence of the future of value account as a positive upshot. Indeed, progress in animal ethics might suggest that a general account of the wrongness of killing might have to say something about non-human animals. A particular account of the wrongness of killing only those individuals that we all believe it is wrong to kill does well to avoid the intricate questions of animal ethics - as we said, it is a statistical fact about human opinion that there is no agreement on that point in either philosophical circles or society at large. But a general account of the wrongness of killing has to say something about non-human animals, especially now that animal ethics has earned its place within moral philosophy. So it is a further advantage of the broad view that it accounts for the wrongness of killing animals.

Someone might be worried that we have gone too far: they might accept that including some non-human animals amongst those to which the account applies might be a welcome consequence, but object that the problem for the future of value account arises when it can no longer distinguish between the wrongness of killing an healthy adult human being and the wrongness of killing a rabbit. So that the application of the future of value account to non-human animals might be seen as a reductio of the original argument. This reductio could take two forms:

1) future of value reasoning shows that it is wrong to kill non-human animals. But it is absurd to think that it is wrong to kill non-human animals. So future of value reasoning must be flawed.

2) Future of value reasoning shows that it is just as wrong to kill a non-human animal as it is to kill a human. It is absurd to think that it is just as wrong to kill a non-human animal as it is to kill a human. So future of value reasoning must be flawed.

We have already addressed 1. Two points here about 2: first, many might actually want to go that far and welcome an account that does not distinguish between the wrongness of killing humans and the wrongness of killing non-human animals. Looking at the animal ethics literature suggests that an account such as the one 2 criticises would not be taken to be going particularly far (7,8,9).

Secondly, I don't think that the broad view of the future of value account must imply that there is no moral difference between killing humans and killing animals. Recall the statement of the broad view: "According to the broad view, one has a future of value just in case, if not killed, one's future will consist, on balance, of experiences one will value" (1). That seems to embrace at least some non-human animals (10). But it does explicitly refer to the subjective evaluation of one's experience. The point about the broad view is, indeed, that we are no longer evaluating and comparing futures and experiences across different individuals; it is enough that the one individual values (or would value) her future experiences.

This still leaves room for various possibilities, all of them short of the view that there is no moral difference between killing humans and killing animals. It may be that the sense in which humans value their experiences is not available to non-human animals. That is, for example, where many introduce consciousness: it is not as if non-human animals do not have phenomenal experiences (think of pain); it is rather that their experiences are not conscious or self-conscious the way human experience is11. One might think that this is a relevant moral difference: not enough to deny that it is wrong to kill some non-human animals (because they have the experience of pain, for example); but enough to deny that there is no significant moral difference between killing humans and killing animals. Alternatively, one classic strategy (from Mill's higher-lower pleasure distinction (12)) is to distinguish between the quality of human experience and the quality of animal experience: again, that might suggest that there is a moral difference between killing humans and killing animals. Still, killing animals would not be morally irrelevant. So there is no reason to think that the broad view of the future of value account will lead to conclusions about killing animals that some might hold to be implausibly demanding.

Summing up, I have argued against Marquis's argument for the narrow view over the broad view and for my earlier suggestion6 that we embrace the more liberal version of the future of value account. This I have motivated by raising some serious problems with the narrow view and by arguing that the implications of the broad view, in particular about the killing of non-human animals, are not problematic and should, rather, be welcomed.

References

1) Marquis D. Strong's objections to the future of value account. J Med Ethics (forthcoming).

2) Marquis D. Why abortion is immoral. J Philos 1989;86:183-202.

3) Strong C. A critique of 'The best secular argument against abortion'. J Med Ethics 2008;34:727-31.

4) Strong C. Reply to Di Nucci: why the counterexamples succeed. J Med Ethics 2009;35:326-7.

5) Di Nucci E. Abortion: Strong's counterexamples fail. J Med Ethics 2009;35:306-7.

6) Di Nucci E. On how to interpret the role of the future within the abortion debate. J Med Ethics 2009;35:651-52.

7) Singer P. Practical Ethics. Cambridge: Cambridge University Press, 1979.

8) Singer P. Animal Liberation. New York: New York Review Book, 1975.

9) Regan T. The Case for Animal Rights. Berkeley: University of California Press, 1984.

10) Allen C. Animal Pain. Nous 2004;38:617-43.

11) Carruthers P. Brute Experience. J Philos 1989;86:258-269.

12) Mill J.S. Utilitarianism. 1863.

Conflict of Interest:

None declared

[This is an elaborated version of a blog on May 13, 2014: http://blog.practicalethics.ox.ac.uk/2014/05/is-home-birth-really-as-safe- as-hospital-birth-woman-centred-care-vs-baby-centred-care/#more-8493] Imagine that you and your partner are having a baby in hospital. Tragically something goes wrong unexpectedly during birth and the baby is born blue. He urgently needs resuscitation if there is to be a chance of preventing permanent severe brain damage. How long would it be reasonable for doctors to wait before starting resuscitation? 15 minutes? 5 minutes? 1 minute? What would be a reasonable excuse for delaying the commencement of resuscitation? They wanted to get a cup of coffee? The mother wanted to hold the baby first? The mother had catastrophic bleeding and this needed urgent attention? If it were my baby, I would not want any delay in starting resuscitation. And the only justification for delaying resuscitation would be some more serious, more urgent problem for another patient, such as the mother. Obstetrics is renowned for the sudden catastrophic events that can occur in a previously low risk healthy pregnancy. These may result in serious morbidity or mortality. Yet when people choose homebirth, delay is precisely what they choose. It is simply not possible to start advanced resuscitation in the home within minutes. And their reason is not typically some relevant competing health concern that necessitates delivery at home. Choosing home birth is choosing delay if some serious problem arises which requires immediate resuscitation. If you could know for certain which babies were going to be born needing resuscitation, and which weren't, you could deliver safely at home. But you can't know for certain. Birth can be a dangerous, unpredictable time. About 0.3% of babies are born with serious medical problems. So someone might argue that the chances of something going wrong at home or in hospital is sufficiently low for it not be necessary to further reduce an already small risk. However, given that around 700 000 women give birth in England and Wales every year, this translates to a large number with serious medical problems. Surely we should try to minimise avoidable severe life long disability? If the numbers were very small, there may be reasons of cost-effectiveness not to try to reduce these to the very minimum. But even if the risk were 1/100,000, when it is your baby that is affected, it is a personal tragedy. In a recent paper, we argued that a neglected outcome of choice of birth place was long term, significant avoidable disability. [1] We argued that there are reasons to believe that the risk of this outcome is higher at home than in hospital, though accurate figures are not available. We called for more research to be done on what we call "future disability." The National Institute for Clinical Excellence has issued a draft guideline for consultation entitled "Intrapartum care: care of healthy women and their babies during childbirth" , [2] Two key recommendations are: * Advise low-risk multiparous women to plan to give birth at home or in a midwifery-led unit (freestanding or alongside). Explain that this is because the rate of interventions is lower and the outcome for the baby is no different compared with an obstetric unit. * Advise low-risk nulliparous women to plan to give birth in a midwifery- led unit (freestanding or alongside). Explain that this is because the rate of interventions is lower and the outcome for the baby is no different compared with an obstetric unit, but if they plan birth at home there is a small increase in the risk of an adverse outcome for the baby. The BBC reported, "Home births were just as safe as other settings for low -risk pregnant women who already had at least one child." [3] This guidance is based on the largest prospective cohort study of the outcome of place of birth, The Birthplace Study. This followed nearly 65000 women. It found, "There were 250 primary outcome events and an overall weighted incidence of 4.3 per 1000 births (95% CI 3.3 to 5.5). Overall, there were no significant differences in the adjusted odds of the primary outcome for any of the non-obstetric unit settings compared with obstetric units. For nulliparous women, the odds of the primary outcome were higher for planned home births (adjusted odds ratio 1.75, 95% CI 1.07 to 2.86) but not for either midwifery unit setting. For multiparous women, there were no significant differences in the incidence of the primary outcome by planned place of birth." [4] This study found that home birth was riskier for first time mothers, but not for low risk women who had had at least one child already. But does that make home birth as safe as hospital birth for multiparous women? This study did not examine the rates of severe long term disability arising from choice of birth place, which we argued is ethically the most relevant outcome. It tells us nothing about the outcome that matters most. What did this study measure as primary outcomes? It adopted a composite measure (that is it pooled a bunch of different outcomes): "stillbirth after start of care in labour, early neonatal death, neonatal encephalopathy, meconium aspiration syndrome, brachial plexus injury, fractured humerus, or fractured clavicle."[4] Why did it do that? Although they studied nearly 65 000 women these adverse events are relatively rare - only 250 occurred in the study. To get statistical significance, they needed to combine these. But these outcomes are very different ethically. Two outcomes included were a fractured clavicle or humerus, but these are fully correctable. They are completely different to permanent severe brain damage. Moreover, the authors don't even grade the severity of these primary outcomes. Encephalopathy can be mild to severe. Some forms will not be associated with any long term severe disability. Taking these factors into account, this study is just underpowered to detect the outcome that matters most ethically - long term avoidable severe disability. The authors state that neonatal encephalopathy and meconium aspiration syndrome are thought to be associated with long-term morbidity and they accounted for 87% of the non-fatal events of the birthplace composite primary outcome . [5] Yet both of these can have variable long term outcomes. Moreover, the composite outcome employed by the Birthplace Study do not include all causes of long term disability. What is needed are long term follow up studies so we know the risks for serious permanent damage of different places of birth. The authors admit as much: "The weaknesses of the study include the use of a composite primary outcome measure, because of the low event rates for individual perinatal outcomes. We cannot rule out the possibility that the use of a composite may have concealed important differences in outcomes between planned places of birth, such as less severe outcomes in a particular setting. However, examination of the distribution of outcomes by planned place of birth did not suggest that this was the case. In addition, although many of the outcomes included in the composite are likely to reflect problems which occur during labour and birth, their long term implications for the baby are uncertain. For example, although moderate and severe neonatal encephalopathy are associated with development of cerebral palsy and long term morbidity, mild encephalopathy has not been associated with detectable longer term impacts." [6] These outcome measures are reasonable surrogates of future disability. But even if there is no difference in these surrogates, this does not tell us of the rates of future child disability. Such studies would have to be very large and conducted over a long period of time at great cost. There are other reasons to believe that this study does not sufficiently inform us of risks of future child disability. They took a random sample of hospitals providing obstetric care. But there are reasons to believe that hospitals will differ in their capacity to offer rapid resuscitation, just as delivery at home cannot offer this to the highest standard possible. A hospital with a neonatal intensive care facility, and highly trained staff readily available, will be able to offer more effective resuscitation than smaller regional centres. (It is for this reason that our friends who are doctors have their babies in hospitals with access to neonatal intensive care). Even if there were no difference between home and hospital with respect to future disability, this would tell us little about the comparison between best obstetric/paediatric care and best home birth care. Two of the authors of the Birthplace Study in response to our article make the point (which is echoed by the BBC and NICE) that hospital based care can be associated with higher rates of intervention, such as Caesarean, which exposes the mother and future pregnancies to risk. [5] For example, the BBC reports: "The guidelines from NICE - the National Institute for Health and Care Excellence - say a home birth may be just as safe for low-risk pregnancies. Hospital labour wards with doctors should be for difficult cases, it says. Otherwise there is a danger of over-intervention, according to NICE." [3] This is a valid concern. However, if true, it would hardly be a good a reason to have home birth. It would be a reason not to over-intervene in hospital! To be sure, getting hospitals to change their practices (some shaped more by concerns about efficiency, or simply bad institutional habits, than optimal health) may not be easy and does indeed need sustained attention to correct. But our focus should be best obstetric care, not second best care. In addition, the harms experienced by women in hospital are of a different kind to those experienced by a baby from hypoxia. For example, treatable infection is very different to life long severe disability. The NICE guideline talks proudly throughout of "Woman-centred care." At very least, this should be "Woman and Baby-centred care". The authors of The Birthplace Study responded to us by saying that although a composite outcome measure could potentially have disguised important differences in outcomes between birth settings, it is pure speculation to suggest that the effect on longer-term disability would be likely to favour hospital birth. [5] But we did not conclude this. We stated that other evidence suggests that long-term disability may be lower with hospital birth. We further suggested that research is required to document the prevalence of long term disability associated with different birth place choices.

Other issues that may impact on whether the primary outcome provides a measure of long term disability include questions about why in this low -risk population of women were 20 of the 32 deaths in the home or Freestanding Midwifery Unit groups? And why did 1 in 5 of the women in the obstetric units group have complications at the onset of birth compared to the 5-7% of the other low-risk women? [7] In addition, what were the number & proportion transported in labour? What was the average, minimum and maximum time interval from the decision to transport the maternal and fetal patient to in-hospital intervention? [8]

Other criticisms of this influential study include that there was a failure to include the lifetime costs for support of severely disabled children, estimated to be ?5 million per child. [9]

In Dahlys' analysis of the published research paper (or supplementary data) he notes that the paper doesn't provide outcome information specifically for women planning a homebirth who wound up transferring to hospital, which is unfortunate. [10]

Tuteur reports that the list of study exclusion criteria employed in this study was far more restrictive than the actual exclusion criteria for homebirth in the UK. Therefore, a substantial proportion of the women who actually had a homebirth were excluded from the study even before it began. Of the 18,269 low risk women planning homebirth at the start of the study, 1346 (7.4%) were excluded from the study despite the fact that they went on to have a homebirth under the auspices of the National Health Service. [11] She notes that the supplemental material includes the outcomes for those women who had homebirths but did not meet the very restrictive criteria of the study. [12] She compares them with the low risk women electing hospital birth who also failed to meet the more restrictive criteria for inclusion in the study. Homebirth doubled the risk of an adverse outcome for both nulliparous women and multiparous women in the "higher risk" group. [11]

She states that this is further confirmation of the central finding of the study. Homebirth doubles the risk of adverse perinatal outcomes. Therefore, the claim of the Royal College of Midwives and other homebirth advocates that ".., [f]or women having a second or subsequent baby ... homebirths appear to be safe for the baby" is not true. [11]

Homebirth is safe only when nothing goes wrong. Since there is no way to predict with complete accuracy whether something is going to go wrong, and facilities, equipment and expertise for complete management of mother and baby are compromised or lacking in the home, homebirth is riskier than hospital birth. [11] The fact is, we just don't know if best practice home birth is as safe with respect to future disability as best practice hospital birth. There is not enough research. There is evidence it is more dangerous for first time mothers (though again this is not in relation to the outcome of future disability) and there are good reasons based on understanding of the risks of child birth and the interventions available to concerned whether there will be some elevated risk at least related to delay in resuscitation associated with transfer. In our article, we reviewed a variety of arguments and existing evidence that suggest risk associated with home birth of future disability will be higher. For example, a meta- analysis which included 12 studies and 500,000 planned homebirths in healthy low risk women showed neonatal mortality tripled. [13] But good direct research has not been done and it would require extremely large studies over many years. The elevation of future child risk associated with home birth, if it exists, is likely to be small. But small risks of tragic outcomes ought to be minimised. For this reason, the Royal College of Obstetricians recommends that home birth only be considered "provided transport arrangements are in place for hospital transfer in the event of an emergency ". [3] But what kind of "transport arrangement" would minimise risk? An ambulance on stand-by? Whatever transport is arranged, it will involve an inevitable delay compared with delivery in a tertiary centre. And minutes can matter for the baby. Some might argue this is excessively risk averse. But the risk of injury to the child from not wearing a seat belt on any single trip, or even over a year is extremely small. Nonetheless, we believe that it is right to minimise this risk by putting on a seat belt. Women should have choice over their place of delivery. But they should make that choice in full knowledge of the facts, arguments and gaps in evidence. There may be reasons to do with cost and distributive justice that preclude every woman and child being offered the best care possible. But where there is a choice, people should at very least know what is known, and what is not known, about the risks and benefits of each option. References: 1 de Crespigny L, Savulescu J. Homebirth and the Future Child. Journal of Medical Ethics Published online first: 22 January 2014. Doi: 10.1136/medethics-2012-101258.

2 NICE guideline. Intrapartum care: care of healthy women and their babies during childbirth. Draft for consultation, May 2014. P10. http://www.nice.org.uk/nicemedia/live/13511/67644/67644.pdf

3 Roberts M. Labour wards not for straightforward births' says NICE. BBC News, Health. 13 May 2014. http://www.bbc.co.uk/news/health-27373543 (accessed 24 June 2014)

4 Birthplace in England Collaborative Group. Perinatal and maternal outcomes by planned place of birth for healthy women with low risk pregnancies: the Birthplace in England national prospective cohort study. BMJ 2011; 343: d7400: 1.

5 Hollowell J, Brocklehurst P. Homebirth and the future child: factual inaccuracies in commentary on the Birthplace study. JME [e- letter] 28 April. http://jme.bmj.com/content/early/2013/10/08/medethics- 2012-101258/reply

6 Birthplace in England Collaborative Group. Perinatal and maternal outcomes by planned place of birth for healthy women with low risk pregnancies: the Birthplace in England national prospective cohort study. BMJ 2011;343:d7400: 4.

7 Falconer T. Re: Perinatal and maternal outcomes by planned place of birth for healthy women with low risk pregnancies: the Birthplace in England national prospective cohort study. JME [e-letter] 1 December 2011. http://www.bmj.com/content/343/bmj.d7400?tab=responses

8 Arabin B. Re: Perinatal and maternal outcomes by planned place of birth for healthy women with low risk pregnancies: the Birthplace in England national prospective cohort study. JME [e-letter] 18 January 2012. http://www.bmj.com/content/343/bmj.d7400?tab=responses

9 Svensson G. Re: Perinatal and maternal outcomes by planned place of birth for healthy women with low risk pregnancies: the Birthplace in England national prospective cohort study. JME [e-letter] 28 November 2011. http://www.bmj.com/content/343/bmj.d7400?tab=responses

10 Dahly. Are homebirths really risky? Statistical Epicimiology. Published 4 December 2011. http://statisticalepidemiology.org/are- homebirths-really-more-risky/ (accessed 24 June 2014)

11 Tuteur A. Birthplace study yields additional disturbing information. The Skeptical OB. Published 19 December 2011. http://www.skepticalob.com/2011/12/birthplace-study-yields-additional.html (accessed 24 June 2014)

12 Hollowell J, Puddicombe D, Rowe R, et al. The birthplace national prospective cohort study: perinatal and maternal outcomes by planned place of birth. Birthplace in England research programme. Final report part 4. Published November 2011. http://www.netscc.ac.uk/hsdr/files/project/SDO_FR4_08-1604-140_V03.pdf (accessed 24 June 2014)

13 Wax JR, Lee Lucas F, Lamont M, et al. Maternal and newborn outcomes in planned home birth vs planned hospital births: a metaanalysis. Americal Journal of Obstetrics and Gynecology, Volume 203, Issue 3, 243.e1 -243.e8; September 2010.

Conflict of Interest:

None declared

Conflict of Interest:

None declared

We thank Dr. Post for his response to our brief report, "Conflict of interest in online point-of-care clinical support websites" (J Med Ethics doi:10.1136/medethics-2013- 101625). We additionally commend UpToDate for many of its policies that protect the integrity of the educational material presented. In light of Dr. Post's comments, we recommend further clarification in the UpToDate conflict of interest policy, particularly in regards to the role of the deputy editor versus the section editor.

As per the UpToDate conflict of interest policy "Deputy Editors may not accept funds from "educational" arms of health industry organizations (such as speaker's bureaus)." We commend this policy and believe as mentioned in our article, that a supervising editor with no conflicts of interest is necessary. In his response, Dr. Post describes the role of the Deputy Editor as one that prevents "inappropriate material from being published in UpToDate topic reviews." This is not, however, mentioned in the UpToDate conflict of interest policy. We believe that this should be added to the conflict of interest disclosure statement. Additionally, the role of the section editor should be discussed. Otherwise, the shared potential conflicts of the author and section editor may remain a concern to the reader. Likewise, the review process described in Dr. Post's final point should additionally be disclosed in the conflict of interest section to improve transparency.

Even if such an accountable review process with multiple named reviewers takes place, it is imperative that readers have such information to support a critical evaluation of the material. This is a flaw in the conflict of interest policy in Dynamed that both our report and Dr. Post's response describe, preventing verification of potential documented and undocumented conflicts. While minimizing the number of conflicts of interest remains essential for unbiased scientific writing, this goal depends on the clear disclosure of conflict-of-interest policies.

Conflict of Interest:

None declared