Ethics in Medicine
Medical ethics will appeal to many temperaments: to the thinker and to the doer; to the philosopher and to the woman or man of action. It deals with some of the big moral questions: easing death and the morality of killing, for example. It takes us into the realm of political philosophy. How should health care resources, necessarily limited, be distributed, and what should be the process for deciding? It is concerned with legal issues. Should it always be a crime for a doctor to practise euthanasia? When can a mentally ill person be treated against his will? And it leads us to the major world issue of the proper relationships between rich and poor countries.
(1) Ethics in Medicine
On why medical ethics is exciting
‘I don’t have a lot of time for thinking about things’ he said with a defensive edge creeping into his tone. ‘I just scatter my hundreds and thousands before the public. Philosophy I leave to the drunks.’
(Ice-cream stall owner, in Malcom Pryce, Aberystwyth Mon Amour)
Modern medical science creates new moral choices, and challenges traditional views that we have of ourselves. Cloning has inspired many films and much concern. The possibility of making creatures that are part human and part from some other animal is not far off.
Reproductive technologies raise the apparently abstract question of how we should think about the interests of those who are yet to be born – and who may never exist. This question leads us beyond medicine to consider our responsibilities towards the future of mankind.
Medical ethics ranges from the metaphysical to the mundanely practical. It is concerned not only with these large issues but also with everyday medical practice. Doctors get caught up in people’s lives, and ordinary life is full of ethical tensions.
An elderly woman with a degree of dementia suffers an acute life-threatening illness. Should she be treated in hospital with all the drugs and technology available; or should she be kept comfortable at home? The family cannot agree. There is nothing in this case likely to hit the headlines; but, as Auden’s Old Masters knew, the ordinary is what is important to most of us, most of the time. In pursuing medical ethics we must be prepared to grapple with theory, allowing time for speculation and the use of the imagination. But we must also be ready to be practical: able to adopt a no-nonsense, down-to-earth, approach.
My own interest in medical ethics started at the theoretical end of the spectrum when studying for a degree that included philosophy. But when I went to medical school my inclination turned more to the practical. Decisions had to be made, and sick people had to be helped. I trained as a psychiatrist and the ethics remained only as a thin interest squeezed into the corners of my working life as doctor and clinical scientist. As my clinical experience grew so I became increasingly aware that ethical values lie at the heart of medicine. Much emphasis during my training was put on the importance of using scientific evidence in clinical decision-making. Little thought was given to justifying, or even noticing, the ethical assumptions that lay behind the decisions. So I moved increasingly towards medical ethics, wanting medical practice, and patients, to benefit from ethical reasoning. I enjoy the highly theoretical, and I like to pursue reasoning back towards the general and the abstract; but I keep an eye to what makes a difference in practice. I discuss the philosophical minefield of the non-identity problem, for example, because I believe it is relevant to decisions that doctors, and society, need to take.
The philosopher and cultural historian, Isaiah
The fox represents those who pursue many ends, often unrelated and even contradictory, connected, if at all, only in some de facto way, [who] lead lives, perform acts, and entertain ideas that are centrifugal rather than centripetal . . . seizing upon the essence of a vast variety of experiences . . . without . . . seeking to fit them into . . . any one unchanging, all-embracing, . . . unitary inner vision.
Are you a hedgehog or a fox?
I am a fox, or at least would like to be. I admire the intellectual rigour of those who try to produce a unitary vision, but I prefer the rich, contradictory, and sometimes chaotic visions of
I believe that medical ethics is essentially a rational subject: that is, it is all about giving reasons for the view that you take, and being prepared to change your views on the basis of reasons. That is why one paper, in the middle of all papers, is a reflection on various tools of rational argument. But although I believe in the central importance of reasons and evidence, even here the fox in me sounds a note of caution. Clear thinking, and high standards of rationality, is not enough. We need to develop our hearts as well as our minds. Consistency and moral enthusiasm can lead to bad acts and wrong decisions if pursued without the right sensitivities. The novelist, Zadie Smith, has written: There is no bigger crime, in the English comic novel, than thinking you are right. The lesson of the comic novel is that our moral enthusiasms make us inflexible, one-dimensional, flat. This is a lesson we need to take into any area of practical ethics, including medical ethics.
What better place to start this tour of medical ethics than at the end, with the thorny issue of euthanasia?
* Professor Doctor Mike Gaze
(2) Ethics in medicine
Euthanasia: good medical practice, or murder?
Good deeds do not require long statements; but when evil is done the whole art of oratory is employed as a screen for it.
The practice of euthanasia contradicts one of the oldest and most venerated of moral injunctions: ‘Thou shalt not kill’. The practice of euthanasia, under some circumstances, is morally required by the two most widely regarded principles that guide medical practice: respect for patient autonomy and promoting patient’s best interests. In the
Outline of the requirements in order for active euthanasia to be legal in the
1. The patient must face a future of unbearable, interminable suffering.
2. The request to die must be voluntary and well-considered.
3. The doctor and patient must be convinced there is no other solution.
Playing the Nazi card
There is a common, but invalid, argument against euthanasia that I call ‘playing the Nazi card’. This is when the opponent of euthanasia says to the supporter of euthanasia: ‘Your views are just like those of the Nazis’. There is no need for the opponent of euthanasia to spell out the rhetorical conclusion: ‘and therefore your views are totally immoral’.
Let me put the argument in a classic form used in philosophy and known as a syllogism:
Premise 1: Many views held by Nazis are totally immoral.
Premise 2: Your view (support for euthanasia under some circumstances) is one view held by Nazis.
Conclusion: Your view is totally immoral.
This is not a valid argument. It would be valid only if all the views held by Nazis were immoral.
I will therefore replace premise 1 by premise 1* as follows:
Premise 1*: All views held by Nazis are totally immoral. In this case the argument is logically valid, but in order to assess whether the argument is true we need to assess the truth of premise 1*.
There are two possible interpretations of premise 1*. One interpretation is a version of the classic false argument known as argumentum ad hominem (or bad company fallacy): that a particular view is true or false, not because of the reasons in favour or against the view, but by virtue of the fact that a particular person (or group of people) holds that view. But bad people may hold some good views, and good people may hold some bad views. It is quite possible that a senior Nazi was vegetarian on moral grounds. This fact would be irrelevant to the question of whether there are, or are not, moral grounds in favour of vegetarianism. What is important are the reasons for and against the particular view, not the person who holds it. Hitler’s well-known vegetarianism, by the way, was on health, not on moral, grounds.
The other, more promising, interpretation of premise 1* is that those views that are categorized as ‘Nazi views’ are all immoral. Some particular Nazis may hold some views about some topics that are not immoral, but those would not be ‘Nazi views’. The Nazi views being referred to are a set of related views, all immoral, that is driven by racism and involve killing people against their will and against their interests. Thus, when it is said that euthanasia is a Nazi view, what is meant is that it is one of these core immoral views that characterize the immoral Nazi worldview. The problem with this argument, however, is that most supporters of euthanasia – as it is practised in the
What is needed is some conceptual clarity.
Clarifying concepts in the euthanasia debate
Let us begin with some definitions. The purpose of these is twofold: to make distinctions between different kinds of euthanasia; and to provide us with a precise vocabulary. Such precision is often important in evaluating arguments and reasons. If a word is used in one sense at one point in the argument, and in another sense at another point in the argument, then the argument may look valid when in fact it is not.
If you study these definitions it will be immediately clear that playing the Nazi card rides roughshod over some important distinctions. The first point is that the term euthanasia, at least as I am suggesting that it should be used, implies that the death is for the person’s benefit. What the Nazis did was to kill people without any consideration of benefit to the person killed. The second point is that euthanasia can be voluntary, involuntary, or non-voluntary.
The third point is that it can be active or passive. Let us start with the first point.
Euthanasia and suicide: some terms
Euthanasia comes from the Greek eu thanatos meaning good or easy death.
X intentionally kills Y, or permits Y’s death, for Y’s benefit.
X performs an action which itself results in Y’s death.
X allows Y to die. X withholds or withdraws life-prolonging treatment.
Euthanasia when Y competently requests death himself, i.e. a competent adult wanting to die.
Euthanasia when Y is not competent to express a preference, e.g. Y is a severely disabled newborn.
Death is against Y’s competent wishes, although X permits or imposes death for Y’s benefit.
Y intentionally kills himself.
X intentionally helps Y to kill himself.
Physician assisted suicide:
X (a physician) intentionally helps Y to kill himself.
Patients’ best interests
Can it be in someone’s best interests to die? I believe it can. The courts believe it can. Most doctors, nurses, and relatives believe it can. The question arises quite frequently in health care. A patient with an incurable and fatal disease may reach a stage where she will die within a day or two, but could be kept alive, with active treatment, for a few weeks more. This situation might occur because the patient gets a chest infection, or because there is a chemical imbalance in her blood, in addition to the underlying fatal disease. Antibiotics, or intravenous fluids, might treat this acute problem although they will do nothing to stop the progress of the underlying disease. All those caring for the patient will often agree that it is in the patient’s best interests to die now rather than receive the life-extending treatment. The decision not to treat is even more straightforward if the patient’s quality of life is now very poor, perhaps because of sustained and untreatable difficulty in breathing – a distressing feeling that is often more difficult to ameliorate than severe pain. If, however, we thought that it was in the patient’s best interests to continue to live, rather than to die within days, we ought to give the life-extending treatment. But we do not think this: we believe it is in her best interests to die now rather than receive the life-extending treatment, because her quality of life, due to the underlying fatal illness, is so poor.
Respecting a patient’s wishes
Most countries that put a value on individual liberty allow competent adults to refuse any medical treatment even if such treatment is in the patient’s best interests; even if it is life-saving. A Jehovah’s Witness, for example, may refuse a life-saving blood transfusion. If doctors were to impose treatment against the will of a competent patient then the doctor would be violating the bodily integrity of the person without consent. In legal terms this would amount to committing a ‘battery’.
Passive euthanasia is widely accepted.
The withholding, or withdrawing, of treatment is widely accepted as morally right in many circumstances. And it is protected in English law. There are two grounds on which it is accepted:
(1) That it is in the patient’s best interests; and
(2) That it is in accord with the patient’s wishes.
Either of these two conditions is sufficient reason to support passive euthanasia.
In common with widespread medical practice, I believe that there are circumstances when it is in a person’s best interests to die rather than to live. I also believe that a competent person has the right to refuse life-saving treatment. Withholding or withdrawing treatment from a patient is justified in either set of circumstances, even though this will lead to death.
If I am right (and the law in
What Dr Cox did
Lillian Boyes was a 70-year-old patient with very severe rheumatoid arthritis. The pain seemed to be beyond the reach of painkillers. She was expected to die within a matter of days or weeks. She asked her doctor, Dr Cox, to kill her. Dr Cox injected a lethal dose of potassium chloride for two reasons:
(1) Out of compassion for his patient, and
(2) Because this is what she wanted him to do.
Dr Cox was charged with, and found guilty of, attempted murder.
(The reason for not charging him with murder was that, given her condition, Lillian Boyes could have died from her disease and not from the injection.)
The judge, in directing the jury, said:
Even the prosecution case acknowledged that he [Dr Cox] . . . was prompted by deep distress at Lillian Boyes’ condition; by a belief that she was totally beyond recall and by an intense compassion for her fearful suffering. Nonetheless . . . if he injected her with potassium chloride for the primary purpose of killing her, or hastening her death, he is guilty of the offence charged [attempted murder] . . . neither the express wishes of the patient nor of her loving and devoted family can affect the position.
This case clearly established that active (voluntary) euthanasia is illegal (and potentially murder) under English common law. It is noteworthy that the patient was competent and wanted to be killed; close and caring relatives and her doctor (as well as the patient) believed it to be in her best interests to die, and the court did not dispute these facts.
The key difference, on which much legal and moral weight is placed, between the case of Dr Cox and the examples of withholding and withdrawing treatment that are a normal and perfectly legal part of medical practice, is that Dr Cox killed Lillian Boyes, and did not simply allow her to die.
Moral philosophers use ‘thought experiments’. These are imaginary and sometimes quite unrealistic situations that tease out and examine the morally relevant features of a situation. They are used to test the consistency of our moral beliefs. The thought experiment that I want you to consider is a case, like the Cox case, of mercy killing.
Mercy killing: the case of the trapped lorry driver
A driver is trapped in a blazing lorry. There is no way in which he can be saved. He will soon burn to death. A friend of the driver is standing by the lorry. This friend has a gun and is a good shot. The driver asks this friend to shoot him dead. It will be less painful for him to be shot than to burn to death.
I want to set aside any legal considerations and ask the purely moral question: should the friend shoot the driver?
There are two compelling reasons for the friend to kill the driver:
1. It will lead to less suffering.
2. It is what the driver wants.
These are the two reasons we have been considering with regard to justifying passive euthanasia. What reasons might you give for believing that the friend should not shoot the driver? I will consider seven reasons.
1. The friend might not kill the driver but might wound him and cause more suffering than if he had not tried to kill him.
2. There may be a chance that the driver will not burn to death but might survive the fire.
3. It is not fair on the friend in the long run: the friend will always bear the guilt of having killed the driver.
4. That although this seems to be a case where it might be right for the friend to kill the driver it would still be wrong to do so; for unless we keep strictly to the rule that killing is wrong, we will slide down a slippery slope. Soon we will be killing people when we mistakenly believe it is in their best interests. And we may slip further and kill people in our interests.
5. The argument from Nature: whereas withholding or withdrawing treatment, in the setting of a dying patient, is allowing nature to take its course, killing is an interference in Nature, and therefore wrong.
6. The argument from Playing God, which is a religious version of the argument from Nature. Killing is ‘Playing God’ – taking on a role that should be reserved for God alone. Letting die, on the other hand, is not usurping God’s role, and may, when done with care and love, be enabling God’s will to be fulfilled.
7. Killing is in principle a (great) wrong. The difference between passive euthanasia and mercy killing is that the former involves ‘allowing dying’ and the latter involves killing; and killing is wrong – it is a fundamental wrong.
How good are these arguments? Let’s consider them one by one.
It is true that in real life we cannot be certain of the outcome. If you rely on argument 1 then you are not arguing that mercy killing is wrong in principle, but instead that in the real world we can never be sure that it will end in mercy. I am happy to accept that we can never be absolutely sure that the shooting will kill painlessly. There are three possible types of outcome:
(a) If the friend does not shoot (or if the bullet completely misses) then the driver will die having suffered a considerable amount of pain – let us call this amount X.
(b) The friend shoots and achieves the intended result: that the driver dies almost instantaneously and almost painlessly. In this case the driver will suffer an amount Y where Y is much smaller than X – indeed Y is almost zero if we are measuring suffering from the moment when the friend shoots.
(c) The friend shoots but only wounds the driver, causing him overall an amount of suffering Z, where Z is greater than X.
It is because of possibility (c), according to argument 1, that it would be better that the friend does not shoot the driver.
We can now compare the situation where the friend does not shoot the driver with the situation where the friend does shoot. In the former case the total amount of suffering is X. In the latter case the amount of suffering is either Y (close to zero) or Z (greater than X).
Thus, by shooting, the friend may bring about a better state of affairs (less suffering) or a worse state of affairs (more suffering). If what is important is avoiding suffering, then whether it is better to shoot or not depends on the differences between X, Y, and Z and the probabilities of each of these outcomes occurring. If almost instantaneous death is by far the most likely result from shooting, and if the suffering level Z is not a great deal more than X, then it would seem right to shoot the driver because the chances are very much in favour that shooting will lead to significantly less suffering.
We can rarely be completely certain of outcomes. If this uncertainty were a reason not to act we would be completely paralysed in making decisions in life. It would be very unlikely; furthermore, that mercy killing in the medical setting (e.g. what Dr Cox did) would lead to more suffering. I conclude that argument 1 does not provide a convincing argument against voluntary active euthanasia.
Argument 2 is the other side of the coin from argument 1, and suffers the same weakness. The question of whether the chance that the driver might survive outweighs the greater chance that he will suffer greatly, and die, depends on what the probabilities actually are. If it is very unlikely that the driver will survive, then argument 2 is not persuasive.
Supporters of argument 2 might counter this conclusion by arguing that the weight to be given to the remote possibility of rescue from the burning lorry should be infinite. In that case, however low the probability of its occurring, the chance should be taken. There are three responses to this argument: first, what grounds are there for giving infinite weight to the possibility of rescue? Second, if we consider that very remote possibilities of rescue justify not shooting then we could equally well conclude that we should shoot. This is because it is also a remote possibility that the bullet, although intended to kill the driver, might in fact enable him to be rescued (e.g. through blowing open the cab door). Third, if argument 2 provides a convincing reason for rejecting mercy killing, it also provides a convincing reason for rejecting the withholding of medical treatment in all circumstances. This is because giving treatment might provide sufficient extension of life for a ‘miracle’ to occur and for the person to be cured and live healthily for very much longer.
The third argument fails because it begs the very question that is under debate. The friend should only feel guilt if shooting the driver were the wrong thing to do. But the point at issue is what is the right and wrong thing to do. If it is right to shoot the driver, then the friend should not feel guilty if he shot him (thus reducing the driver’s suffering). The possibility of guilt is not a reason, one way or the other, for deciding how the friend should act. Rather we first have to answer the question of what is the right thing to do and only then can we ask whether the friend ought to feel guilty.
Argument 4 is a version of what is known as the ‘slippery slope argument’. This is such an important type of argument in medical ethics that I will consider it in more detail later. I will distinguish two types of slippery slope – the logical, or conceptual, slope; and the empirical, or in-practice, slope. The types of reason needed to counter a slippery slope argument depend, as we shall see, on which type of argument is being advanced.
Arguments 5 and 6
The arguments from Nature and from Playing God have, like the slippery slope argument, a more general application in medical ethics. I will consider them in more detail later.
Of all the arguments considered, it is only argument 7 that views killing as wrong in principle.
(3) Ethics in medicine
Is mercy killing wrong in principle?
At this stage we need to get clear what ‘killing’ means. Those who believe that mercy killing, but not the common medical practice of passive euthanasia, is wrong in principle do so on the grounds that mercy killing involves actively causing death rather than failing to prevent it.
(4) Ethics in Medicine
Why undervaluing ‘statistical’ people costs lives
Whether happiness be or be not the end to which morality should be referred – that it should be referred to an end of some sort, and not left in the dominion of vague feeling or inexplicable internal conviction, that it be made a matter of reason and calculation, and not merely of sentiment, is essential to the very idea of moral philosophy . . . (J. S. Mill, London and Westminster Review, 1838)
The cash value of life
In January 1997 Tony Bullimore was attempting to sail round the world in the Vendée Globe race. He had reached the dangerous and cold waters of the Southern Ocean, over 1,500 miles south of the Australian coast, when his boat was capsized by hurricane force winds and enormous waves. He spent four days trapped under its hull before he was rescued in the largest and most expensive such operation ever undertaken by the Australian defence forces. How much money should a civilized society be prepared to spend in order to save a life? Is the answer ‘whatever it takes’, or should there be a limit? When is the chance of success too low even to attempt a costly rescue operation?
Let me pose a more general question. What is the cash value of a human life? This question is a disturbing one to ask but, paradoxically, there are situations where avoiding the question may cost lives, and allocating scarce medical resources is one of them.
There is no health care system in the world that has sufficient money to provide the best possible treatment for all patients in all situations, not even those that spend relatively large sums on health care (see box). New and better treatments are being developed all the time. On average, in the UK, about three new medicines are licensed each month. Almost all have some benefit over existing treatments and some will extend people’s lives. Many of these new medicines are expensive. When is the extra benefit worth the extra cost? This question must be asked by all health care systems, whether private systems, such as ‘managed care’ in the US, or publicly funded systems, such as the British National Health Service. If the best treatment cannot always be provided then choices have to be made. The general question of how our limited health care resources should be distributed is one of the most important in medical ethics. The quality and quantity of thousands of people’s lives will be affected by the answers that we give.
National expenditure on health: examples of some of the wealthier nations
Country - % GDP - per capita purchasing power ($)
Australia - 8.6 - 2085
Canada - 9.3 - 2360
France - 9.4 - 2043
Germany - 10.3 - 2361
New Zealand - 8.1 - 1440
Norway - 9.4 - 2452
United Kingdom - 6.8 - 1510
United States - 12.9 - 4165
Quality of life
Some medical treatments have little or no effect on life-span but improve quality of life: hip replacement for osteoarthritis is an example. One rather deep problem that faces us in thinking about the right way to distribute health resources is how we compare and evaluate the relative importance of improving quality of life vis-à-vis extending it. I am not going to tackle this issue, nor the problems associated with the measurements of quality of life in the first place. I will focus exclusively on life-extending treatments since there are more than enough problems in thinking about allocating resources to these treatments alone. There are many examples of life-extending treatments. Surgery for appendicitis extends life because without such surgery most people would die.
Breast cancer screening can extend life because early detection and treatment can increase life-span. High blood pressure increases the risk of death from heart attack and stroke. Treatment that lowers blood pressure reduces, although it does not eliminate, this risk. Renal dialysis keeps those people alive whose kidneys no longer function adequately. Each year of dialysis is a year more life.
In control of a budget
Imagine that you are in charge of a health service for a particular population. You have a limited budget – you cannot afford the best treatment for all of the people all of the time. You have decided how to spend most of your budget and you have a few hundred thousand pounds left uncommitted. You sit down with your advisers to consider the best way of spending this last remaining tranche of money. There are three possibilities and you must choose one of them.
The possibilities are:
(1) A new treatment for bowel cancer that gives the relevant patients a small but significant chance of increased life-expectancy;
(2) A new drug that lowers the chance of death from heart attack in people with genetically induced raised blood cholesterol;
(3) A new piece of surgical kit that ensures a lower mortality from a particularly difficult kind of brain surgery.
On what basis do you choose between these possibilities?
One approach that has a lot going for it is to say: there is no good reason to prefer one person’s year of life over another person’s, or to give any priority to people who would benefit from the bowel cancer treatment over people who have the genetically induced high blood cholesterol or to people with the brain tumour. In each case people stand to die prematurely and in each case the treatment increases the chance that they will live for longer. What we should do, therefore, is to spend the money so that we can ‘buy’ as many life years as possible. By doing this we are treating everyone fairly: we are valuing one year of life equally, regardless of whose life it is.
The distribution problem
Even amongst people (like me) who are attracted by this approach there is an issue that needs to be faced: the ‘distribution problem’. Take a look at the three interventions described in the box.
Choosing between three interventions
Intervention - 1 benefits 10 people total life years gained: 35
Intervention - 2 benefits 15 people total life years gained: 30
Intervention - 3 benefits 2 people total life years gained: 16
Suppose that all these interventions cost the same and that we can only afford one of them. Suppose further that the distributions are as follows. The two people who are benefited by intervention 3 will enjoy 8 more years of life each. Of the ten people who are benefited by intervention 1, the average benefit is 3.5 years and the range is 2–4 years. Of the fifteen people who are benefited by intervention 2, the average benefit is 2 years and the range is 1 to 3 years. Which of the three interventions should we go for?
If we think that what we should do is to ‘buy’ the maximum number of life years that we can (the maximization view) then we should put our money into intervention 1 because we buy 35 life years, which is more than we will get if we spend the money on either of the other two interventions. Some might argue that intervention 2 is preferable because we help more people (15 as opposed to 10) although each person gains fewer extra years of life. Still others might argue that intervention 3 is the best option because the two people who are helped receive a really significant gain (eight years of life) whereas no one gains more than four years of life with either of the other two options.
The question of whether it is only the total number of life years that matters, or whether the way in which those years are distributed between people is important, is known as ‘the distribution problem’. Those who reject the maximization view have to specify how they balance the value in helping more people, but each gaining relatively less, against the value in helping fewer people, but each gaining relatively more. Except at extremes I am generally happy to go with maximizing the total number of life years and not worry too much about their distribution.
In being generally happy with using resources to maximize total number of life years I am in a minority - and no health care system in the world behaves remotely in this way. One problem with my position (the maximization view) takes us right back to Tony Bullimore and his attempt to sail round the world. My position gives no moral weight to what has been called ‘The Rule of Rescue’ - and yet this rule seems, intuitively, to be right.
The rule of rescue
The ‘rule of rescue’ is relevant to a situation where there is an identified person whose life is at high risk. There exists an intervention (‘rescue’) which has a good chance of saving the person’s life. The value that is at the heart of ‘the rule of rescue’ is this: that it is normally justified to spend more per life year gained in this situation than in situations where we cannot identify who has been helped.
Consider two hypothetical, but realistic, situations in health care.
Intervention A (saves anonymous ‘statistical’ lives) A is a drug which will change the chance of death by a small amount in a large number of people. For example, out of every 2,000 people in the relevant group, if A is not given than 100 people will die over the next few years. If A is given then only 98 will die. Although we know that drug A will prevent deaths we do not know which specific lives will be saved. Drug A is cheap – the cost per life year gained is Euros 20,000. One example of a medical treatment like this is treatment that lowers moderately raised blood pressure. Another example is a class of medicines known as statins that lower blood cholesterol. Lowering blood pressures, and lowering cholesterol, reduce risk of heart attack, stroke, and death.
Intervention B (rescues an identified person) B is the only effective treatment for an otherwise life-threatening condition. Those with the condition face a greater than 90 per cent chance of death over the next year if not given B. If given B then there is a good chance of cure – say 90 per cent. B is expensive. The cost per life year gained is Euros 50,000. Renal (kidney) dialysis is an example of this type.
There are three, potentially relevant, differences between intervention A and intervention B. The first is that B saves lives within the next year, whereas the benefits of A are not realized for many years. This difference has some moral relevance. A few of those who might benefit from intervention A will die from some quite independent cause before any benefit from A could be gained.
There are also problems in calculating the cost per life year gained when at least some of the costs of the intervention are borne years before the benefits are seen. This is because of monetary inflation. Both these effects can be allowed for in the calculation of cost per life year gained. Having made such allowances, there seems no good reason to value the saving of life years in the future any less than saving life years now.
The second difference between the interventions is that B will almost certainly save the lives of the relevant patients, but A only has a low probability of doing so. Thus B might be seen as giving greater benefit to individuals than A. I will argue, in a moment, that this is false.
The third difference is that intervention B benefits identifiable people. Intervention A benefits a proportion of patients within a group (e.g. those with raised blood pressure), but we cannot know who within the group will benefit (although we may know the likely proportion that will benefit).
According to the rule of rescue it may be right for a health care system to fund intervention B but not intervention A, even though B is more expensive in terms of life years gained. For example, the rule of rescue would provide justification for spending more per life year gained on treatments such as renal replacement therapy, than on treatments like statins.
In practice this is exactly what health care systems do. The British National Health Service provides renal dialysis at costs over Euros 50,000 per life year gained, whilst paying for statins only for those with very high cholesterol levels. This is despite the fact that treatment with statins for those with moderately raised cholesterol levels would cost only about Euros 10,000 per life year gained. In other words, if the money spent on some people for renal dialysis were, instead, spent on some people with moderately raised cholesterol, five times as many life years could be gained. But we don’t do it – because we would feel that we had condemned the person needing dialysis to death; whereas all we would be doing in the case of statins is slightly lowering an already quite small chance of death.
The most powerful reason in support of the rule of rescue is that in the typical case the identified person, like Tony Bullimore, stands to gain a significant increase in chance of life, whereas in the typical case of saving anonymous ‘statistical’ lives no one stands to gain more than a small decrease in probability of death. I will put this argument in favour of the rule of rescue as strongly as I can. I will then say why I do not agree with it.
(6) Ethics in medicine
People who don’t exist; at least not yet
The minutest philosophers, who, by the by, have the most enlarged understandings, (their souls being inversely as their enquiries) show us incontestably, the HOMUNCULUS . . . may be benefited, – he may be injured, – he may obtain redress; – in a word, he has all the claims and rights of humanity, which Tully, Puffendorf, or the best ethic writers allow to arise out of that state and relation.
The story of medical ethics begins before conception. In the opinion of Tristram Shandy, a person’s character, and the life he will enjoy, is shaped by the parents’ thoughts during copulation. Tristram complains:
I wish either my father or my mother, or indeed both of them, as they were in duty both equally bound to it, had minded what they were about when they begot me; had they duly considered how much depended upon what they were then doing; – that not only the production of a rational Being was concerned in it; but that possibly the happy formation and temperature of his body, perhaps his genius and the very cast of his mind: – and, for aught they knew to the contrary, even the fortunes of his whole house might take their turn from the humours and dispositions which were then uppermost . . . Pray, my Dear, quote my mother, have you not forgot to wind up the clock?
– Good G—! Cried my father, making an exclamation, but taking care to moderate his voice at the same time,
– Did ever woman, since the creation of the world, interrupt a man with such a silly question?
The Human Fertilization and Embryology Act 1990 (HFEA) – the law that governs assisted reproduction services in the UK – requires doctors to mind what they are about when they help a woman to conceive a child. The Act states: ‘A woman shall not be provided with treatment services unless account has been taken of the welfare of any child who may be born as a result of the treatment (including the need of that child for a father) . . . ’
A great deal of brouhaha was created in the British press when a post-menopausal woman aged 59 years went to a private fertility clinic in Italy to be helped to conceive a child (in fact she subsequently gave birth to twins). ‘Think of the poor children who will be born’ was one response ‘they will be the laughing stock of their friends when they are met at the school gate by such an elderly mother’. According to one member of the Human Fertilization and Embryology Authority (which oversees fertility clinics), concern for the welfare of the potential children rules out fertility treatment for elderly women.
The welfare of children is so important a consideration in our moral thinking that the wording of the HFEA may seem unproblematic: but this is not so. When assisting conception it is not the welfare of an actual child that is under consideration, it is the welfare of a child that may exist at a later time, if indeed there will later exist any such child at all. It turns out that a consideration of the welfare of children who may exist at a later time is a very slippery customer indeed.
The analogy with adoption
In the early days of in-vitro fertilization (IVF) – the technique that led to the idea of test-tube babies – a Manchester woman was removed from the IVF waiting list when it was discovered that she had a criminal record involving prostitution offences. The hospital concerned had a policy in place (this was a couple of years before the HFEA was enacted). This policy stated that couples wanting IVF ‘must in the ordinary course of events, satisfy the general criteria established by adoption societies in assessing suitability for adoption’.
In effect this policy means that if a person seeking IVF would not be considered suitable as an adoptive parent, she should not be provided with assistance to reproduce. And underlying this policy, presumably, is the idea of the welfare of the child who might exist at a later time. But does the analogy between adoption and assisting reproduction hold?
In the case of adoption we have a child (child X) and a number of possible adoptive parents: A, B, C etc. Suppose that we have good reason to believe that parents A will be better parents than B, C, etc. and that child X is likely to have a better life if we choose parents A than if we choose any of the other parents (B, C, etc.). Assuming that judgments about the likely quality of parenting can be made (and such judgments have to be made by adoption agencies) then we act, as far as we can judge, in child X’s best interests in giving child X to parents A.
Now compare this situation of adoption with that of assisting reproduction. Suppose that couples A, B, C, etc. come for help with fertility treatment. All these couples are likely to be perfectly reasonable parents but we have good reason to believe that couple A are likely to be better parents than couples B, C, etc. Which couple should we help? Would we not be acting in the best interests of the child who may come to exist if we helped parents A, on the grounds that, as far as we can judge, the child would be happier with couple A than with couples B, C, etc.?
It is not, however, as simple as this. There is no kingdom, as far as I am aware, of potential children waiting to be allocated to a particular set of parents. If we help couple A to conceive, then one child (child a) will come into existence. If we help couple B then a different child (child b) will come into existence. What sense can we make of assessing the interests of the child that may exist at a later time? If we help couple B then child b would come to exist and have a good start in life but not as good as child a would have done.
If we have the resources to help only one couple, which couple should we choose, if our only criterion is what is in the best interests of the child who will come to exist? It is tempting to say that the best interests of the child would be served by helping couple A. But this is wrong. It will be a different child depending on which couple we help. It is in potential child as best interests for us to help couple A, but in potential child by are best interests to help couple B. If we focus on the interests of the child who may exist at a later date the question that needs to be asked is: are these interests better served if he or she is born to these parents or if he or she never exists at all?
The question, put this way, is of course rather odd since it asks us to compare existence with non-existence. Perhaps a better question is: if there were later to exist a child to this couple, would it have a reasonable expectation of a life worth living? I will come back to these issues in the next section. The key point for the present discussion is that the possibility of ‘this’ potential child being born to any other (possibly better) parents does not arise. This, crucially, is where the analogy with adoption breaks down. If we have the resources to help only one couple then an argument could be made for choosing to help couple A. The argument is as follows: if we help couple A then the child that will exist (child a) will be happier (on the best prediction) than the child (child b) who would have existed had we helped couple B. If there are no other relevant grounds for choosing between the various couples then it is better to act in such a way as to bring about the existence of the happiest children that we can. We are, in this case, most likely to bring about the existence of the happiest child that we can by helping couple A rather than couples B, C, etc. We should, therefore, help couple A. In choosing to help couple A we are acting against the best interests of the child who would have existed in the future had we helped couple B instead. Our choice to help couple A is not on the grounds of an individual’s best interests but in order to make the world a better place. The child who will actually exist in that ‘better world’ (i.e. child a) will have a better life than the different child (child b) who would have existed had we helped couple B rather than couple A.
This point can be made more strongly by considering the following analogy. Suppose that a hospital delays the admission of a patient who requires non-urgent surgery in order to admit a patient requiring an urgent operation. No one would maintain that it was in the best interests of the first patient that her surgery be delayed. On the contrary, it is against her best interests. The justification for acting against her best interests is in order to benefit the patient who needs urgent surgery. Since a choice has to be made, the decision to give priority to the patient in more urgent need seems the right one.
We seem to have found an argument that justifies the initial intuition that, in the case of assisting reproduction, we should help couple A rather than couples B, C, etc. (assuming that we have the resources to help one only). This argument is not based on the idea of acting in the best interests of the child who may be born. It is not based on following the guidelines from the HFEA or from St Mary’s Hospital in Manchester. Instead, the argument is based on the idea of welfare maximization: that we should act so as to bring into existence as happy children as we can. Does it matter that the reasons are different, if the decision is the same? The answer is that it does, both in theory and in practice.
Comparing existence with non-existence
We have been assuming that we can help only one of the couples A, B, C, etc. But often this is not the case. The 59-year-old woman who went to Italy and conceived twins bore the costs herself. The clinic did not have to choose between her and someone else. The outcry in the British press was not on the grounds that some other couple would not receive help as a result of her being assisted to conceive. The outcry was on the grounds that it was against the interests of the potential child (i.e. any child who might be born) that she be helped to conceive at all. If we focus solely on the interests of the potential child, the question, I have suggested, that needs to be asked is: are the interests of this potential child better served if he or she is born tothese parents, or if he or she never exists at all? But this is a very strange question. Does it make any sense to compare existence (in whatever state) with non-existence? Some have said such a comparison is like dividing by zero – it appears to make sense at first sight, but it is a function without meaning. Others think that as long as the child will not have an appalling life then it is in the child’s best interests to exist, on the grounds that, on the whole, existence is a positive thing. Perhaps some, like Montesquieu, of a more pessimistic disposition, take the opposite view and see existence, on balance, as a negative experience.
If those who say that one cannot compare existence with nonexistence are correct, then the criterion of the best interests of a potential child is meaningless. But this view faces a difficulty. Let us suppose, for the sake of argument, that were couple J to have a child that child would suffer immensely (perhaps from some dreadful genetic condition). The child would live in constant pain and finally die, to the relief of all, at the age of one. So the life of this child would be one year of constant pain followed by death. In these circumstances it does seem to make sense to say that it would be wrong to help couple J conceive such a child on the grounds that to do so would be against the interests of the child who would exist.
It may be possible to make sense of this judgment without having to ‘divide by zero’. Over any period of life one can ask whether, overall, the experiences are positive or negative. The zero line here is such that life above zero is overall worth living for the person concerned and life below zero is not worth living. In the case of the child who would be born to couple J, his life, overall, would rate as below zero. It is for this reason that we can say that it is in his best interests not to be born. In saying this we do not rely on the problematic comparison of non-existence with existence, but on being able to make a judgment as to whether the life it is predicted that he would have would, overall, be above or below zero.
The argument that the post-menopausal 59-year-old woman should not be helped to conceive, on the grounds that to do so would be against the best interests of the potential child, falls apart, whichever view you take on this issue.
1. If it makes no sense to compare existence with non-existence then it makes no sense to argue that in helping the woman conceive one is acting against the best interests of the potential child. For on this view one cannot argue anything on the basis of best interests, since on this view it is meaningless to compare the interests in not existing with the interests in existing.
2. If, on the other hand, it does make sense to judge whether it is in the interests of a child (who may exist in the future) to exist, and if that judgment is essentially whether the predicted life will be, overall, a positive experience, then the question to be asked is this: is the predicted life of a child born to this 59-year-old woman, overall, likely to be positive?
If, like me on a bad day, you take a rather gloomy view of existence, then perhaps you think it is not in the interests of the child, who may come to exist, for the woman to be helped to conceive. But it was not such a view that prompted the outcry against helping the post-menopausal woman to conceive. Such a view would, after all, justify refusing to help almost all couples seeking help in reproducing. A more balanced view would be that being teased at school might make a child unhappy but hardly justifies the claim that it means that overall his life would not be worth living. Where
courts have had to decide whether it might be in the best interests of very young children to be allowed to die rather than have life extending treatment they have set the standards very high: that is, the life has to be very bad for the courts to decide that it would be in the child’s best interests to be allowed to die. The outcry at helping the post-menopausal woman to conceive was based on the grounds that the life of the child who may exist as a result of the treatment would not go as well as children born to a younger mother. But that, as I have argued, is not relevant to the question of the best interests of the child who would come to exist were we to help the woman.
That child could not exist as the child of a younger woman.
Identity-preserving and identity-affecting actions
There is a fundamental distinction that arises from this discussion: that between an identity-preserving and an identity-affecting action or decision. An example of an identity-preserving action is when a pregnant woman drinks large amounts of alcohol. The drinking of the alcohol in this example does not affect the identity of the foetus. If the child is subsequently born with some brain damage as a result of the mother’s alcohol intake that child has been harmed by the alcohol intake.
An example of an identity-affecting action is when a woman delays reproduction from, for example, 30 to 40 years of age. A different child will be born as a result of her delay. When a doctor chooses to help couple A to conceive, rather than couple B, she is making an identity-affecting decision.
What is the effect of the identity-affecting nature of an act on the morality of that act? This is a question that was first asked in the context of the analysis of fundamental moral theory. It is a question that is becoming of increasing importance to doctors.
The non-identity problem and identity affecting interventions
Derek Parfit called this issue the non-identity problem. He explains the problem using the example of ‘the 14 year old girl’. He writes: This girl chooses to have a child. Because she is so young, she gives her child a bad start in life. Though this will have bad effects throughout this child’s life, his life will, predictably, be worth living.
If this girl had waited for several years, she would have had a different child, to whom she would have given a better start in life.
Suppose that we tried to persuade this girl that she ought to wait . . . ‘You should think not only of yourself, but also of your child. It will be worse for him if you have him now. If you have him later, you will give him a better start in life.’ . . .
We failed to persuade this girl . . . Were we right to claim that her decision was worse for her child? If she had waited, this particular child would never have existed. And, despite its bad start, his life is worth living . . . ‘If someone lives a life that is worth living, is this worse for this person than if he had never existed?’ Our answer must be No . . . When we see this; do we change our mind about this decision? Do we cease to believe that it would have been better if this girl had waited, so that she could give to her first child a better start in life? . . . We cannot claim that this girl’s decision was worse for her child. What is the objection to her decision? This question arises because, in different outcomes, different people would be born. I shall therefore call this the Non-Identity Problem.
Parfit’s example raises many issues other than the non-identity problem, not least of which is what is in the interests of the girl herself. I want to set these other issues to one side. I give bellow some further medical situations in which the non-identity problem arises. In all these cases it can certainly be argued that it would be better if the decision were made that would lead to the birth of whichever child would be likely to have the better life. Such an argument could be based on the idea of maximizing overall welfare. In none of the cases, however, can an argument be based on the interests of the potential child. Nor can it be claimed, whichever decision is made in the three cases, that the child born has been harmed by the decision.
The non-identity issue has an important impact on what doctors should do. Where the doctor aids an act, such as in prescribing during pregnancy a drug that may harm a foetus, then such harm provides a good reason for the doctor to refuse to prescribe the drug even when the woman wants it and it is appropriate treatment.
Prescribing this drug is an example of an identity-preserving action. But when the doctor’s action is an identity-affecting action that may lead to a child being born with a handicap then there is no child who has been made worse off than she could have otherwise been.
In societies that give considerable weight both to patient autonomy and reproductive choice, doctors should not normally override a woman’s choice in situations where no person is harmed; and in identity-affecting decisions, or acts, no person is harmed (unless the handicap is so severe that the child’s life, overall, would not be worth living). Such a conclusion goes against normal intuition. In this case, it seems to me, normal intuition is wrong: it is based on a false metaphysics.
Three clinical examples that involve the non-identity problem
1. Preimplantation genetic testing
Hypothetical case 1: ‘deafening’ an embryo. A couple with a genetic condition causing deafness wishes to have a child who is also deaf. This is so that the child is part of the ‘deaf community’.
The woman becomes pregnant. Genetic testing shows that the foetus does not have the gene causing deafness: it is likely to become a normal child. Suppose that a drug is available that if taken by a pregnant woman will cause a normal foetus to become deaf. It has no other effect and is otherwise completely safe for both embryo and mother. The couple decides that the woman should take this drug in order to ensure that their child is born deaf.
(a) Would the couple be morally wrong to choose to take the drug?
(b) Would a doctor be wrong to prescribe the drug at the couples’ request?
(c) If the parents did take the drug and their child were born deaf, would the child have a morally legitimate grievance against the parents, and/or the doctors?
I imagine that most people will answer ‘yes’ to these three questions. Now consider the following hypothetical case.
Hypothetical case 2: choosing a ‘deaf embryo’. A couple with a genetic condition causing deafness wishes to have help with conceiving. A number of embryos are created, using IVF (the sperm fertilizes the egg in a laboratory and outside the woman’s body, and the fertilized egg is then implanted into the woman’s uterus/womb). These are genetically tested to see which have the ‘deafness gene’. Embryo A is a genetically normal embryo. Embryo B has the ‘deafness gene’ but is otherwise genetically normal. The couple chooses to have embryo B implanted and subsequently give birth to a deaf child: child B. (If you consider that the embryo has the full moral status of a person, vary the example to involve egg, rather than embryo, selection.)
(a) Are the couple morally wrong to choose, for implantation, embryo B rather than embryo A?
(b) Would doctors be acting wrongly to accede to their request?
(c) Does child B have a morally legitimate grievance against the parents and/or the doctors?
At first sight it seems wrong for the couple to choose to have a deaf child when they could have had a child with normal hearing, and wrong for doctors to allow such a choice. The principal reason why this seems wrong is that such a choice would be harmful to the child. But this is false: it is not harmful to the child because the choice of which embryo to implant is an identity-affecting choice.
2. Delaying pregnancy
A 35-year-old woman hopes in the long run to become a mother, but not yet. She wants to delay pregnancy for another four years until she has finished a degree course. She knows that she is more likely to conceive a child with Down’s syndrome if she delays pregnancy. (Down’s syndrome is caused by an extra chromosome over the normal number, i.e. 47 rather than 46. Most people with Down’s syndrome have some degree of learning difficulty.) She asks her doctor for a prescription for the contraceptive pill. The doctor prescribes the pill for the next three and a half years. After this the woman becomes pregnant and has a child with Down’s syndrome. Did the doctor’s act, in prescribing the contraceptive pill, harm the child?
3. Treating acne
Acne is a skin condition that typically affects adolescents. It is characterized by spots and small pustules that are distributed over the face. Most adolescents experience mild acne but some suffer a much more severe form. Severe acne, if left untreated, can lead not only to psychological problems but also to permanent facial scarring. Sometimes the only effective treatment for severe acne is a drug called isotretinoin.
There is one, very important, unwanted effect of isotretinoin: it may cause foetal damage if a woman is taking the treatment during pregnancy. Children may be born with congenital malformations mainly of facial appearance or of the heart.
Because of the significance of these unwanted effects on a foetus it would normally be considered wrong for a doctor to prescribe isotretinoin to a woman with severe acne known to be pregnant, even if the woman wanted the treatment, because of the harm to the foetus, or at any rate the child that the foetus will become.
What should a doctor do, however, in circumstances where a patient is not pregnant, but might become so while taking the drug? The advice that is given to doctors on this issue is that they should only prescribe the isotretinoin if the woman will reliably delay pregnancy until after she has stopped taking the isotretinoin. In some situations this will require the doctor to prescribe the isotretinoin only in combination with the contraceptive pill.
On this view it is right for a doctor to prescribe isotretinoin to a non-pregnant woman if she will reliably delay pregnancy until after the course of isotretinoin (typically six months to a year); but wrong to prescribe it if she will not reliably delay pregnancy. The intuition is that if she does not delay pregnancy then she has harmed the child, but if she does delay pregnancy then she has not harmed the child. Once again, however, it will be a different child. If she becomes pregnant, and the child is born with a handicap, it cannot be claimed that the child has been harmed as a result of the woman’s not delaying pregnancy. For if the woman had delayed pregnancy that child would not have existed at all.