Colleges, London El 2AD. Len Doyal, senior lecturer in medical ethics. Daniel Wilsher, King's Fund researchfellow in medical ethics. Correspondence to:.
so, the decision to discontinue advanced cardiac life support should be made by the physician in charge, but all others involved in the attempt should usually be consulted.
3 Robertson C. The precordial thump and cough techniques in advanced life support. Resuscitation 1992;24:133-5. 4 Bossaert L, Koster R. Defibrillation: methods and strategies. Resuscitation 1992;24:21 1-25. 5 Robertson C, Holmberg S. Compression techniques and blood flow during
Members of the working party were: Douglas Chamberlain (England; chairman), Leo Bossaert (Belgium), Pierre Carli (France), Erik Edgren (Sweden), Lars Ekstrom (Sweden), Svein Hapnes (Norway), Stig Holmberg (Sweden), Rudy Koster (Netherlands), Karl Lindner (Germany), Vittorio Pasqualucci (Italy), Narciso Perales (Spain), Martin von Planta (Switzerland), Colin Robertson (Scotland), Petter Steen (Norway).
6 Carli P, Hapnes SA, Pasqualucci V. Airway management and ventilation. Resuscitation 1992;24:205-10. 7 Hapnes SA, Robertson C. CPR-drug delivery routes and systems. Resuscitation 1992;24: 137-42. 8 Lindner KH, Koster R. Vasopressor drugs during cardiopulmonary resuscitation. Resuscitation 1992;24:147-53. 9 Koster R, Carli P. Acid-base management. Resuscitation 1992;24:143-6. 10 von Planta M, Chamberlain DA. Drug treatment of arrhythmias during cardiopulmonary resuscitation. Resuscitation 1992;24:227-32. 11 Steen AP, Edgren E, Gustafson I, Fuentes CG. Cerebral protection and postresuscitation care. Resuscitation 1992;24:233-7. 12 Holmberg S, Ekstrom L. Ethics and practicalities of resuscitation. Resuscitation
1 Basic Life Support Working Party of the European Resuscitation Council. Guidelines for basic life support. Resuscitation 1992;24:103-10. 2 Advanced Life Support Working Party of the European Resuscitation Council. Guidelines for advanced life support. Resuscitation 1992;24: 111-21.
cardiopulmonary resuscitation. Resuscitation 1992;24:123-32.
1992;24:239-44.
(Accepted 29April 1993)
Withholding cardiopulmonary resuscitation: proposals for formal guidelines Len Doyal, Daniel Wilsher Working with members of the Royal London Trust and its medical council, Len Doyal and Daniel Wilsher have composed a set of guidelines governing the making of decisions to withhold resuscitation from patients. The guidelines describe the procedures that should be followed when giving orders for non-resuscitation and the clinical, legal, and moral criteria that should be satisfied before such orders are issued. The authors hope that these guidelines will be of help to those responsible for the creation of hospitals' policies for non-resuscitation.
Joint Department of Human Science and Medical Ethics, London and St Bartholomew's Hospitals' Medical Colleges, London El 2AD Len Doyal, senior lecturer in medical ethics Daniel Wilsher, King's Fund research fellow in medical ethics Correspondence to: Mr Doyal. BMJ 1993;306: 1593-6
BMJ VOLUME 306
Although there has been some discussion in the United Kingdom of hospital policy on withholding cardiopulmonary resuscitation from patients who suffer cardiac arrest, no consensus has yet emerged.`3 Generally decisions continue to be taken according to the clinical judgment of those caring for a patient without regard to more formal guidelines. Over the past year this informal approach has come under scrutiny, largely as a result of a letter sent by the chief medical officer to all consultants in England and Wales. This asked for the formulation of clear policies on the making of decisions to withhold resuscitation.4 The chief medical officer further indicated that consultants should make their policy clear to all junior staff. His letter was precipitated by the upholding of a complaint made to the parliamentary ombudsman by the son of an elderly woman who had been given "Not for resuscitation" status by a junior doctor without consultation.5 There is evidence that this was not an isolated incident. A study conducted in a district general hospital suggested that considerable confusion existed over the resuscitation status of patients.6 Decisions were poorly documented or in some cases not documented at all. This meant that nursing staff were unaware of some patients' not for resuscitation status, potentially resulting in inappropriate cardiopulmonary resuscitation. Discussions with colleagues suggest that arbitrary differences may exist in assigning not for resuscitation status depending on the specialty of the consultant in charge.
Professional and legal duty to act to save life It has been argued that a failure to attempt resuscitation is an omission and therefore acceptable because 12 JUNE 1993
"nothing" is done. Nature is allowed to take its course. This argument has little legal or moral basis.' In general, when there is a recognised professional duty to act to save life, not to do so in the face of cardiac arrest is a prima facie breach of the law. Morally an omission constitutes a choice, itself an action, which may or may not be culpable. Like any other actions, omissions must therefore be justified by the acceptability of their consequences-in this instance the death of a patient. If there was evidence that such a patient might have survived resuscitation the doctor who issued the not for resuscitation order could face criminal proceedings. This would depend, in part, on whether reasonable grounds existed for the decision not to resuscitate. Even if the physician was found not guilty further charges of professional misconduct might be brought by the General Medical Council. Clarification about what constitutes a legally and morally appropriate policy for cardiopulmonary resuscitation is therefore required. There is some relevant case law specifying the circumstances under which it is legitimate to withhold life prolonging treatment. Morally, the increasing emphasis within medicine on respect for individual autonomy potentially clashes with a tradition of clinical discretion, which continues to deprive patients of any knowledge of their cardiopulmonary resuscitation status. If such discretion is to be justified it must be against a background of acceptable legal and moral principles. The BMA, in conjunction with the Royal College of Nursing, has recently published a policy on nonresuscitation.8 While helpful in some respects, it lacks the clarity and completeness necessary to inform good clinical practice. Our aim in this paper is to formulate and defend more comprehensive guidelines, which can be incorporated into publicly stated policies for cardiopulmonary resuscitation within British hospitals.
Grounds for non-resuscitation There are three occasions when we consider nonresuscitation to be acceptable. The first is when patients deemed competent to give informed consent to medical treatment indicate that they refuse resuscitation after having been told about the probable consequences of cardiac arrest and resuscitation. This might be done by means of an "advanced directive," given recent judicial approval of this concept.9 The 1593
The third acceptable justification for non-resuscitation applies equally to competent and non-competent patients and does not require informed consent. The clinical condition of some patients is such that the probability of successful resuscitation approaches zero.'7 In such cases cardiopulmonary resuscitation can rightly be said to be futile as it will not benefit the patients. Indeed, it may well further harm them even if it is "successful" by resulting in severe brain damage. To the extent that these patients can be identified in advance this provides sufficient reason for issuing a not for resuscitation order. No consent is morally or legally required in such circumstances. It is not part of a doctor's duty of care to administer useless or harmful treatment.'819 We must emphasise, however, that a high degree of proof is required for this justification to be valid-cardiopulmonary resuscitation must be very unlikely to succeed. If there is significant uncertainty about the. likely outcome then resuscitation must be attempted unless a competent patient has consented to not for resuscitation status or cardiopulmonary resuscitation is not in the best interests of a noncompetent patient. difficulty in practice is discussing the possibility of death from cardiac arrest with a patient in the first place. This is rarely done, and many reasons are given for not informing patients about resuscitation. We know of no convincing moral argument, however, that competent patients have any less right to know that they are being considered for non-resuscitation than they have to other general information about their options for treatment. Patients may, of course, refuse resuscitation by implication, as part of a general rejection of life sustaining treatment.'01' If the risks associated with resuscitation were communicated with the sensitivity that one would expect in other instances of breaking bad news, it is not clear that the ordeal would be as harrowing as is often suggested by clinicians who resist the idea. In the United States, for example, obtaining the informed consent of patients for a not for resuscitation status is common practice in many hospitals.'2 Of course, a discussion of resuscitation and its potential aftermath might itself be dangerous to the health of a patient. This is a strong argument against obtaining informed consent provided there is clinical evidence that it could be dangerous. A non-resuscitation order is also acceptable when a patient is not competent to give consent and a clinician judges resuscitation to be against the patient's best interests. The problem is deciding what circumstances warrant the judgment that death is better than continued life. Morally, it has been argued that nonresuscitation is compatible with respect for human rights if a patient's mental capacity is so severely and permanently impaired that he or she cannot flourish in even minimal ways."3 The same conclusion is reached with the argument that it is immoral to inflict long term pain and suffering on a patient by giving a treatment about which the patient is permanently unable to make a choice.'4 Case law in neonatal medicine similarly suggests that non-resuscitation is justified if there is good evidence that a neonate is inevitably going to die soon despite medical or surgical intervention." The same holds true-although here the legal judgments are more qualified-if a neonate is irreversibly and severely brain damaged but not actually dying.9 There is no reason why legal principles developed in this clinical context should not apply equally to all non-competent patients. The only English case relating to an adult is that of Anthony Bland. The House of Lords held that it was not unlawful to withhold life sustaining artificial nutrition and hydration from a patient in a persistent vegetative state.'6 1594
Withholding resuscitation because of resources Some clinicians have said in private that the real reason for not for resuscitation orders was often concem about the resource implications of a successful resuscitation with a poor and prolonged clinical outcome. Such comments underline current confusion about what the clinical and moral criteria should be for deciding the allocation of resources. Until the moral grounds for the finance of resuscitation by purchasers are clarified, it is inappropriate for clinicians to refer to resources when justifying decisions to withhold resuscitation. Such references are also legally ill advised. In a recent case, however, it was stated that providers could not be ordered to deliver particular treatments because when they "find that they have too few resources . . . it is then their duty to make choices."20 The difficulty in applying this judgment to resuscitation is that the procedure is relatively inexpensive, although the clinical outcome can sometimes prove costly. If equipment for cardiopulmonary resuscitation is available then it should be provided to all patients who fall outside the criteria listed previously. These include patients whose quality of life may be deteriorating but for whom resuscitation is not yet clearly against their best interests (for example, mildy demented patients). Such patients might legitimately be denied access to advanced cardiopulmonary resuscitation only if a higher authority had deemed it too expensive to provide a crash team on the site where these patients were being treated. Clinicians cannot withhold resuscitation from some patients for reasons of lack of resources if a crash team is available to other patients on the same ward. Proposed guidelines for non-resuscitation We suggest that a policy based on our proposed guidelines conforms as closely as possible to acceptable moral and legal dictates and optimises clinical discretion. The guidelines were developed through lengthy negotiation with clinical colleagues at the Royal London NHS Trust, many of whom put considerable thought and effort into the process. The guidelines already form the basis of the trust's policy on nonresuscitation and are beginning to be applied throughout the hospital. We hope to conduct an audit of clinical practice in relation to these guidelines in the near future. BMJ VOLUME 306
12JUNE 1993
A-PROCEDURES FOR GIVING NOT FOR RESUSCITATION ORDERS
(1) Not for resuscitation status means that if a patient suffers a cardiac arrest the crash team will not be called and neither advanced nor basic cardiopulmonary resuscitation will be given. It has no implications for any other clinical decisions conceming a patient's management. (2) Only consultants and senior registrars ("senior clinicians") possessing MRCP or FRCS or its equivalent may give a not for resuscitation order. If more junior staff, however, are responsible for a cardiopulmonary resuscitation procedure which is clearly failing then, in the absence of a senior clinician, there is no obligation for them to continue. (3) When a not for resuscitation order is made by a senior registrar the agreement of the consultant in charge of the case, or that of the consultant's named senior deputy, should be obtained at the earliest opportunity. (4) If a senior registrar has any doubt about making a decision the case must be referred to the consultant in charge. (5) The date and time of giving a not for resuscitation order should always be prominently entered into both the patient's medical and nursing notes, as well as the clinical justification for the order and the patient's documented consent when applicable. (6) Not for resuscitation orders should be reviewed routinely by senior medical staff: every 24 hours unless this is inappropriate. (7) Senior nursing staff and, when appropriate, other clinicians concemed with a case should be consulted before not for resuscitation status is decided. Final responsibility for the decision rests with the senior clinician who takes it. (8) When appropriate, consultations with patients or their relatives, or both, should be considered before decisions are made about not for resuscitation status. (9) Decisions about not for resuscitation status should be based on the clinical, legal, and moral conditions listed below. B-CLINICAL CONDITIONS FOR GIVING NOT FOR RESUSCITATION ORDERS
(1) A not for resuscitation order is clinically acceptable only if it is in "the best interests of the patient." This requires clear evidence that (a) a patient is irreversibly close to death in the short term or (b) resuscitation presents an unacceptably high probability of death or severe brain damage if the procedure is successful. (2) A senior clinician may think that a not for resuscitation order should be issued on grounds other than those in condition B1. In Britain this is legally debatable except in the case of patients in the persistent vegetative state. When applied to a patient with other irreversible and severe brain damage, however, there is a wide moral consensus that the order is still acceptable. Some existing case law points in the same direction, but until there is further legal clarification an order for non-resuscitation based on such grounds should be made only by the consultant in charge. C-LEGAL AMD MORAL CONDITIONS FOR GIVING NOT FOR RESUSCITATION ORDERS FOR COMPETENT PATIENTS
(1) Most patients who have given informed consent to treatment will be neither irreversibly close to death in the short term nor present an unacceptably high probability of death or very serious brain damage
following cardiopulmonary resuscitation. They should be given cardiopulmonary resuscitation and can be regarded as having given their implied consent unless they state otherwise. (2) If such a patient specifically indicates to a senior BMJ VOLUME 306
12 JUNE 1993
clinician and a witness that he or she does not wish resuscitation to be attempted the patient's views should be honoured. The clinician and the witness must be of the view that the patient was competent to make the decision and did so of his or her own volition.2' (3) A senior clinician may think that attempted resuscitation will not be in the interests of a competent patient on grounds other than those in condition B 1. Examples are prognoses of a future poor quality of life or of death which is not imminent. In such a case informed consent for a not for resuscitation order must be obtained. It should be given in the presence of a witness and can be implied in the form of a stated wish by the patient "not to go on" or some equivalent. (4) If a senior clinician is unsure about a patient's true wishes explicit consent must be obtained. The only exception is when the clinician judges that obtaining consent might pose a serious and immediate danger to the patient. Permission should then be sought to share general medical information with relatives, and, if possible, attempts should be made with their help to determine the patient's wishes. (5) If a senior clinician believes that cardiopulmonary resuscitation is in the best interests of a competent patient who has, however, expressed the desire "not to go on" the clinician must obtain explicit informed consent to non-resuscitation before making an order to that effect. (6) There may be patients (for example, those with dementia) who are judged to be competent to make some decisions but not competent to make others. If such a patient communicates a desire not to be kept alive to a senior clinician and a witness a not for resuscitation order may be issued provided that the patient is judged to have been competent when the wish was expressed. D-LEGAL AND MORAL CONDITIONS FOR GIVING NOT FOR RESUSCITATION ORDERS FOR NON-COMPETENT PATIENTS
(1) If a senior clinician judges an adult natient not to be competent to make a decision about whether resuscitation should be withheld the best interests of the patient should be defined with reference to conditions B 1 and B2. In other clinical circumstances patients should ordinarily be resuscitated. The one exception to this is if a patient has previously issued an advance directive. This should generally be honoured, but, in doubtful cases, it is advisable to seek legal and professional advice (for example, from the BMA or a defence society). (2) In the case of infants or children legal proxies -usually parents or those with parental authorityassume the rights of consent for patients, which have been outlined above. Decisions about nonresuscitation should be taken according to the same criteria which govem those conceming adults. Parents are under a duty to consent to treatment only if it is in the best interests of their child. (3) Relatives have no legal rights in the treatment of adult patients. Therefore, if no reliable indications of a patient's wishes exist senior clinicians must base their decisions about non-resuscitation on conditions B1 and B2. For moral and evidential reasons, however, discussions with relatives concerning non-resuscitation may be helpful if they are deemed appropriate by the senior clinicians involved. (4) If relatives insist on the resuscitation of a noncompetent patient despite the existence of conditions B 1 or B2 the case should be referred to the consultant in charge, who should explain the reasoning for nonresuscitation. If possible the wishes of relatives should be taken into account. Conversely, if relatives insist on non-resuscitation without the existence of conditions B I or B2 their request should be refused. 1595
E-PATIENTS UNDERGOING SURGERY OR GENERAL ANAESTHESIA
(1) When there is no existing not for resuscitation order resuscitation should be attempted unless, as a result of an intraoperative complication, one of the clinical conditions for non-resuscitation (B 1 or B2) is clearly satisfied. (2) When there is an existing not for resuscitation order the patient should not be resuscitated if an arrest occurs during anaesthesia.22 The one exception is when the reason for the order was that resuscitation presented an unacceptably high risk (condition B 1(b)) and the probability of success is now judged to be acceptable. F-INFORMATION ON RESUSCITATION FOR INPATIENTS
The literature given to inpatients when they are admitted should include a statement that it will be assumed that they consent to attempted resuscitation in the event of cardiac arrest unless they indicate otherwise to their clinician. Conclusion Some may argue that any attempt to produce guidelines constraining doctors' choices for withholding treatment is misguided or even harmful. There are fears that guidelines will be unworkable because they cannot cope with the nuances that arise across the wide range of clinical circumstances. Of further concem is the possibility that they will fetter clinical discretion in a situation when it is most needed. These are legitimate reservations that we have sought to accommodate in our guidelines. These are drawn in broad terms to allow many different situations to be assessed with reference to the key principles they embody. The guidelines are meant to provide clinicians with as much freedom as possible consistent with defensible legal and moral reasoning. The importance of conforming with the law in clinical practice goes without saying, at least for the purposes of this paper. The need to be able to justify decision making morally-to have good reasons for one's actions over and above any specific set of legal or social conventions-is also becoming increasingly accepted in the profession. Clinicians already make decisions to withhold resuscitation according to a range of personal principles, in which they have considerable discretion. Rules are not incompatible with individual interpretation; they invite it. The issue is whether the rules by which crucial clinical decisions are made are open to public scrutiny. We have placed less emphasis upon patients'
autonomy than have analogous American guidelines conceming non-resuscitation. While attempts should be made to obtain informed consent when a not for resuscitation order is contemplated on the grounds of quality of life, there are situations when this is not required. Furthermore, even when it is mandatory the patient does not have to give explicit agreement. We argue that consent to non-resuscitation can be implied in some cases from patients' general views about their goals for treatment. The production of formal guidelines will not in itself lead to changes in clinical behaviour. This will happen only when the guidelines are understood and accepted by those who would use them. We hope that at the very least this paper will stimulate further debate about how to put legal and moral theory into practice. We thank colleagues at the Royal London Hospital Trust and especially thank R D Cohen, E Dickinson, Lesley Doyal, A H McDonald, A D W MacLean, F W Marsh, C Reisner, M Swash, D L Wingate, and D Morgan. 1 SaundersJ. Who's for CPR?JR Coil Physicians Lond l992;26:254-7. 2 Evans TR, ed. ABC of resuscitation. 2nd ed. London: BMJ, 1990:66-7. 3 Holmberg S, Ekstrom L. Ethics and practicalities of resuscitation. Resuscitation
1992;24:239-44. 4 Chief Medical Officer. Letter to all consultants on resuscitation policy. London: Department of Health, 1991. (PLJCMO(9 1)22.) 5 Communications surrounding a decision not to resuscitate a patientW258/89-90. In: Report of the health service commissioner, 1990/1991. London: HMSO, 1991:12. 6 Aarons EJ, Beeching NJ. Survey of "Do not resuscitate" orders in a district
general hospital. BMJ 1991;303:1504-6. 7 Kennedy I, Grubb A. Medical law: test and materials. London: Butterworths, 1990:936-41. 8 British Medical Association, Royal College of Nursing in association with the Resuscitation Council (UK). Decisions relating to cardiopulmonary resuscitation. London: BMA, 1993. 9 Airedale NHS Trust v Bland (1993) 1 All ER 858, HL. 10 Schade SG, Muslin H. Do not resuscitate decisions: discussions with patients. J Med Ethics 1989;15: 186-90. 11 Loewy EH. Involving patients in do not resuscitate (DNR) decisions: an old issue raising its ugly head. J Med Ethics 1991;17:156-60. 12 President's Commission for the Study of Ethical Problems in Medicine and Biomedical Behavioral Research. Deciding to forego life-sustaining treatment. Washington, DC: US Government Printing Office, 1983:231-55. 13 Doyal LT. Neonatal surgery and the morality of selective non-treatment. In: Freeman N, Burge D, Griffiths M, Malone P, eds. Surgery of the newborn. London: Churchill Lvingstone (in press). 14 R C (a minor) [ 1989] 2 All ER 782, CA. 15 RJ (a minor) [1990] 3All ER 930, CA. 16 Bedell SE, Delbanco TL, Cook EF, Epstein FH. Survival after cardiopulmonary resuscitation in the hospital. NEngl7Med 1983;309:569-76. 17 George AL, Folk III BP, Crecelius PL, Barton Campbell W. Prearrest morbidity and other correlates of survival after in-hospital cardiopulmonary arrest. AmJMed 1989;87:28-33. 18 Tomlinson T, Brody H. Futility and the ethics of resuscitation. JAMA 1 990;264: 1276-80. 19 Tomlinson T, Brody H. Ethics and communication in do-not-resuscitate orders. N Engl7Med 1988;318:43-6. 20 RJ (a minor) [1993] 9 BMLR 20. 21 RT [1992[ 3 WLR 782. 22 Walker RM. DNR in the OR: resuscitation as an operative risk. JAMA 199 1;266:2407-12.
(Accepted 29 March 1993)
THE ONE MESSAGE I WOULD LIKE TO LEAVE BEHIND You must care enough By showing that you care you get things done, even to the extent of getting government offices to work really quickly. One of my jobs was to examine schoolchildren and issue permits to let them do odd jobs like paper rounds. This requires them to carry heavy bags full of newspapers and magazines. I was worried when a lot of shops were demolished to make way for a new road: no paper shops were left so vans were sent to the area with bags of papers and the children collected them from the van. At the weekend there were no shops at which to pay the bills. So the children were told to collect the money from the houses. It only needed a bully to cause mayhem. I asked for an urgent meeting with the medical officer of health, the chief education officer, and the chief constable and asked to get this practice stopped. I got nowhere.
1596
Within about three months a bully boy held a knife to a boy's throat and demanded all the money. I learnt about it and asked for another meeting. Action was immediate. Within 72 hours an addition to the bylaws was suggested and accepted by the secretary of state for the environment: "No schoolchild may act as the agent for the collection of money." The message is, "If you care enough you can get things done, and quickly, even though from the bottom rung of the ladder."-ROBERT CRAWFORD is a retired senior clinical medical officer in public health in Shipley We are delighted to receive submissions of up to 600 words on A paper (or patient) that changed my practice, A memorable patient, The one message I would like to leave behind, or related
topics.
BMJ VOLUME 306
12 JUNE 1993
