Do-not-attempt-cardiopulmonary resuscitation decisions in critical care: The gap between theory and practice

Abstract

The provision or withholding of Cardiopulmonary Resuscitation (CPR) is often a source of much controversy. Do-Not-Attempt CPR (DNACPR) decisions have been discussed extensively from an ethical and legal point, not to mention featuring heavily in the news. We aim to discuss various aspects of DNACPR decision-making with particular relevance to the Intensive Care environment. In doing so, we highlight important concepts associated with decision-making and common challenges encountered. Firstly, we provide a broad oversight of the current context in which such decisions are made before describing the principal ethical points of consideration and their relevance. This is followed by insights from a legal point as well as practical suggestions for approaching such decisions. It is hoped that these will help to assist clinicians with making such complex decisions under challenging circumstances.

Introduction

The incidence of Out of Hospital Cardiac Arrest (OOHCA) is 55 per 100,000 population. ¹ Survival in these patients has decreased, with recent data suggesting a 30-day survival of 7.8%. ² In-hospital cardiac arrest has an incidence of 1 to 1.5 per 1000 admissions and is associated with slightly better outcomes but even in this instance, only 23.6% of patients survive to discharge from hospital. ¹ Neurological outcomes are harder to define but favourable outcomes have been reported in as low as 5% of patients following cardiac arrest, particularly in those receiving CPR for longer than 15 min. ³ In this article, we describe the context in which DNACPR decisions are made in the Intensive Care Unit (ICU) and present the ethical and legal framework for such decisions in the UK, aiming to highlight potential challenges and areas for future research.

Cardiac arrest on ICU

The majority of comatose patients following a cardiac arrest are admitted to the ICU and this can account for almost 5% of all admissions. ⁴ Cardiac arrest within critical care itself occurs in approximately 2% of ICU admissions and has a high mortality of approximately 70%.^5,6 Mostly they occur on the day of admission and at times of anticipated risk such as tracheal intubation, lateral positioning for nursing care and during transfer. ⁷ Common precipitating factors include hypoxaemia, metabolic disorders, hypovolaemia and as an adverse effect of life-sustaining devices. On the one hand, this is to be expected with the most unwell patients in the hospital concentrated in a single area. Yet with invasive monitoring, mechanical organ support and the presence of highly skilled medical personnel, one would assume that most cardiac arrests are foreseeable, preventable and treatable. Therefore, it is somewhat surprising that the ICU can account for the majority of in-hospital cardiac arrests. ⁸

The literature around long-term survival and neurological outcomes following an in-ICU cardiac arrest is relatively sparce – this limits the available guidance for ICU clinicians called to make decisions about whether to offer CPR or not. ⁹ In a study of ‘very elderly’ (more than or equal to 90 years) patients who suffered a cardiac arrest whilst in ICU, survival to discharge from ICU was 12%. ¹⁰ There were no survivors at discharge from hospital. Whilst a small and limited patient group, this demonstrates the inefficacy of CPR in certain patient populations. In a French study, only one in six patients were alive at 6 months with a favourable neurological outcome (defined as Cerebral Performance Category 1 or 2). Factors in common amongst these patients included a shockable rhythm at the time of cardiac arrest, short time until return of spontaneous circulation (ROSC) and fewer organ failures prior to arrest. ⁷ Such poor outcomes remain frequent despite evidence suggesting that adherence to key resuscitation quality metrics is high within the ICU (with the exception of early defibrillation) as perhaps would be expected. ¹¹ In such studies, many patients would likely have had a DNACPR decision in place at the time, precluding attempts at resuscitation. Therefore, the true likelihood of successful resuscitation within the critical care population is inevitably much lower than described.

It is within this context that we should ask what the purpose of a DNACPR decision in ICU is? By nature of the circumstances prompting admission, it is unlikely to be the optimal setting for patient centred discussions and advanced care planning (see later). Perhaps it can prevent CPR being offered in circumstances where it would most likely be unsuccessful? In such a heavily monitored environment where experienced healthcare professionals are located, is it possible to facilitate contemporaneous decision-making at the time of cardiac arrest? Or is it to signpost to family members that a loved one is unlikely to survive their critical illness and transitioning to end of life care may need to be considered? Certainly, a DNACPR decision can often be the initial focus in discussions about establishing ceilings of treatment where a patient is deemed unlikely to recover.

Challenges around DNACPR decisions

Whilst without CPR following a cardiac arrest, death is somewhat inevitable, performing CPR is not without immediate and longer-term consequences. These include thoracic complications such as rib and sternal fractures, bleeding and abdominal visceral complications. ¹² Survival if it is achieved is associated with long term morbidity – both physical and psychological – with a high prevalence of debilitating depression, severe anxiety and post-traumatic stress disorder. ¹³ A greater concern is that inappropriate and unnecessary CPR may constitute a violation of the integrity of the body after death. This can be an affront to an individual’s human dignity and may even be considered as physical assault. The moral, spiritual and religious consequences of such actions are unknown. However, such an intrusive action as unwarranted CPR appears at odds with many of the commonly accepted principles of caring for the human body after death has occurred. The associated harm argues strongly against CPR as a default option especially where there is minimal prospect of success.

However, there are unwanted consequences of having a DNACPR decision. There is evidence that the mere presence of a DNACPR decision can negatively impact the standard of care for patients with regards to nursing and medical interventions. ¹⁴ Indeed, whilst cardiopulmonary resuscitation commonly refers to mechanical chest compressions alone, the presence of a DNACPR decision may also inadvertently influence decisions around intubation and ventilation, administration of intravenous fluids and/or vasoactive drugs ¹⁵ or even sampling for blood cultures. ¹⁶ Such unforeseen effects often undermine trust in the relationship between patients, their advocates and clinical teams. An awareness of these inherent biases and open transparent communication is important in ensuring that DNACPR decisions alone do not automatically extend into other aspects of care.

Paradoxically, it is worth noting that for some patients, this may be a desirable effect and an appropriately made DNACPR decision can encourage less invasive interventions and a less ‘medicalised’ death. In certain circumstances, this may help to promote patient-focused goals of care such as time with family or the possibility of care at home. The introduction of ReSPECT conversations has moved towards much more patient focused discussions with medical professionals ¹⁷ with a view to delivering exactly this.

An essential aspect of DNACPR decisions is the need for discussion with patients. Much emphasis is placed on the central concept of the right to self-determination and there has been a move away from the historical notion of medical paternalism. Even for patients without capacity, there still remains the need to involve them in decision-making as far as practical. However, there is consistent evidence that the process of DNACPR decision-making as currently reported is somewhat flawed. ¹⁸ The timing and location of DNACPR discussions varies significantly and may often require adjustment to the individual concerned. Despite the need to involve patients and their advocates in such discussions, there remains a significant proportion of decisions where this has not happened sufficiently. Factors that contribute include a desire to avoid causing distress or conflict – both with patients and their families. ¹⁹ Indeed, a report by the Ombudsman into DNACPR decisions highlighted that patients and families were often not told about such decisions. ²⁰ It also identified a lack of accessible information given at the time or before DNACPR discussions take place as well as issues with record-keeping and documenting decisions. A number of training needs were recommended with the aim of improving communication and record-keeping with respect to DNACPR decisions and greater patient involvement.

Less clear is to what degree such involvement extends. The purpose of a true discussion about a DNACPR decision is essentially either to furnish the patient with an understanding of the likely burdens, benefits and outcomes of CPR in order to elicit the patient’s wishes such that they have the option to refuse such treatment in advance. Or having provided the above, to engage the patient in deciding whether those burdens and harms likely to be suffered, alongside the level of recovery achievable, are acceptable to them. This is done through exploring their beliefs, values and wishes. In contrast, where CPR is not considered to offer any realistic prospect of success such that a DNACPR decision has already been made, the focus of discussion is communicating what has already been decided and explaining reasons. Here, patient involvement may be restricted to being informed of the outcome of a decision rather than actively participate in making it. Importantly, as will be discussed later, this can afford the patient the opportunity to seek a second opinion. We did not include a patient’s perspective on DNACPR decisions in the current manuscript, as we chose to focus on the ethical and legal challenges a clinician might face. However, we firmly believe that patients’ and relatives’ voices are an important piece that is missing from this complicated puzzle.

Ethical considerations

There are several bioethical considerations involved in making DNACPR decisions. First and foremost is the concept of autonomy. This has been defined as self-rule that is free from both controlling interference by others and from limitations, such as inadequate understanding, that prevent meaningful choice. ²¹ In practice, this will entail furnishing the patient to understand a key decision about their life. However, there are challenges to truly promoting autonomous choice. There is significant variation in health literacy and understanding of decisions over resuscitation. The public perception of the success of CPR can be hugely overinflated and bears little resemblance to what happens in practice. ²² Similarly, there is considerable variation in the ability of clinicians to sensitively but comprehensively communicate the risks and benefits of CPR in a manner that the patient or their family understand. The desire to avoid distressing language may contribute – clinicians sometimes talk about ‘keeping comfortable’ and ‘avoiding jumping on chests’, whilst not intentionally misleading, one can appreciate that patients may not perceive this as a discussion about CPR. Indeed, the promotion of true autonomous choice is dependent upon ensuring the individual concerned has a sufficient understanding of what they are choosing.

Importantly, one cannot assume willingness to participate in decisions about CPR. On occasions, patients and their relatives may not want to discuss or even consider the possibility of their death. ²³ In this study, a significant proportion of relatives of intensive care patients did not wish to participate in decision-making. In such circumstances, it can be challenging to identify where this is the case let alone to ascertain previous or likely wishes about a patient. In addition, it should be remembered that opinions may change over time. ²⁴ An uncertain prognosis can sometimes hinder decision-making and the relationship between autonomous choice and the passage of time is complex. Hence, it is difficult to know whether previously expressed opinions can accurately reflect one’s wishes at the time of making of such decisions.

Next, we can consider the principle of dignity. This has often arisen in conjunction with discussions about the process of dying and the desire to avoid burdensome, life-prolonging medical treatment. It can be difficult to define. One approach is to distinguish ‘basic or absolute dignity’ as the intrinsic moral worth of all people from personal dignity which is socially constructed, individualistic and transient. Dignity can be used to both defend and justify diametrically opposed actions or practices. Whether dignity is an extension of the sentient personal autonomy described above or a distinct ethical concept in its own right continues to be debated. ²⁵

Another ethical principle of relevance is that of non-maleficence – the principle of refraining from causing intentional harm towards another human being. Providing treatment that is considered ‘futile’ would seem to conflict this – no doubt this remains a central concern of many clinicians. However, the concept of futility itself is often debated – its meaning can be interpreted in a plethora of ways. ²⁶ What makes a treatment ‘futile’ can be difficult to define. Non-maleficence can apply in terms of not just physical harm, but psychological harm as well – this can arise from discussions around treatment as much as treatment itself. In addition, as mentioned previously, a DNACPR decision has been reported to also adversely affect clinical care in terms of other interventions. ¹⁴

Finally, there is the principle of Trust. This is considered to be the cornerstone of the doctor-patient relationship. However, arguably it has been eroded through changes in continuity arising from fragmented treating teams as well as recent high profile judicial cases. Patients have an expectation that clinicians will act in their best interests and reflect commonly shared norms such as respect and integrity. Clinicians are spending less time with their patients – in hospitals, not infrequently the initial interaction could be a conversation about DNACPR. In this context, it is easy to see how conflict can occur.

This list is by no means exhaustive. Other considerations weigh heavily such as the need to uphold justice and equity of access, the duty to preserve life and the obligation to act to only bring about benefit to an individual. There are ethical considerations relevant to those caring for such patients with the moral injury caused by invasive treatments with little or no prospect of success contributing to an increasing prevalence of burnout within the profession. Ultimately, it should be remembered that ethical principles represent a framework to cope with uncertainty under challenging conditions where outcomes are unknown and competing interests exist.

DNACPR as an advance decision

In view of the above, it is not surprising then that autonomy and the right to self-determination are central to decision-making at the end of life and therefore to DNACPR decisions. Advance decisions are intended to promote these principles and ensure that individuals receive only treatments that they would wish for themselves. The principle of ‘best interests’ enshrined in the Mental Capacity Act 2005 is in itself intended to safeguard autonomous choice. By explicitly declaring which treatments are acceptable and which are not, there can be improved clarity over making patient-centred choices in the event that the individual in question subsequently lacks capacity. Such decisions need not appear rational or reasoned to outside parties, the emphasis remains on the individual making the decision. For an advance decision with regards to life-saving treatments to be valid, there are specified criteria. The subject must have capacity at the time of making the decision, they must do so willingly and not under any form of duress and must provide a written statement in the presence of a witness that they understand the consequences of such a decision in that it is to apply even if their life is at risk. Any advanced decision must be specific with regards to the treatment in question and is only applicable in the circumstances that were envisaged when the decision was made. Of note, advance decisions constitute advanced refusals of treatment, there cannot be a right to demand treatment where it is not offered. In this way, there is the potential for conflict where the desire to support one’s autonomous choice must be balanced against the clinician’s concerns for avoiding harmful interventions. How to establish which ethical principle should prevail is inherently a very subjective determination. Where there is a documented, valid and applicable Advance Decision in place refusing CPR in the event of cardiac arrest, this is considered legally binding.

It should be remembered that a DNACPR decision is a means of communicating in advance when CPR should not be offered to a patient – either because this would be against their expressed wishes, or it is considered to have little prospect of success. ²⁷ When documented, it is often referred to as a DNACPR ‘order’ although they are not ‘orders’ per say but advisory guides. Within the UK, unless constituting an Advanced Decision, such a document has no legal authority – it merely serves as a record of previous decisions and/or wishes. Ultimately, unless the expressed or perceived wishes of the patient preclude it, the decision to offer or withhold CPR falls to the healthcare responders at the time of cardiac arrest. Whilst they would be likely to act in accordance with a DNACPR decision if present, there is no obligation to do so. In this way, it is more a recommendation to clinicians to help them make a best interest decision at the time of cardiac arrest, and perhaps would be better referred to as such. Interestingly, there is considerable variation in the reported frequency of DNACPR decisions between different regions with clinicians in some countries making such decisions considerably more than others. ²⁸ This may subsequently lead to variations in expectations from patients and families – and potentially an associated risk of conflict.

The ReSPECT framework

Determining a patient’s wishes about CPR and indeed other life-sustaining treatments is essential for promoting autonomy. However, the ICU is arguably not the best place to start DNACPR conversations, as patients are often unable to participate themselves and families may be struggling with emotional challenges. As with all advanced decisions, discussions about resuscitation should start ideally before the acute decline in health and be continuously re-visited. The ReSPECT (Recommended Summary Plan for Emergency Care and Treatment) tool is one such process, which facilitates ongoing exploration of patients’ views and wishes around their health. ¹⁷

This arose out of observations regarding deficiencies in conversations about DNACPR decisions, misinterpretations of such decisions, a misplaced focus on treatments rather than outcomes and significant variability in clinical practice. It is an evidenced based tool that is consistently evaluated and revised in response to feedback. ²⁹ ReSPECT is currently being used across many hospitals in the UK and aims to:

• a) put each individual at the centre of the conversation,

• b) achieve shared understanding between person and clinician,

• c) focus on treatments to be given, not just ones to be withheld,

• d) record agreed clinical recommendations

• e) be recognised across all boundaries.

Firstly, a shared understanding of the current situation is established. This goes both ways – ensuring that the patient understands their diagnosis and prognosis and that the doctor understands the patient’s current home set up and own views of their quality of life. A shared understanding of the patient’s values and fears is then explored. This highlights what matters to the patient and allows their individual wishes or concerns to be discussed. Finally, there is the opportunity to document the recommended treatments that are likely to result in the patient’s preferred outcomes. The balance between strategies prioritising extending life and those prioritising comfort is agreed with the patient. A decision about DNACPR is only one aspect of this – clearly there are further treatments beyond this that merit consideration. These discussions can then inform decisions about treatment in the event of critical illness both prior to admission to and following prolonged treatment on Intensive Care. Indeed, a ReSPECT conversation can take place at any time, and a public engagement campaign to ‘expect ReSPECT’ may help encourage patients to think about what health outcomes they fear and value.

DNACPR and the law

The legal aspects of DNACPR decision-making have been clarified in the judgement from R (Tracey) v Cambridge University Hospital NHS Foundation Trust and others [2014]. ³⁰

The facts were as follows:

Mrs T suffered from terminal lung cancer and had a life expectancy of around 9 months. On 19th Feb 2011, she was involved in a road traffic accidence and suffered a cervical fracture amongst other injuries. She was admitted to Cambridge University Hospital. Due to chronic respiratory problems, Mrs T was placed on a ventilator but did not respond to treatment for a chest infection. A decision made to take Mrs T off the ventilator. At the time, the focus of discussions was on extubation or otherwise. There were two attempts at extubation. The court was provided with evidence that Mrs T expressed a wish for full active treatment. On the 27th February a DNACPR notice was prepared for Mrs T. It was disputed as to whether Mrs T was consulted or agreed with this decision. Subsequently, Mrs T was successfully weaned off the ventilator and her condition improved such that her daughter was able to return to her home abroad.

On 2nd March, Mrs T’s family became aware of the DNACPR decision that had been made and objected to the DNACPR notice saying this was not what they or Mrs T wanted. As a result, it was withdrawn. Shortly afterwards, Mrs T withdrew from wanting to discuss resuscitation status and unfortunately her condition deteriorated. She no longer had capacity and a second DNACPR notice was imposed after discussion with family. Mrs T died on 7 March 2011, with no CPR given.

Proceedings

Mrs T’s family brought a claim against the Trust alleging a breach of her Article 8 European Convention of Human Rights – that is the right to respect for her private and family life – because by imposing the DNACPR notice, it had failed to adequately to consult Mrs T or family; notify Mrs T of the notice; offer a second opinion; make its DNACPR policy available to her; and have a policy which was clear and unambiguous. A further claim was made against the Secretary of State for a breach of Article 8 rights by failing to publish clear national guidance regarding DNACPR decisions. The claim against the Secretary of State was dismissed. However, the claim against the Trust partly succeeded, the Court of Appeal holding that it was unlawful not to involve Mrs T in the decision to impose the first DNACPR notice. This was in breach of Article 8 ECHR because there should be a presumption in favour of P’s involvement, given the life and death decision. However, it would be inappropriate (and not required under Article 8) to involve P if to do so would cause P to suffer physical or psychological harm. This is different to causing distress. Finally, the Court held that where it is considered that attempting CPR is futile, you should inform P of that decision, enabling P to seek a second opinion where possible.

The Master of the Rolls from the Court of Appeal said the following:

Drs should be wary of being too ready to exclude patients from the process on the grounds that their involvement is likely to distress them. Many patients may find it distressing to discuss the question whether CPR should be withheld from them in the event of a cardio-respiratory arrest. If however the clinician forms the view that the patient will not suffer harm if she is consulted, the fact that she may find the topic distressing is unlikely to make it inappropriate to involve her.

From this, it can be seen that for patients with capacity, discussions about DNACPR decisions must be had where appropriate as early as possible. One should explain the risks, benefits and burdens of CPR following national guidance such as that from the Resuscitation Council UK. ³¹ There is a need for sensitivity to the possibility that some patients may not want to have that discussion. Will telling the patient cause psychological harm or merely cause them some distress? The judgement in Tracey indicates that only where there is a genuine concern over harm to the patient should they be excluded from discussions about DNACPR decisions. All discussions, and the reasons for not having such a discussion if appropriate, should be documented clearly and using the correct forms.

For patients who lack capacity, the legal framework is discussed in the case of Winspear v City Hospitals Sunderland NHS Foundation Trust [2015]. ³² Briefly, this concerned the case of Carl Winspear, a 28-year-old gentleman with cerebral palsy, epilepsy and several spinal deformities who presented with a likely chest infection. A DNACPR decision was made by the treating team on the basis that CPR would not have any realistic prospect of success due to his pre-existing medical conditions and frailty. This was documented at 3am and it was not discussed with his mother nor anyone representing his interests. His mother brought a claim against the Trust for making such a decision without any consultation. The claim succeeded with the judge highlighting the importance of the best interest approach. He explained that there was nothing to suggest that it was not practicable nor appropriate to have attempted to contact Mrs Winspear at the time of making the DNACPR decision. Failure to do so constituted a breach of the patient’s Article 8 rights.

This case highlighted the need for shared decision making and consulting others to consider the interests of the individual lacking capacity beyond just the medical ones. One needs to consider personal wishes and feelings as well as their likely values. When making a best interests decision and consulting with family or friends, the question is not what you would want for the patient but what the patient would want themself? There should be a discussion with others as if talking to the patient, giving information about the nature of CPR and the burdens and risks for the patient. It should be made clear that they are not deciding on the clinical benefits of CPR. Instead, efforts should be made to invite their input on the patient’s overall wishes, values and feelings before deciding what would he/she would want were they able to make the decision themselves. In doing so, it is essential to explain the limits of what treatments are on offer and available to the patient.

Disagreements

A situation may occur where a patient with capacity or his/her proxy decision-maker requests CPR is attempted despite burdens and risks, but the medical consensus is that CPR would not be clinically appropriate. Here, one should explore the reasons for the request and their understanding of the consensus view. If after further discussion, you still consider attempting CPR would not be of overall benefit for the patient, you are not obliged to offer to attempt CPR in the circumstances envisaged. You should explain your reasons and any other options available including their right to seek a second opinion or take legal advice. Second opinions should be facilitated where appropriate, with support given to the family and/or their family in obtaining such. It is best that everyone involved approves who should provide that second opinion and, ideally, has an input into the questions asked of that clinician before they give their opinion. That can improve the chances of all accepting that individual’s opinion once given. Of course, that is not always possible. There are currently no national guidelines with regards to specific mandated requirements for a second opinion to be considered valid.

It should be remembered that a patient cannot demand treatment that is not clinically appropriate nor available. There are occasions when an application to the Court of Protection can be made, seeking a declaration as to what treatment (or lack of treatment) is in a patient’s best interests. In such circumstances, a Court can only consider available treatment options though. If there is no-one prepared to provide CPR to that patient, CPR is not an available option. There might be occasions though where the clinical team would not ordinarily provide CPR unless a judge of the Court of Protection indicates that such is in a patient’s best interests. In those circumstances, CPR is, in theory, an available treatment.

It is important to document everything in the medical records and any advance care plan. This should also be shared with community medical teams and the ambulance service where possible. It should be emphasised that the DNACPR decision applies only to CPR. Importantly, a DNACPR decision is not legally binding and does not override clinical judgement about CPR where the circumstances of the cardiac arrest are not those envisaged when the DNACPR decision was made. For example, if the patient suffers a cardiac arrest from an easily reversible cause – for example, induction of anaesthesia during a planned procedure, it may be reasonable to perform CPR even in the presence of an active DNACPR decision.

Conclusion

In recognition of the need to promote individual autonomous choice and avoid traditional medical paternalism, instruments such as Advanced Decisions and DNACPR decisions can have a significant role. In the UK, unlike in other European countries, DNACPR decisions or recommendations are a recognised means of recording previously expressed wishes or a clinical decision with regards to the benefits and burdens of CPR. However, despite the theoretical benefit of such decisions, several challenges remain, especially in the highly technical ICU environment. These include the relevance of DNACPR decisions in continuously monitored patients, the conflict they may create for a relatively rare occurrence and the need to balance contrasting ethical principles. More data around the mortality and neurological outcomes of patients suffering in-ICU cardiac arrests, and further exploration of the perceived obstacles by clinicians and families around DNACPR implementation, are necessary to improve communication and promote true autonomous choice where appropriate.