Decompressive craniectomy is a technically straightforward procedure whereby a large section of the cranium is temporarily removed in cases where the intracranial pressure is dangerously high. While its use has been described for a number of conditions, it is increasingly used in the context of severe head injury. As the use of the procedure increases, a significant number of patients may survive a severe head injury who otherwise would have died. Unfortunately some of these patients will be left severely disabled; a condition likened to the RUB, an acronym for the Risk of Unacceptable Badness. Until recently it has been difficult to predict this outcome, however an accurate prediction model has been developed and this has been applied to a large cohort of patients in Western Australia. It is possible to compare the predicted outcome with the observed outcome at 18 months within this cohort. By using predicted and observed outcome data this paper considers the ethical implications in three cases of differing severity of head injury in view of the fact that it is possible to calculate the RUB for each case.
- Risk of unacceptable badness
- decision making
- quality/value of life/personhood
- informed consent
- allocation of health care resources
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- Risk of unacceptable badness
- decision making
- quality/value of life/personhood
- informed consent
- allocation of health care resources
Decompressive craniectomy is a technically straightforward procedure whereby a large section of the cranium is temporarily removed. Its use has been described in a number of pathological conditions where the intracranial pressure is intractably raised and the patient is likely to die without surgical intervention. In the context of ischaemic stroke, for many years the possibility that the use of this procedure would produce an increasing number of very severely disabled survivors was a major source of discussion.1 2 However, the results of recent trials have demonstrated that when used selectively, the procedure not only increases the number of survivors but also the number of patients with a favourable functional outcome.3–5
For patients with traumatic brain injury the question remains unanswered.6 7 Severe brain injury is a major scourge in modern society and heroic measures to try and save lives are becoming more widely used.8–11 Unfortunately, as in any acute intervention in the face of uncertainty, the concept of the RUB (Risk of Unacceptable Badness)12 becomes pertinent. This is especially so in the context of a severe neurotrauma when there is a more global cerebral injury. A decompressive craniectomy in this situation risks leaving the patient with an unacceptable outcome requiring long term care in a clinical condition that the patient would themselves find unacceptable. There is therefore a challenging ethical question to be asked about whether to proceed or abstain from intervention where the outcomes can vary so dramatically between two differing pathological conditions.
This problem is currently being addressed by two ongoing randomised controlled trials evaluating decompressive craniectomy for severe head injury.13 14 If the results of these trials confirm that the procedure is beneficial in some cases, then the clinical decision-making will need to reflect that knowledge. The responsible clinician or team will need to ask:
Should consideration be given to an intervention decompressive procedure not being performed because there is a significant chance that the outcome is likely to be poor and would be unacceptable to the person on whom the procedure was carried out.
Until recently, outcome has been difficult to predict and this problem has had to be considered in a relatively surgical fashion—that one must do what seems necessary to safeguard life and enhance survival. Any discussion of whether the intervention was of real benefit to the patient has had to occur after the event when the person has either made some recovery or been left in a state that, had s/he have been asked, they would probably regard as unacceptable.15 In an attempt to address this problem many factors important in outcome prediction have been proposed. Traditionally, the presenting Glasgow Coma Score has been used10 16 17 however, recent studies have demonstrated its limitations given the changes in management of patients with head injury over the past decade.18 19 Aggressive pre-hospital treatment involving early sedation and mechanical ventilation tend to obscure the real Glasgow Coma Score (GCS) assessment thereby limiting its predictive significance. Other studies have emphasised the importance of age8 9 17 pupillary reaction,16 17 extracranial injuries17 20 and radiological appearances16 17 and while a number of prediction models have been developed most are relatively cumbersome, based on limited empirical data and tend to emphasise survival as the clinical end point rather than quality of life.21
The CRASH collaborators have provided the first user-friendly web based outcome prediction model that had been internally and externally validated on a very large number of patients, in both high and low income countries.22 By applying this model to a cohort of patients in Western Australia, it has been demonstrated that outcome can be predicted with a relatively high degree of accuracy.23 24 Specifically the long term risk of poor functional outcome can be predicted, converting the RUB from philosophical concept, into a clinical reality. For the 137 patients in the cohort study24 who have clinical follow-up at 18 months, the RUB can be calculated:
When the predicted risk of an unfavourable outcome at 6 months was less than 60%, the observed outcome at 18 months revealed 58 of 59 (98.3%) survivors had a favourable outcome and only one patient survived with an unfavourable outcome (the RUB=1.7%). A favourable outcome is defined as a Glasgow Outcome Scale (GOS) of good outcome (able to return to work or study) or moderate disability (independent but unable to return to work or study). An unfavourable outcome is defined as a GOS of severe disability (dependent, requiring full time care), vegetative (unaware of surrounding environment) or dead.
When the predicted risk of an unfavourable outcome at 6 months was between 60 and 80%, at 18 months, 10 of 45 survivors had an unfavourable outcome (the RUB=22.2%).
When the predicted risk of an unfavourable outcome at 6 months was greater than 80%, at 18 months, 19 of 22 survivors had an unfavourable outcome (the RUB=86.7%). Of the three survivors with a favourable outcome all were moderately disabled.
This data can help with the difficult ethical decision that the clinician faces when considering a decompressive craniectomy in a patient with a severe head injury. There are two ways of reflectively using experience to address the ethical problem:
The RUB—an acronym for the Risk of Unacceptable Badness.12 This reflects the likelihood of a patient surviving a severe head injury but being left severely disabled: a condition which they would find unacceptable. The implication being that, were it possible, the physician could not reasonably presume that they would have informed consent for life saving but non restorative operative intervention;
Illustrative case histories and clinical intuition.
The RUB (the risk of living in an unacceptable state if you live at all) can be calculated (and is above) for the three groups given the plausible assumption that severe disability or persistent vegetative state would be unacceptable to most people.15 25 While accepting the obvious limitations of the GOS, it does provide a standardised, reproducible assessment tool. Someone who remains severely disabled by definition requires full time care and while this category covers a relatively broad functional range, many of these patients are young people who are eventually placed in a nursing home facility. Many are only mobilised in a wheelchair, they have severe contractures and are incontinent. They are fed via a gastrostomy tube or with assistance and they can only follow single stage commands. This represents to most people the RUB, an unacceptable and, to the families, a possibly unbearable outcome.
Someone who recovers to moderate disability cannot by definition return to work or study but can live independently. These patients can have significant physical and neuropsychiatric problems such as poor memory, attention, executive/cognitive function, behavioural control and regulation of mood. There is also the accompanying social disruption and isolation, with the associated loss of self-esteem. While accepting of these problems, it would be difficult, if not impossible, to suggest a decompressive procedure should not be offered because the patient and their families will have to come to terms with these difficulties.
This ethical consideration arguably should be part of the discussion that allows the operation to go ahead for any person who is a possible candidate for decompressive craniectomy—notice that it is completely separate from any decisions of a rationing type.
Three cases focus a case based discussion, as in (ii).
Case history 1
A 17-year-old male fell while intoxicated. His GCS was noted to be seven (E1, M5, V1), both pupils were reactive, and he had no other significant injuries. CT scan of the brain revealed diffuse cerebral swelling and patent fluid spaces (indicating that brain swelling was not extreme). Despite aggressive medical management, the intracranial pressure (ICP) was dangerously high. On day five following the injury he had a decompressive craniectomy. Following surgery his intracranial pressure stabilised and he was discharged from intensive care 10 days following the injury. He made a rapid recovery and went home on day 25. At 6 month follow-up he resumed his studies as an apprentice welder.
The predicted risk of an unfavourable outcome at 6 months was 46.1%.
Case history 2
A 52-year-old female fell from a moving vehicle while intoxicated. She was transferred to a local hospital and her GCS was 11 (E2, M6, V3). Both pupils were reactive and she had sustained multiple fractured ribs and lung contusions. CT scan of the brain revealed diffuse cerebral swelling and midline shift. Her intracranial pressure remained stable for 24 h but despite aggressive medical management the ICP measurements remained high. Thirty-six hours following the injury, she had decompressive surgery following which the intracranial pressure remained stable and she remained in an intensive care unit mainly because of her lung injury. She was discharged to the ward on day 15 and remained an inpatient for 20 days during which time there were significant problems requiring 24 h nursing cover and intermittent use of physical restraints. She was then transferred to a rehabilitation facility where she made a slow recovery and was discharged home 5 months following the injury. At 18 month follow-up she remained moderately disabled with ongoing neurocognitive problems. She has been able to return home and can look after herself; however, she has significant neuropsychiatric problems. Her behaviour is erratic and unpredictable and she is prone to unprovoked episodes of violence, most commonly directed at her mother. It is thought that she is unlikely to return to her studies at university.
The predicted risk of an unfavourable outcome at 6 months was 71.8%.
Case history 3
A 28-year-old was involved in a motor vehicle accident of such violence that she was ejected from the vehicle. Her GCS was four (E1, M2, V1) and both pupils were unreactive but she had no other significant injuries. CT scan of the brain revealed diffuse cerebral swelling, loss of fluid spaces and midline shift. She was taken to theatre 5 h after the accident and the larger haemorrhages decompressed. Post-operatively the ICP remained below 20 and she was weaned from ventilatory support and discharged to the ward on day 10 following the injury. Unfortunately she made a poor recovery. At 1 month following the injury she opened her eyes spontaneously but would not track objects. She would flex upper and lower limbs to painful stimuli but remained reliant on her tracheostomy to clear upper airway secretions. A percutaneous gastrostomy tube was placed for artificial nutrition and hydration. She had sufficient neurological function to participate in rehabilitation. She remained an inpatient for 180 days and was discharged home but at 18 month follow-up she remained severely disabled and reliant on round-the-clock nursing care. She was unable to stand unaided and needed two people to help her transfer for bed to wheelchair. She had developed severe contractures and was incontinent. She could mumble a few words but could not manage more than single-stage commands.
The predicted risk of an unfavourable outcome at 6 months as 93.6%.
There is little doubt that acute procedures such as decompressive craniectomy are becoming an established tool in the management of severe head injury.8–11 However, while numerous studies suggest that clinical outcome can be improved, an injury can be so severe that whatever is done any survivors will be left severely disabled. The RUB has an obverse concept, substantial benefit—an outcome that now or in the future the patient would consider worthwhile. The latter is the basis of authorisation for treatment to which any decision-maker, including family or guardian, should always hold themselves. The RUB makes us consider the risk of severe disability, a prolonged state of indignity or suffering that a patient would find intolerable and substantial benefit makes us consider that outcome as the patient would see it. The CRASH prediction data brings scientific evidence to bear on the ethical dilemmas of rescue in a particular set of cases and allows families and clinicians to get away from the thought that ‘any chance at life is better than none’. The chance of life may be bought with a terrible cost or risk in the groups with the adverse CRASH data so that it is not a chance of ‘life as we know it’. To ignore the ethical problem means that fraught and difficult decisions have to be made down the line: do we continue ICU care? What should be taken as an indicator that we should stop striving officiously to keep alive? Does the beginning of rescue imply that we have to push it to the limit or is there a point where we should say ‘We have given it our best shot and we are not going to win here?’ These questions should be confronted and even pre-emoted by entering into what is best thought of as a ‘contract of care’ or ‘trial of treatment’ so that everybody expects the issue of whether to persist in the face of the eventualities that arise in the patient's care.
The studies referred to provide statistical data about probable outcomes following acute neurosurgery and illustrate the ethical issues surrounding emergency practice in high morbidity areas. While statistics should not replace clinical judgement, the prediction of an unfavourable outcome at 6 months can be used as an index of injury severity. By comparing the predicted risk (ie, the severity of the head injury) and the observed outcome (ie, what is likely to be achieved following decompressive surgery) the clinician is in a better position to discuss realistic outcome expectations among decision makers who are often searching for realistic help from the emergency team in making these difficult choices.
The three cases make the ethics vivid. In the group represented by case 1, most survivors had a good outcome at 18 month follow-up (98.3%). While there is a risk of moderate or severe disability, it would be unwarranted to withhold intervention given the observed high probability of a good outcome. In the second case there is considerable variability in observed outcome and the clinician can clearly state prior to intervention that the outcome is unpredictable. While there is a possibility of a good outcome the patient may likewise remain severely disabled or vegetative and that risk needs to be borne in mind when decisions are made. In such a case it becomes understandable that ordinary people, if per impossibile they could be consulted, might opt not to run the risk, especially when we link this data to stage of life criteria. We see intuitively that a young person with a better chance of an acceptable, even if burdensome outcome, might legitimately be credited with an interest in taking their chances, whereas to make that decision for an older person might look far less appealing. This group poses the most difficult clinical problems but also allows us to recognise, when the ethical complexity is taken into account, that good clinicians might vary in their recommendations and that subsequent medico-legal or ethical reflection should acknowledge the uncertainties. That recognition would, we are certain, be of considerable reassurance to those faced with advising patients and families in such trying circumstances.
In the final group of cases, the data provides clearer guidance. The clinician can state prior to intervention that if the patient survives, there is a small possibility of survival with moderate disability but the most likely outcome is severely disability or persistently vegetative state, a state that most would regard as unacceptably bad.
There remain, of course, unanswered questions and important ethical challenges.
The heroic journey: if a significant number of patients improve over the following years, this could be likened to a difficult journey to a state of increasing quality of life rather than an unacceptably bad state. But if indeed a few patients do improve or (as seen in very rare instances) actually ‘recover’ to full function is that too much to ask of anyone and what are the resource implications of such ‘against the odds’ heroism? We have argued that the invoking of such resource implications in early clinical decisions could be obviated by better directed use of the dramatic interventions where the ethical decisions are related primarily to the authorisation required for treatment on the basis of presumed consent.
The further ethical implication for a clinician or multi-disciplinary team involved in these rescue decisions where there are high stakes for the person who has to endure their illness is our attitude to decision-making. There is an urge to rescue by acute care specialists and family but perhaps that is best accommodated by a trial of treatment whereby they do their best and revisit the decision once the effect of the intervention is seen. This occurs quite naturally when transitions in care (such as tracheostomy after initial stabilisation, when the patient is moved from ICU to subacute care and when there are life threatening problems such as chest infection) arise if the situation is properly described at the outset. Such points of review and reconsideration arise less frequently as care goes on but the team who establishes the practice of realistic updates and outcome predictions will find that decision makers feel less distressed and insecure and more empowered in their decision-making on the patient's behalf. At each stage the pattern set by the RUB is useful: ‘Where are we going?’ ‘How certain are we?’ ‘Ought we to reconsider what the patient would have wanted?’
Acute care teams need to try and take account of what a patient's decision might be if that patient could be asked. Only then can they presume ongoing consent for their interventions. The two thoughts of substantial benefit and unacceptable badness are useful pointers that allow decisions to be made that accord with best ethical reasoning and patient centred values. The process of decision-making is one of partnership in which the healthcare team acts responsibly and draws upon their expertise and the evidence they have to contribute to a collaborative discussion, between themselves and then with the patient's decision-makers, as to what they should do. This is a patient-centred ethic of partnership in practice. It is noteworthy that the major legal decisions in this area concur with these principles and help us to deal with a much stressed and often confused situation with some clarity and compassion. At the end of the day, an ethically sound decision should be able to be included in the stories of those involved in such a way as to respect their integrity. That constraint ensures that, for the person whose life may end and for those who are left grieving there is closure and a sense of having done what is right in a tragic situation.
Competing interests None.
Ethics approval This study was conducted with the approval of the Royal Perth Hospital Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.