Article Text
PDF
Abstract
Objective To assess the issue of ‘retrospective consent’ among a cohort of patients who had survived with unfavourable outcome and to assess attitudes among next of kin regarding their role as surrogate decision makers.
Methods Twenty patients who had survived for at least 3 years with an unfavourable outcome following a decompressive craniectomy for severe traumatic brain injury were assessed with their next of kin. During the course of a semistructured interview, participants were asked whether they would have provided consent if they had known their eventual outcome. They were also asked for general comments regarding all aspects of the clinical journey. Eighteen patients had next of kin who were available for interview. For two patients, there was no longer any family involvement.
Results Of the 20 patients, 13 were able to provide a response and 11 felt that they would have provided consent even if they had known their eventual outcome. Of the 18 next of kin who were able to express an opinion, 10 felt that they would have provided retrospective consent.
Conclusions Many patients appeared to have adapted to a level of disability that competent individuals might deem unacceptable. This does not necessarily mean that such outcomes should be regarded as ‘favourable’, nor that decompressive craniectomy must be performed for patients with predicted poor outcome. Nevertheless, those burdened with the initial clinical decisions and thereafter the long-term care of these patients may draw some support from the knowledge that unfavourable may not necessarily be unacceptable.
- Clinical Ethics
- Informed Consent
- Neuroethics
Statistics from Altmetric.com
Introduction
There continues to be considerable debate regarding the role of decompressive craniectomy in the management of severe traumatic brain injury.1–3 There would appear to be little doubt that it can be a life-saving intervention;4–6 however, the concern has always been that many patients may survive with a level of disability that they and their families feel to be unacceptable.7–9 The clinical and ethical issues that arise from this dilemma are compounded by the limitations inherent in contemporary outcome assessment. Traditionally, the most widely used outcome measure has been the Glasgow Outcome Scale,10 and more recently, the Extended Glasgow Outcome Scale,11 both of which are felt to be robust and reproducible assessment tools.12 They do however tend to focus on the physical and functional aspects of recovery but give limited insight into neuropsychological and psychosocial impairments. In addition, the individual categories cover very broad ranges of neurological function and provide limited information on overall quality of life.13 ,14 Notwithstanding these shortcomings, numerous investigators have used these assessments and outcome is generally dichotomised into favourable or unfavourable. The obvious implication is that the aim of the procedure is to achieve a favourable outcome with an unfavourable neurocognitive outcome being unacceptable. Indeed, the senior authors have argued on a numerous occasions that this is a reasonable position to adopt on behalf of a notional ‘objective’ patient.8 ,9 ,15
However, the results of a recent study may require some of these assumptions to be called into question or perhaps at least re-examined.16 This study explored the issue of retrospective consent among a small group of patients who had survived with severe disability (unfavourable outcome) following a decompressive craniectomy for severe traumatic brain injury. Patients were asked whether they would have given consent for surgical intervention if they had known their eventual outcome. Among those who were able to respond to the question, most indicated that they would provide ‘retrospective’ consent so raising the question as to whether ‘unfavourable’ is necessarily ‘unacceptable’.17 ,18
The aim of this study was to qualitatively analyse some of the responses provided by the patients and their families and explore some of the ethical issues.
Methods
The patients in this study were drawn from a cohort of 147 patients who had required a decompressive craniectomy for severe traumatic brain injury between the years 2004 and 2009 and who were at a minimum of 3-year follow-up (figure 1).16 Thirty-nine of these patients had previously been adjudged to be either severely disabled or in a vegetative state at 18 -month follow-up. After obtaining hospital research ethical committee approval, these patients and their next of kin were approached and invited to participate in the study.
Flow diagram demonstrating the method by which patients were selected.
The participants and their next of kin were assessed either in the outpatient department or their own home. The assessments took the form of a semistructured interview administered by two independent occupational therapists who had not been involved in the patient’s primary care (CJ and KK).
The functional assessments performed were The Extended Glasgow Outcome Score (figure 2), The Barthel Index and the Short Form 36 (the full results of these assessments have previously been reported16). In addition, the issue of retrospective consent among these patients and their next of kin (or primary carers) was assessed by asking the following questions:
The Extended Glasgow Outcome Scale.
Questions to the patient
At the time of your initial injury you were unable to give any indication that you would or would not agree to what was considered ‘life-saving’ surgery. In retrospect (looking back), would you have agreed to have the surgery performed knowing your eventual outcome?
Questions to the next of kin or primary carer
Relatives that had been involved in providing consent for the initial decompressive surgery were asked the following questions:
Do you feel you were given sufficient information regarding the possible long-term outcome of the surgery? Specifically the possibility of survival with severe neurological disability?
If you had known that your relative would have had this outcome, would you have provided consent for the surgery?
At the time of the injury your relative was not able to provide consent or permission for the surgery. Knowing your relative as they were prior to the injury and knowing how they are now, do you think they would have provided consent if it had been possible to ask them?
Do you have any general comments regarding any aspect of your relatives care?
At each stage, patients and their carers were asked to provide additional comments regarding each issue.
Statistical analysis
Given the small number of the sample size and the qualitative nature of the information, no formal statistical analysis of the data was performed.
Results
Twenty patients or their next of kin consented to participate, 7 patients had died and 12 patients were either lost to follow-up or declined the invitation.
The characteristics of the 20 participants and the 7 patients who had since died are shown in table 1. Among the 15 patients who had previously been adjudged to be severely disabled, none had improved to a level of moderate disability. The five patients who had been in a vegetative state remained so. The comments from the next of kin are numbered according to which question they were specifically asked (1–3). Responses labelled (4) are general comments provided at the end of the interview.
Characteristics of the 20 participants and the 7 patients that had since died
Of the 13 patients who were able to provide a response, 11 felt that they would have provided consent even if they had known their eventual outcome. Of the 18 next of kin who were able to express an opinion, 10 felt that they would have provided retrospective consent (figure 3). Two patients no longer had any contact with their families.
Responses to the following questions:
Responses from the patients Q1 (pt) At the time of your initial injury you were unable to give any indication that you would or would not agree to what was considered ‘life-saving’ surgery. In retrospect (looking back), would you have agreed to have the surgery performed knowing your eventual outcome?
Responses from the next of kin Q1 (NOK) Do you feel you were given sufficient information regarding the possible long-term outcome of the surgery? Specifically the possibility of survival with severe neurological disability? Q2 (NOK) If you had known that your relative would have had this outcome would you have provided consent for the surgery? Q3 (NOK) At the time of the injury your relative was not able to provide consent or permission for the surgery. Knowing your relative as they were prior to the injury and knowing how they are now, do you think they would have provided consent if it had been possible to ask them?
Quite aside from the distribution of retrospective decisions for or against surgery, there are some interesting qualitative responses of interviewees that cluster around opinions regarding the decision to surgically intervene and that of the eventual outcome (table 2).
Qualitative comments and responses from next of kin regarding information received and retrospective consent
Discussion
Should decompressive craniectomy be performed?
A decompressive craniectomy represents an aggressive surgical intervention with well-documented morbidity.19 ,20 It is usually only considered once medical therapy has failed to control intracranial hypertension and the patient is thought to be unlikely to survive without surgical intervention. When this occurs in the often-fraught setting of acute neurotrauma, there are a number of interrelated and sometimes conflicting issues that can arise.
Family and friends (and the clinicians in some cases) may focus on the life-saving nature of the intervention and fail to recognise that the surgery may, in certain circumstances, be far from restorative.4–6 The clinicians burdened with the decision to surgically intervene may be faced with the possibility that they may leave a patient with an unacceptable level of disability.21 Finally, the patients themselves will be acutely injured and as such will be in no state to provide an opinion as to what they would or would not deem to be an acceptable outcome.
Counselling and initial decision making
There were a number of difficulties encountered when consideration was being given to surgical intervention. Many respondents felt that outcome was often dichotomised into life or death, and there was insufficient information provided regarding prognosis and long-term quality of life. A second issue was the emotional strain felt by many of the respondents and the effect that this had on their ability to provide informed consent.
Predicting long-term outcome
It may well be that the possibility of survival with severe neurological disability had been discussed and the respondents have either forgotten or did not take in the information at the time. Alternatively, the fact that many respondents felt that they were not given sufficient information about what to expect may merely be a reflection of genuine uncertainty on behalf of the treating clinicians. The difficulty will always be determining at what point in the timeline of these difficult events, initial speculation becomes reasonably accurate prognostication and also establishing exactly what was communicated and understood at the time.
Until recently outcome prediction had to be considered in a relatively subjective fashion based on individual prognostically important factors such as age, initial Glasgow Coma Score (GCS), pupillary reactivity, extracranial injuries and radiological features.4–6 However, recently developed web-based outcome prediction models have gone some way to improve this process.22 ,23 The CRASH collaborators (corticosteroid randomisation after significant head injury) model provides a percentage predicted risk of an unfavourable outcome22 (defined by the Glasgow Outcome Scale as dead, persistent vegetative state or severely disabled). The higher the predicted risk, the greater the degree of primary brain injury and this can serve as a surrogate index of injury severity. Comparing the predicted risk of unfavourable outcome with the observed long-term outcome can provide an objective assessment of outcome following decompressive surgery (figure 4).24 ,25
A bar chart demonstrating how the CRASH collaborators web-based outcome prediction model can be used to stratify patients according to the severity of primary brain injury. As the prediction of an unfavourable outcome (used as a surrogate index of injury severity) increases (x axis), the numbers of patients who are observed to be either severely disabled, vegetative or dead also increase. Numbers within the bar chart represent absolute patient numbers. (Reproduced with kind permission, Mary Ann Liebert, Inc. Publisher.)
Using the predicted risk of an unfavourable outcome allows patients to be stratified according to injury severity, and this type of data analysis can be useful when considering treatment efficacy especially among large groups of patients. There are, however, limitations when applying mathematical prediction models to individual cases and this is highlighted by the data in table 1. A number of patients had a relatively low predicted risk of unfavourable outcome, and yet their eventual outcome was that of severe disability or a vegetative state.
We would certainly agree with the CRASH collaborators that this type of predictive model should not replace clinical judgement; however, we do believe that it can provide some supportive information when considering realistic outcome expectations. For example, once a patient has a prediction of an unfavourable outcome of greater than 80%, the degree of primary brain injury is such that the most likely outcome for the patient if they survive is one of severe disability. Indeed, the responses of many of the participants in this study would seem to indicate that more emphasis should have been placed on the issue of poor quality survival and the patient's conception of benefit rather than the surgery proceeding as a fait accompli with the ethical issues sidelined or not even mentioned.
Consent for decompressive craniectomy: prospective and retrospective
The role of family members in providing consent was another issue that was highlighted as particularly problematic. In the context of decompressive craniectomy, the normal framework of autonomy and informed consent are inapplicable because patients are acutely injured and will be in no position to express views about either their treatment or eventual outcome. In addition, most of these patients will be young adults who are unlikely to have expressed a view or documented their wishes regarding treatment choices were they to be placed in such a position. In these circumstances, either the clinician (who is legally responsible in most jurisdictions) must act in the best interests of the patient or the relatives must provide substituted judgement regarding management decisions and outcome.
The considerations here are not only clinical but also ethical and they require us to exercise well considered judgement about our actions. Before we expose a patient to the possibility of survival with severe disability, we have an obligation to establish that they would regard this as acceptable.
Substituted judgement
In an attempt to explore this issue, several studies have used the type of data presentation as shown in figure 4 to assess opinion among competent individuals regarding the possibility of survival with severe neurological disability.26 In a number of semistructured interviews, a large group of healthcare workers were provided with a number of scenarios involving severe traumatic brain injury in which consideration was being given to performing a decompressive craniectomy. In one of the scenarios, the participants themselves were the injured party and the implication, which was clearly stated, was whether they would provide consent for surgery given that survival may come at the cost of living with severe disability. Initially no prognostic information was provided other than clinical details, and opinion was recorded on a visual analogue that was marked at one end ‘strongly agree’ on the other end ‘strongly disagree’. Participants were then shown the CRASH model, the prediction of unfavourable outcome and the observed outcome as shown in figure 4. They were then asked to reconsider their response. The study demonstrated that participants were very risk aversive and once the prediction of unfavourable outcome was greater than 80% very few felt that they would provide consent based on the likelihood of survival with severe neurological disability.
It could of course be argued that this was a relatively hypothetical situation and not representative of the real-life tension that occurs in the context of acute neurotrauma. However, it could equally be argued that in the real-life acute setting, hasty decisions that cannot be mitigated, may be made under the pressure of limited time and perhaps limited competency (due to emotional distress) to reflect on the true implications of their decisions. Thus, the responses of the participants in these studies may be a better reflection of how many people genuinely feel about this outcome for themselves.
While there are limitations when making somewhat abstract statements such as “I would rather die than live with severe disability”, this cannot detract from the fact that people have a right to make decisions regarding their healthcare when they are competent to make a judgement regarding personal risks and benefits. Based on this reasoning and in the absence of a patient’s known wishes, it could be argued that once a patient’s prediction of unfavourable outcome is greater than 80%, consideration might be given to withholding intervention because the most likely outcome is that of severe disability and most people would find this unacceptable.
The results of the current study would not however necessarily support this position.
The eventual outcome
Patient’s best interests
Despite the initial psychological and emotional trauma faced by patients and their relatives, one of the fundamental questions on which this study aimed to focus was whether the patients themselves felt that the ‘heroic journey’ had been worthwhile. In order to reduce bias, all patients and their families were independently assessed by occupational therapists who had not been involved in the patient’s initial care. The assessments were all performed at a minimum of 3-year follow-up, so patients would have had ample time to address early issues of anxiety, depression and dependence and perhaps learn to adapt to their disability. In this context, when patients were asked whether they would have provided consent to what was believed to be a life-saving intervention knowing that they would survive with the disability that they have, most patients that could provide an answer provided a positive response. Based on these findings, it is perhaps difficult to adopt the position that life-saving surgical intervention was not in these patients’ ‘best interests’.
Retrospective consent
These findings were, at least to the senior authors, unexpected and could on face value lend support to the ongoing use of decompressive surgery even if there are an increasing number of survivors with severe neurocognitive disability. However, this does require a degree of qualification.
There can be little doubt that obtaining a positive response when asking a patient whether they would agree to an intervention that has enabled them to stay alive, although with a considerable alteration in functional status, is certainly a testament to the human will to survive and adapt to adverse and challenging circumstances. It may well be that they have adapted to a level of neurological disability that they might previously have deemed unacceptable and this has been observed in previous studies.27 ,28 Whether this has come at the expense of diminished cognitive capacity is unknown and is perhaps of questionable relevance. It is also apparent that many relatives (but certainly not all) were grateful for the intervention and would agree to the intervention again despite the eventual outcome. However, it would be injudicious and perhaps misguided to interpret this as a variation of the consenting process and therefore a validation of the surgical intervention no matter what the eventual outcome without consideration being given to a number of limitations of this study.
Study limitations
In the first instance, when answering the question “Are you glad to be alive?” one is inclined to say, “As distinct from what?” if one is not in fact suicidal and this is perhaps the most obvious limitation of this type of assessment.
Secondly, there is the issue of selection bias. While 11 out of 13 patients interviewed (85%) felt that they would provide retrospective consent, overall there were 39 eligible patients. Twenty patients agreed to participate so it could be argued that nine patients (out of 20 recruited to the study) were unable to answer questions because of the severity level of disability and were therefore none responders. This brings down the percentage of ‘positive responders’ for 85% to 55% (11 out of a possible 20). If the seven patients who have subsequently died are included as non-responders, the percentage of positive responders falls to 41% (11 out of 27 patients) and if the whole eligible cohort is included the percentage of positive responders falls to 28% (11 out of 39 patients).
It will not be possible to determine the precise number of patients who would provide a positive response; however, there will always be a tendency to selection bias. For a patient to provide any sort of response, they must have retained sufficient cognitive function such that they can process the information provided, form an opinion and provide some sort of response.
In view of these limitations and given that we are asking the patient to consider the stark option of not being alive at all, it is hard to know what significance we should put on the opinions expressed. The ethical question is whether asking for consent retrospectively is a valid reflection of the patient's ‘critical interests’, a term coined by Dworkin29 to refer to those interests and care choices that most closely reflect our values and our personal autonomy as unimpaired human competent individuals. This viewpoint would certainly challenge the validity of retrospective consent and notwithstanding the human proclivity to adapt to disability would not necessarily amount to vindication for surgical intervention if the most likely outcome is severe disability.
Informed consent prior to medical intervention forms one of the fundamental tenets of modern medicine,30–32 and this dictates that people must have the right to decide what outcome they feel to be acceptable. A contemporary illustration of this issue is provided by Dax Cowart, himself a survivor of terrible burns sustained in a gas explosion.33 He quotes a passage from Mark Twain's A Connecticut Yankee in King Arthur's Court, in which a mother of a plague victim addresses the would-be saviours of her daughter.It might bring her back to life. None that be so good and kind as ye are would do her that cruel hurt. Thou go on thy way and be merciful friends that would not hinder.
He has become a renowned speaker in the North American health sector, arguing forcefully for the right of patients to refuse treatment in the face of a subjectively unacceptable prognosis, a stance that finds support among some of the most outspoken of our respondents.
Conclusions
This study has demonstrated some of the difficulties encountered when considering outcome following potentially life-saving but not necessarily restorative surgery for severe traumatic injury. It has highlighted the need for better guidance and support to be given for those involved in the often fraught and emotionally charged setting of acute neurotrauma especially when considering issues such as consent. It has also drawn attention to the element of uncertainty an earlier acknowledgement that survival with severe disability is a possibility so that discussions can focus on realistic outcome expectations. Finally and perhaps most importantly was the finding that most patients in this study appeared to have adapted to a level of disability that many competent individuals would feel to be unacceptable.
This does not necessarily mean that such outcomes should be regarded as ‘favourable’, nor that decompressive craniectomy must be performed for patients with predicted poor outcome. Nevertheless, categorising survival with severe disability as unfavourable outcome should probably remain as such and attaching determinative weight to ‘retrospective consent’ should not validate an intervention that is most likely to leave a person in a state that they might have previously assessed as unacceptable when competent to make that judgement, merely because they may subsequently adapt to that condition. However, having acknowledged the uncertainty and acted in good faith, those burdened with the initial clinical decisions and thereafter the long-term care may draw some support from the knowledge that unfavourable may not necessarily be unacceptable.
References
Footnotes
Contributors All authors contributed equally to this manuscript.
Competing interests None.
Ethics approval Sir Charles Gairdner Hospital SCGH HREC No: 2011-33.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Clinical ethics
Other content recommended for you
- Ethical considerations for performing decompressive craniectomy as a life-saving intervention for severe traumatic brain injury
- Neurotrauma and the rule of rescue
- Decompressive craniectomy for severe traumatic brain injury reduces mortality but increases survival with severe disability
- Acute subdural haematoma in the elderly: to operate or not to operate? A systematic review and meta-analysis of outcomes following surgery
- Neurotrauma and the RUB: where tragedy meets ethics and science
- IMPACT and CRASH prognostic models for traumatic brain injury: external validation in a South-American cohort
- Which treatments are safe and effective to reduce intracranial pressure following severe traumatic brain injury?
- Adverse events associated with microsurgial treatment for ruptured intracerebral aneurysms: a prospective nationwide study on subarachnoid haemorrhage in Sweden
- Neurological outcomes after traumatic cardiopulmonary arrest: a systematic review
- Traumatic brain injury in adults