Over the past few decades, there has been increasing attention paid to medical decisions at the end of life. Medical practitioners frequently have to make decisions that affect the timing of a patient’s death,1-10 and the frequencies with which medical decisions at the end of life occur differ between countries.11,12 Determinants of attitudes and practices relating to such decisions include unbearable pain and suffering,13 incompetence,14 and patient requests,15 as well as characteristics of the treating medical practitioner, such as sex, age, religion and specialty.1,5,16-19
There is ongoing debate about the involvement of medical practitioners in a range of medical end-of-life decisions, such as withholding or withdrawing life-sustaining treatment,6,20 use of drugs for possibly life-shortening alleviation of pain and symptoms,21 or active euthanasia and physician-assisted suicide.22-24 Given the crucial role that medical practitioners play in this debate, ongoing assessment of their attitudes and practices related to end-of-life decision making is important.
An international study involving researchers in six European countries and Australia was therefore conducted. Results of the study, comparing attitudes and practices across the seven countries have been reported. Findings were that the main predictors of end-of-life decision making were a request from a competent patient, short life expectancy and uncontrolled pain;25 that there was considerable variation between countries in requirements of institutional ethics committees;26 and that most of the variation in responses was accounted for by “country”.27
Ethical approval was obtained from the Behavioural and Social Science Ethical Review Committee of the University of Queensland.
After ensuring strict confidentiality and privacy, a maximum of 300 medical practitioners were drawn from the membership registers of Australian and Australasian professional colleges of specialties frequently involved in the care of dying patients.
The structured questionnaire used in all seven countries was piloted in Australia. Respondents were asked to provide demographic information, nominate the main clinical specialty in which they were working, and to answer a series of questions based on four hypothetical cases. The cases, three specified courses of action and the conditions under which respondents would take the specified actions are shown in Box 1. Respondents were also asked if they would prescribe drugs to enable the patient to end his or her life if that was what the patient requested. The nature of this question precluded asking about conditions such as requests from relatives, or practitioners taking the action on their own initiative.
Questionnaires were coded and data were entered into EPI Info, version 6 (Centers for Disease Control and Prevention, Atlanta, Ga, USA) and transferred to SPSS, version 10 (SPSS Inc, Chicago, Ill, USA) for analysis. Double entry was undertaken to ensure accuracy. Weighting factors were used to correct for the different sampling fractions of each specialty and response percentages in the different strata.
Questionnaires were sent to 2964 medical practitioners. After deleting those who were “no longer at this address”, had retired, were overseas or deceased, we had a possible sample of 2896. Completed questionnaires were received from 1540 respondents, giving a response rate of 53%. The 62 respondents who were no longer working as medical practitioners were removed, giving a valid sample of 1478 respondents (Box 2).
Respondents were predominantly male (78%), and more than 70% were aged 40 or more years; 32% had no religious affiliation or spiritual philosophy; 45% claimed affiliation with the three mainstream Australian Christian churches (Anglican, Catholic and Uniting Church), while 15% belonged to a range of other religions and 8% held a non-religious philosophy. Fifty-four per cent of respondents stated that their belief/philosophy was very important or important in their professional attitude towards end-of-life decision making, while 46% said their belief/philosophy was not very or not at all important.
Respondents’ self-reported clinical specialties are shown in Box 2; some respondents nominated a primary specialty which was not in accord with their college membership.
Response rates ranged from 42% for general practice to 62% for oncology; for palliative medicine, there was a sample of 28, but 29 respondents self-identified with this specialty (Box 2).
Differences between specialty groups did not reach significance for any of the specified actions shown in Box 1 when the patient was competent (Cases 1 and 2), or for increasing drug therapy when the patient was no longer competent (Cases 3 and 4). However, differences reached or approached significance for providing terminal sedation and for giving drugs to hasten the end of the patient’s life in Cases 3 and 4. The range of responses among the specialty groups are shown in Box 3.
Increasing drug therapy: As shown in Box 4, most respondents in every specialist group would accede to a current request from a competent patient (Cases 1 and 2) to provide relief from pain and symptoms by increasing drug therapy, even if this might hasten the end of the patient’s life. In both Cases 1 and 2, palliative care specialists were least likely to report that they would do so, with oncologists also least likely in Case 2. There was greater support from respondents for taking this action on their own initiative in Case 1 than in Case 2.
Providing terminal sedation: There was less support for providing terminal sedation, for all four Cases and under all three conditions (Box 5), than for increasing drug therapy which might hasten death. For Case 1 (competent patient in pain, with life expectancy of 2 weeks), geriatricians were least likely to provide terminal sedation, either at the request of the patient or on their own initiative, obstetricians/gynaecologists were the most likely to do so when the patient requested it, and palliative care specialists were most likely to do so on their own initiative; all respondents were more willing to provide it on their own initiative than at the request of relatives. For Case 2 (competent patient with pain controlled, but with symptoms of tiredness, shortness of breath, and bedridden with a life expectancy of 3 months), oncologists were least likely, and obstetricians/gynaecologists most likely, to provide terminal sedation at the patient’s request, and geriatricians were least likely and palliative care specialists most likely to do so on their own initiative, again with all respondents more willing to provide terminal sedation on their own initiative than at the request of relatives.
Giving drugs to end a patient’s life: As expected, giving drugs explicitly to hasten the end of the patient’s life received lower levels of support in all four Cases and under all conditions than either of the other specific actions (Box 6). For competent patients (Cases 1 and 2), differences between the specialty groups reached significance under the condition of a direct request from the patient in both Cases, with palliative care specialists and oncologists least likely to do so, and anaesthetists most likely to do so. Differences reached significance under the condition of respondents taking this action on their own initiative only for Case 1 (ranging from 3.4% for general practitioners and palliative care specialists to 13.6% for surgeons and 13.9% for anaesthetists). In both Cases 1 and 2, almost all respondents were more likely to undertake the action on their own initiative than at the request of relatives.
Physician-assisted suicide: The proportion of respondents who reported that they would prescribe drugs to enable the patient to end his or her life if that was what the patient requested was 13% in Case 1 and 10% in Case 2. Significant differences by specialty for this question were revealed by χ2 analysis (Box 7).
Overall, our results reflect strong medical support for patient self-determination, but also the acceptance of decision-making responsibility by relatives once competence is lost. Intensifying drug therapy for pain, which could hasten death, was strongly supported for competent patients, but there was far less support for providing terminal sedation, even if requested by a competent patient. This may reflect a feeling that providing terminal sedation is a more direct method of hastening death than increasing pain relief, and is thus less acceptable within medical culture. However, we also note that 26.1% of medical practitioners would be willing to provide terminal sedation on their own initiative to competent patients with poorly controlled pain and 2 weeks to live. For these respondents, the extreme conditions of Case 1 appear to extinguish the need for a request from the patient before taking such action. For patients who have lost their competence, the proportion of medical practitioners willing to institute terminal sedation rises (although not to the levels of willingness to intensify pain treatment), and this is the same if requested by relatives or on the doctor’s own initiative.
Like those of other studies,2 our results indicate that low but significant proportions of medical practitioners would give drugs intentionally to hasten death in different situations, despite the fact that this is unlawful in Australia. Overall, more respondents would, on their own initiative, intentionally hasten the death of incompetent patients than competent patients, but the fact that any medical practitioners would take such unilateral action in relation to competent patients is cause for concern. While we believe that the case descriptions, actions and conditions presented to respondents were unambiguous, it is possible that some of the respondents interpreted “on your own initiative” to mean that they would be willing to broach the subject of giving drugs to hasten death, and to then respond positively to a competent patient’s request — still unlawful, but ethically far less problematic.
Palliative care specialists, oncologists and geriatricians were least likely to respond positively to a request from a competent patient for drugs to hasten death, or to give drugs intended to hasten death in cases of incompetent patients who requested this through an advance directive, irrespective of life expectancy. These low figures were also replicated for these specialties in relation to practitioners being willing to hasten a patient’s death on their own initiative. This reflects expected low levels of support in these groups for active euthanasia, either voluntary or non-voluntary,28 but the differences within each of these groups in relation to practitioners’ willingness to initiate different processes which lead to death is important. While actions traditionally perceived as direct hastening of death are avoided by these groups for all patients, other actions with the same result but purportedly with different intentions (eg, terminal sedation) would sometimes be taken without reference to competent patients.
Our results indicate that perceptions concerning the causation of death as well as aspects of medical culture influence attitudes towards medical decisions at the end of life. The relief of suffering is important to all specialties, and a quarter of respondents were prepared to offer terminal sedation to competent patients with poorly controlled pain and a poor prognosis. However, the specialties which are most closely involved in caring for patients nearing death — oncology, palliative care and geriatrics —were less likely to act in ways that might be perceived as actively hastening death, and more likely to act in ways which may be understood as fulfilling the duty to relieve symptoms as a medical necessity.
Since euthanasia and physician-assisted suicide are unlawful in Australia, the approach by these three specialties could be interpreted simply as being more consistent with the existing legal climate. However, as the same groups are more prone to unilateral decision making, sometimes without reference to competent patients, their avoidance of active hastening of death could also reflect a generally weaker recognition of patients’ wishes.
Our findings point to the need to include decision-making theory and practice within medical ethics curricula, and to facilitate more discussion between specialties about medical decisions at the end of life. In addition, further collaborative research on the attitudes of patients, carers and the general community to medical end-of-life decisions would provide useful comparative perspectives between the attitudes of physicians and those of the communities in which they practise.
1 Hypothetical cases, specified actions and conditions
Each case relates to a 71-year-old patient with extensive brain and bone metastases who has undergone burdensome chemotherapy twice.
Case 1: The patient is clear-headed and can still communicate well. You estimate the patient’s life expectancy to be no more than 2 weeks. The patient has pain that is difficult to control, despite the use of analgesics in high doses.
Case 2: The patient is clear-headed and can still communicate well. You estimate the patient’s life expectancy to be at least 3 months. Pain can be adequately controlled, but the patient is extremely tired, short of breath and bedridden.
Case 3: The patient is drowsy or subcomatose and communication is not possible. You estimate the patient’s life expectancy to be no more than 2 weeks. Pain can be adequately controlled, but the patient is extremely tired, short of breath and bedridden.
Case 4: The patient is drowsy or subcomatose and communication is not possible. You estimate the patient’s life expectancy to be at least 3 months. The patient has pain that is difficult to control, despite the use of analgesics in high doses.
Would you intensify the alleviation of symptoms by using drugs, taking into account the probability or certainty that this could hasten the end of the patient’s life?
Would you give drugs, such as benzodiazepines or barbiturates, to keep the patient in deep sedation until death occurs (referred to henceforth as terminal sedation)?
Would you administer drugs with the explicit intention of hastening the patient’s end of life?
Conditions under which respondents would take the specified actions
If that is what the patient requests (Cases 1 and 2) or requested in an advance directive (Cases 3 and 4).
Without first informing the patient (Cases 1 and 2) if that is what the relatives request (all Cases).
On your own initiative to reduce suffering.
Response options were on a five-point scale (yes; probably; undecided; probably not; no).
3 Range of responses among specialties for three specific actions at the request of relatives
4 Percentage (95% CI) of each specialist group answering yes/probably to the specific action of increasing drug therapy with the probability of hastening death
5 Percentage (95% CI) of each specialist group answering yes/probably to the specific action of providing terminal sedation
6 Percentage (95% CI) of each specialist group answering yes/probably to the specific action of giving drugs to hasten the end of a patient’s life
7 Percentage of each specialist group answering yes/probably to the question on physician-assisted suicide*
* Applicable only to competent patients (Cases 1 and 2). |
|||||||||||||||
Received 22 August 2007, accepted 13 December 2007
Abstract
Objective: To compare attitudes and practices of Australian medical practitioners, by specialty, to a range of medical decisions at the end of life.
Design, setting and participants: As part of an international study, in 2003, a structured questionnaire was mailed to 2964 medical practitioners drawn from membership registers of Australian and Australasian professional colleges. Data from 1478 questionnaires were statistically analysed using validated instruments.
Main outcome measures: Practitioners’ willingness to comply with requests from patients and/or their relatives for symptom relief which might also hasten death; provision of terminal sedation and euthanasia, or willingness to provide these on their own initiative.
Results: Respondents reported being much more willing to comply with a patient’s request for increasing symptom relief, even at risk of hastening death, than for terminal sedation. Over a quarter of respondents would provide terminal sedation to competent patients on their own initiative. A small number of respondents would intentionally hasten death. There were significant differences by specialty for all three actions. Oncologists, palliative care physicians and geriatricians were least likely to actively hasten death, and more likely to act unilaterally to relieve symptoms as a medical necessity.
Conclusions: Perceptions about the causation of death and aspects of medical culture appear to influence physicians’ attitudes towards medical decisions at the end of life. Our findings have implications for medical education, interprofessional communication and discussion between the medical profession and the community.