Pergamon

PII: S0277-9536(96)00430-3

Sot'. Sci. Med. Vol. 45, No. 6, pp. 887 892, 1997 (" 1997 Elsevier Science Ltd. All rights reserved

Printed in Great Britain 0277-9536/97 $17.00 + 0.00

THE ETHICS OF EUTHANASIA--ATTITUDES AND PRACTICE AMONG NORWEGIAN PHYSICIANS

R E I D U N FORDE,"-' O L A F G J E R L O W A A S L A N D ~~ and E R I K F A L K U M '

'The Research Institute, Norwegian Medical Association, N-0107 Oslo, Norway, -~Center for Medical Ethics, University of Oslo, Oslo, Norway and ~Center for Health Administration, University of Oslo,

Oslo, Norway

A ~ t r a c ~ T h e ethical guidelines of the Norwegian Medical Association strongly condemn physician participation in euthanasia and assisted suicide. A previous study on attitudes towards euthanasia in the Norwegian population, however, indicates that a substantial part of the population is quite liberal. This study explores Norwegian physicians" attitudes towards and experience with end of life dilemmas. Sixty-six percent of a representative sample of 1476 who received postal questionnaires responded. They confirmed that Norwegian physicians actually seem to hold quite restrictive attitudes towards euthanasia. Seventeen percent answered yes to a question of whether a physician should have the opportunity to actively end the life of a terminal patient in great pain who requests this help, while 4% agreed that the same could be done to a chronically ill patient with great pain and a poor quality of life who otherwise would have several more years to live. Six percent of the physicians had performed actions intended to hasten a patient's death, while 76% said that they at least once had treated patients even if they had felt that treatment should have been discontinued. A multiple logistic regression analy- sis showed that internal medicine specialists, surgeons and psychiatrists were significantly more restric- tive than their colleagues in laboratory specialties, and that physicians educated abroad and those with negative attitudes towards patient autonomy had more liberal attitudes towards euthanasia, when gen- der and time since graduation from medical school were controlled for. ~') 1997 Elsevier Science Ltd

Key words--euthanasia, physicians, patient autonomy

INTRODUCTION

Ethical problems related to end of life decisions have received substantial attention in medical jour- nals as well as in public debate during recent years. Euthanasia is usually defined as the physicians's deliberate administration of a lethal dose, and should be separated from actions such as increasing doses of morphine which may hasten death, but where the primary aim is to alleviate pain and not to end the life. Further, euthanasia should be separ- ated from withholding or withdrawing life-sustain- ing treatment that only serves to prolong life without reversing the underlying condition.

More liberal attitudes seem to be developing towards euthanasia in a number of countries in the Western world (Kinsella and Verhoef, 1993;

Anderson and Caddell, 1993; Cohen et al., 1994, Ward and Tate, 1994; Folker et al., 1996). Contrary to this tendency, the Assembly of the Norwegian

Medical Association recently unanimously passed a new paragraph in the association's Ethical Guidelines condemning all involvement by phys- icians in euthanasia and assisted suicide. However, ethical guidelines and physicians' attitudes and practice in this field are not necessarily consistent. Little has so far been uncovered about Norwegian physicians' actual attitudes towards and experiences

with end of life decisions. This paper is an attempt to meet these shortcomings.

The patient's right to self-determination has been a most central argument in favour of physician as- sistance in life termination (Callahan, 1992; Lowy et al., 1993; Jochemsen, 1994). Accordingly, positive attitudes towards patient autonomy should be re- lated to liberal attitudes to euthanasia. So far, this issue has rarely been included in euthanasia studies. Further, the ideals of beneficence and compassion with the suffering have been used as arguments in favour of euthanasia (Lowy et al., 1993). Since the physician's compassion may increase with his occu- pational experience with suffering, his specialty or type of work is likely to influence his attitude towards ethical end of life dilemmas. Psychiatrists have previously been found to be significantly more in favour of euthanasia than other specialists, and physicians in specialties most likely to be con- fronted with the issue of euthanasia have been found to be significantly more opposed to it than physicians in less exposed specialties (Cohen et al.,

1994). Among health professionals, years of practice and previous experience of withholding or with- drawing treatment tend to be related to a more restrictive attitude towards euthanasia (Anderson and Caddell, 1993). Compassion for the patient may also be influenced by the physician's experience

887

888 Reidun Forde et al.

with his own or a close relative's suffering. We have not been able to trace studies investigating the poss- ible impact of the physician's own disease experi- ence on attitudes towards end of life dilemmas.

A relationship between religious activity and a restrictive attitude towards euthanasia is documen- ted both for physicians (Kinsella and Verhoef, 1993) and for other health professionals (Anderson and Caddell, 1993). It has also been shown that the country of graduation from medical school makes a difference (Kinsella and Verhoef, 1993). The cul- tural impact of medical school and the first years as a professional are likely to have an important bear- ing on the physician's subsequent attitudes and practice in ethical issues like euthanasia.

In this paper we will present Norwegian phys- icians' attitudes to and experience with end of life dilemmas, and how certain professional and per- sonal characteristics may have impact on how these difficult ethical dilemmas are perceived and solved.

MATERIAL AND METHODS

In 1993, four questions on end of life dilemmas were asked to a representative sample of 1476 mem- bers of the Norwegian Medical Association through a postal survey. The questions were part of a com- prehensive study on health, well-being, work satis- faction, professional autonomy and lifestyle among Norwegian medical doctors, described in more detail elsewhere (Aasland et al., 1997).

Two questions were on attitudes towards eutha- nasia:

1. A patient has a painful incurable disease and is close to death. The patient asks the doctor to help him/her to die. In your opinion, should the doctor be permitted to bring the patient's life to an end painlessly?

2. A patient has an incurable disease causing chronic pain and extreme discomfort which severely limits his/her quality of life. The patient has a duly reflected wish to die and requests the doctor's help to carry out this wish. In your opinion, should the doctor be permitted to bring the patient's life to an end painlessly? Response alternatives: Yes, No, Don't know

2. These two questions were previously used in a general population survey in Norway (Vigeland, 1991). Two related questions on experience were:

3. Have you as a physician performed actions intended to hasten a patient's death?

4. Have you actively treated a patient with a term- inal disease when, according to your own opinion, it was preferable to discontinue the treatment?

Response alternatives: Very often, Quite often, Sometimes, Seldom and Never.

Other independent variables

The physicians' attitude towards patient auton- omy was measured with a "paternalism index", based on the sumscore of the following five statements:

1. The doctor is the expert and should make the choices in most clinical situations.

2. The patient should primarily receive the infor- mation that increases the acceptance of the doc- tor's judgement.

3. It is usually an extra strain on the patient to have to take part in decisions on treatment.

4. Even if the patient may disagree with the doctor in a certain situation, she will usually afterwards appreciate that the doctor did not diverge from his clinical judgement.

5. Due to insufficient knowledge, the patient's autonomy in certain situations will be a threat to his health.

Response alternatives were: Agree completely (1), Agree to a certain extent

(2), Uncertain (3), Disagree to a certain extent (4), Disagree completely (5).

The internal consistency of the five item paternal- ism scale was fairly good (Cronbach's alpha 0.69). In the analyses the scores were inverted and added, giving a scale from 5, "Not paternalistic" to 25, "Very paternalistic". This scale was dichotomized with a score of 15 as the upper limit of the low paternalism category.

Other potential predictors were thought to be gender, age, family situation, place of living, work setting, specialty, general health and well-being, own experience of serious disease like cancer, cardi- ovascular disease or depression, religious activity, attitude towards patient autonomy, country of graduation from medical school, and actual experi- ence with end of life decisions (questions 3 and 4).

The main focus of this study is physicians' atti- tudes towards euthanasia, with special emphasis on negative attitudes. Consequently, we grouped the respondents in those who answered no and those who answered yes or don ' t know to the question on physicians' attitudes towards euthanasia in the case of a painful and lethal condition (question 1). Logistic multiple regression modelling was used to investigate the effect of a number of potential pre- dictor variables, including questions 3 and 4, on whether the physicians were clearly against euthana- sia or not.

Analyses were performed with the BMDP statisti- cal package. All associations were measured as odds ratios (OR), which approximate the relative risk of being against euthanasia. Goodness of fit of the model was measured by the Hosmer-Lemeshow X 2 test (Hosmer and Lemeshow, 1989), where the observed and predicted frequencies for up to ten cells are compared.

Ethics o f euthanas ia 889

Table I. Representativity of study sample (%)

Study sample All physicians P-value (n = 980) (n = 12479)

Gender female 28.9 25.5 male 71.1 74.5 Age group under 40 43.2 39.5 40-59 50.7 54.1 60 and over 6. I 6.5

Work situation hospital 54.9 53.1 general practice 26.9 28.8 private specialist 3.9 4.9 practice research 4.8 4.9 administration 4.8 4.4 other 4.8 3.8

0.039

0.072

0.332

RESULTS

Nine-hundred and eighty usable questionnaires were returned, giving a response rate of 66.4%. An additional 18 cases had missing data on some of the variables, and were not included in the multivariate analyses. The representativity of the sample with regard to gender, age and workplace is given in Table 1, where the respondents are compared to all members of the Norwegian Medical Association in 1993.

Apart from female physicians being slightly over- represented, no other significant differences were found.

Sixty-five percent of the Norwegian physicians were negative to a permission to end the life of a terminal patient in great pain, while 17% answered yes to the question and 18% answered don't know. 4% answered yes and 12% answered don't know to the same question in the case of a chronic and pain- ful, but not lethal condition. Six percent of the

respondents had at least once performed an action intended to hasten a patient's death. Of these, 1.5% stated that they had performed such an action quite often or sometimes, while 4.5% claimed to have rarely done such an act. Seventy-six percent of the physicians had treated a terminal patient despite their own conviction that treatment should be dis- continued at least once, and 30.6% answered that they had done this now or then or more frequently.

Group differences

Table 2 shows group differences related to gen- der, years since graduation, country of education and specialty concerning attitudes to euthanasia. There are no clear gender or age differences. Physicians educated abroad are more likely to answer yes to both attitude questions (P = 0.001 for the terminal patient question). Psychiatrists and physicians working in the internal medicine special- ties have the most restrictive attitudes.

Concerning experience with prolongation of futile treatment, 32% of the women and 21% of the men said that they had never treated a patient after they felt that the treatment should have been stopped ( P = O.OO4).

Multivariate analyses

Preliminary logistic regression analyses indicated that gender, family situation, place of living, reli- gious activity, present work setting, subjective well- being, experience of serious disease or having reluctantly extended a patient's life (question 4) were not clearly related to attitude towards eutha- nasia. With the exception of gender, these vari- ables were excluded from the model. Likewise, since the intercorrelation between the experience of

Table 2. Group differences regarding physicians' attitudes towards euthanasia in two patients in great pain who both request the physi- cian's help to die (% with numbers in parentheses)

Terminal patient Non-terminal patient No Yes Don't know No Yes Don't know

Gender female (282) 66.7 male (697) 64.3

Years since graduation 0 5 (145) 60.0 6-10 (183) 60.7 11 15 (167) 67.5 16-20 (137) 69.3 21 25 (138) 69.6 over 25 (178) 65.2

Country of graduation Norway (707) 67.9 abroad (214) 58.1 Specialty psychiatry (58) 77.6 internal med (154) 72.7 public health (37) 70.3 fam. medicine (141) 66.2 surgeons (140) 64.3 laboratory med (54) 50.0 all physicians (980) 65.0

15.2 18.1 86.2 3.2 10.6 17.6 18.1 82.3 4.6 13.1

19.3 20.7 83.4 7.6 9.0 16.9 22.4 80.9 3.8 15.3 16.6 16.0 84.4 2.4 13.2 10.2 20.4 92.0 0.7 7.3 18.1 12.3 84.8 2.9 12.3 19.1 15.7 78.1 7.3 14.6

14.0 18.t 85.6 3.7 10.7 24.7 17.2 79.4 6.1 14.5

8.6 13.8 86.2 5.2 8.6 13.0 14,3 90.3 0.6 9.1 10.8 18.9 91.9 2.7 5.4 17.6 16.2 80.1 2.8 17.0 20.0 15.7 83.6 5.6 10.7 22.2 27.8 74.1 5.6 20.4 16.9 18.1 83.4 4.2 12.4

890 Reidun Forde et al.

Table 3. Estimated odds ratios (OR) with 95% confidence intervals (CI) for answering no to whether a dying patient should be helped to a pain free death (962 Norwegian physicians)

Variable Level (n) OR 95% CI P-value

Gender male (663) 1 female (269) 1.34 0.96 1.85 .08

Years after graduation 0-5 (144) 1 6-10 (179) 1.14 0.72 1.82 11-15 (169) 1.55 0.89 2.72 16-20 (135) 1.75 0.92-3.34 21 25 (134) 2.21 1.15-4.25 over 25 (171) 1.55 0.83 2.92 .27

Specialty laboratory medicine (52) 1 family medicine/GP (136) 1.84 0.93-3.65 internal medicine

2.75 1.40 5.39 specialties (149) surgical specialties ( 131 ) 2. l 3 1.08~..24 psychiatry (55) 3.08 1.29-7.32 public health,/occupational 1.89 0.75-4.72 med. (36) non specialists (373) 1.94 0.97 3.86 .11

Country of education abroad (209) 1 Norway (723) 1.53 1.09 2.15 .02

Paternalism high (427) 1 low (505) 1.5 1.12-2.00 .004

"'Intentionally hastened yes (59) 1 death"

no (873) 3.04 1.73-5.34 .0001 Reluctantly extended yes (712) 1 patient's life

no (220) 0.81 0.58-1.13 .22

Goodness-of-fit ;(2 (Hosmer-Lemeshow) 8.69, df8 , P = 0.369.

having prolonged a patient's life against one's own conviction and the experience of having performed euthanasia was relatively high ('f = 0.71 for the dichotomized variables), we decided to control for this variable. The interaction between these vari- ables, however, was not a significant predictor. Thus, we have the following potential predictors and levels: Gender, years since graduation from medical school (0-5, 6-10, 11 15, 16-20, 21-25, over 25), specialty (laboratory disciplines including radiology and pathology, family medicine/general practice, internal medicine with sub-specialties including pediatrics and oncology, surgery with sub-specialties including obstetrics and gynaecol- ogy, psychiatry, public health/occupational medi- cine, and non-specialists, of which the majority are younger physicians in the process of specializ- ation), country of education (Norway or abroad, the great majority of those educated abroad are also native Norwegians), attitudes towards patient autonomy (high or low paternalism) and finally, experience with end of life decisions (dichotomized responses to questions 3 and 4; no = 1, seldom/ sometimes/quite often/very often = 0).

Table 3 shows the result of the multiple logistic regression analysis. Physicians in internal medicine, surgeons, and psychiatrists were significantly more restrictive than laboratory specialists, who constitute the reference group. Being educated in Norway and scoring low on paternalistic attitudes were also sig- nificant predictors of being against euthanasia, when other variables were controlled for. The physicians who graduated 21-25 years before the survey (i.e. in 1968-1972) had a significantly higher probability of

being against euthanasia than their younger and older colleagues. Not having performed euthanasia proved to have the strongest effect in this adjusted model, with an odds ratio of 3.04 (95%, CI 1.73- 5.34). Having treated a patient after they felt the treatment should have been stopped, however, did not add predictive power to the model. The model fits the data quite well.

DISCUSSION

This is the first systematic study of Norwegian physicians' attitudes to and experience with end of life dilemmas. It indicates that Norwegian phys- icians have restrictive attitudes towards euthanasia and very limited practical experience with it. Problems related to ending active treatment to a terminally ill patient are quite prevalent.

Study limitations

A postal questionnaire with simplifying questions and preformed response categories cannot elucidate all the aspects of these difficult ethical dilemmas. Although the questions have limitations we decided to use the same wording as in the 1988 Norwegian population survey (Vigeland, 1991) in order to be able to compare physician attitudes with those of lay groups. Obviously, the questions could have been formulated more unambiguously and nuanced, especially with regard to comparisons with other physician studies. However, the study was primarily seen as an attempt to elucidate a difficult issue which so far has suffered from lack of empirical data.

Ethics of euthanasia 89 |

Considered in isolation, question i ("In your opinion, should the doctor be permitted to bring the patient's life to an end painlessly?") might have been interpreted as whether the treatment of a terminal patient should aim at the least possible pain. In that case, there is hardly a moral dilemma attached to it. However, this interpretation seems unreasonable when the preceding sentences are taken into account. Further, the positively loaded and possibly leading expressions "duly reflected wish" and "painlessly" may have produced a bias in the yes-direction.

The wording of the question "Have you as a physician performed actions intended to hasten a patient's death?" has serious weaknesses. It is likely that actions consisting of withdrawing life sustain- ing treatment in hopelessly ill patients (and where death is the immediate result) have been experi- enced by the participating physicians as actions resulting in the death of a patient (Downie and Calnan, 1994).

On the other hand, since euthanasia is illegal, one should expect under-reporting, even though the study was anonymous. For the same reason, a ten- dency to suppress that life actually actively and intentionally has been shortened may also have led to under-reporting. However, it seems reasonable to assume that colleagues in favour of legalization of euthanasia may have used this survey as an oppor- tunity to show that this practice actually does already occur in Norway.

Attitudes

Norwegian physicians seem to be more conserva- tive than their colleagues in other Western countries from which we have empirical data. Norwegian physicians also seem to be more conservative than the Norwegian population as a whole. In the popu- lation survey from 1988 (Vigeland, 1991), 36% thought that a physician should be permitted to actively hasten death in a terminal patient, and 17% in a chronic, non-lethal disease. This study also reported data for various groups of university students, and, interestingly, the medical students were the most restrictive.

The amount of time the patient has left to live influences the attitudes towards euthanasia of the Norwegian physicians as well as the population (Vigeland, 1991). Dutch physicians have used short time left to live as an argument when euthanasia done without explicit request from the patient was presented (Pijnenborg et al., 1993). One vital ques- tion related to this is how "a short time left to live" is to be understood. A study from the U.S. shows that among physicians who were in favour of legalization of euthanasia, 64% thought that, in order to justify euthanasia, the patient should have a life expectancy of less than 6 months (Cohen et al., 1994). Twenty-one percent of these physicians also agreed that euthanasia or assisted suicide may

be appropriate if external factors such as not wanting to be a burden on the family or not wanting to deplete the savings, led to the patient's request despite adequate pain control and quality of life.

Contrary to findings in the same U.S. study (Cohen et al., 1994), Norwegian psychiatrists are more restrictive towards euthanasia than their col- leagues. Our study confirms previous findings that colleagues in the medical specialties, among them oncologists and haematologists, who in their daily work have close and continuous contact with dying patients, hold more restrictive attitudes than phys- icians in laboratory specialties.

Surprisingly, we found that physicians with the most paternalistic attitudes towards patient partici- pation in medical decisions also held the most lib- eral attitudes towards euthanasia. We have no explanation for this, however, it may be speculated that it is easier for physicians with paternalistic atti- tudes to make exceptions from ethical and pro- fessional guidelines. This thought-provoking result calls for more comprehensive predictive models in future research.

As mentioned previously, religious activity has been found to predict restrictive attitudes towards euthanasia (KinseUa and Verhoef, 1993). We have not been able to find such a relationship. A British study, however, found that the attitudes held by some doctors with religious beliefs seemed to be at variance with their behaviour (Ward and Tate, 1994).

The model reveals an interesting pattern with regard to years of practice, in that those who graduated around 1970 are more restrictive. We have no logical explanation for this, but a specu- lation could be that they represent the first classes of medical students that experienced the growing public awareness of physician fallibility (Rothman, 1991). They also represent the "'hippie" or "Vietnamese War" generation with idealistic atti- tudes and strong political awareness.

Experiences

Only a small minority of Norwegian physicians theoretically accept euthanasia, and even fewer report that they have ever performed an act intended to advance a patient's death. The fact that the act is illegal is the most obvious explanation for this. No Norwegian doctor has ever been taken to court for euthanasia.

Although the legal restrictions are the same in Norway and Great Britain, Norwegian physicians seem to practice euthanasia to a lesser extent than their British colleagues (Ward and Tate, 1994), a discrepancy that also bears witness to the complex- ity of attitude formation in this field. Denmark and Norway have cultures that may be compared and the same legal restrictions on euthanasia. A recent study from Denmark shows that 34% of Danish

892 Reidun F~rde et al.

physicians find the administration of a lethal dose on the request of the patient ethically acceptable, and 5% of these have administered a lethal injec- tion at the patient's request at least once (Folker et al., 1996). These 5% may be compared to the 6% in our study. However, unlike the Danish study the question used to explore Norwegian physicians's ex- perience with euthanasia (" . . .performed actions intended to hasten a patient 's death?") are not pre- cise. Accordingly, it is most likely that among the 6% yes answers to this question are also hidden actions of treatment withdrawal in hopelessly ill patients indicating that Norwegian physicians may have even more limited experience with euthanasia than their Danish colleagues.

A study from the state of Washington showed that although a small majority of physicians were in favour of legalizing euthanasia, only 33% were actually willing to perform the act (Cohen et al., 1994). This pattern is also found elsewhere (Kinsella and Verhoef, 1993; Ward and Tate, 1994).

A less dramatic ethical problem, but according to our findings a quantitatively larger problem, is the question of when treatment of a hopelessly ill patient should be terminated. Three out of four Norwegian physicians have given treatment which they did not feel justified in such a situation, indi- cating that stopping futile treatment is perceived as ethically difficult. We regard this experience mainly as caused by a system defect, through lack of guide- lines or a defensive attitude. Having done "every- thing possible" is always safer than making specific choices of termination, especially when there is also an element of peer pressure.

We found no relation between physicians' attitudes to euthanasia and their experiences of having "over-treated" patients. This is somewhat surprising. One might assume that doctors in favour of euthanasia would be able and willing to terminate futile treatment more easily. When this is not confirmed, it may be explained by the assumption that those who defend euthana- sia tend to experience "over-treatment" more strongly than the physicians who are restrictive to a change of treatment which more quickly will lead to the patient 's death. The last cat- egory will probably more rarely define a treat- ment as futile. The fact that female physicians both tend to have more restrictive attitudes towards euthanasia and more rarely experience

that they treat patients after the treatment should have been stopped, indicates the likeli- hood of such an explanation,

This study has confirmed that national differences influence the attitudes to and experience with cross cultural ethical dilemmas such as euthanasia. The results of our study may shed light on the complex- ity of the attitude formation of such a complicated and difficult ethical question.

REFERENCES

Aasland, O. G., Olff, M., Falkum, E., Schweder, T. and Ursin, H. (1997) Health complaints and job stress in Norwegian physicians. The use of an overlapping ques- tionnaire design. Social Science & Medicine. In press.

Anderson, J. G. and Caddell, D. P. (1993) Attitudes of medical professionals toward euthanasia, Social Science & Medicine 37, 105-114.

Callahan, D. (1992) When self-determination runs amok. Hastings Cent Report, March-April, 52-55.

Cohen, J. S., Fihn, S. D., Boyko, E. J., Jonsen, A. R. and Wood, R. W. (1994) Attitudes toward assisted suicide and euthanasia among physicians in Washington State. New England Journal of Medicine 331, 89-94.

Downie, R. S. and Calnan, K. C. (1994) Healthy Respect. Ethics in Health Care. Oxford University Press, Oxford.

Folker, A. P., Holtug, N., Jensen, A. B., Kappel, K., Nielsen, J. K. and Norup, M. (1996) Experiences and attitudes towards end-of-life decisions amongst Danish physicians. Bioethics 10, 233-251.

Hosmer, D. W. and Lemeshow, S. (1989) Applied Logistic Regression. Wiley, New York.

Jochemsen, H. (1994) Euthanasia in Holland: an ethical critique of the new law. Journal of Medical Ethics 20, 212-217.

Kinsella, T. D. and Verhoef, M. J. (1993) Alberta Euthanasia survey: 1. Physicians' opinions about the morality and legalization of active euthanasia. Canadian Medical Association Journal 148, 1921-1926.

Lowy, F., Sawyer, D. M. and Williams, J. R. (1993) Canadian physicians and euthanasia: 4. Lessons from experience. Canadian Medical Association Journal 148, 1895-1899.

Pijnenborg, L., van den Maas, P. J., van Delden, J. J. M. and Looman, C. W. N. (1993) Life-terminating acts without explicit request of patient. Lancet 341, 1196-1199.

Rothman, D. J. (1991) Strangers at the Bedside. A History of How Law and Bioethics Transformed Medical Decision Making. Basic Books, New York.

Vigeland, K. (1991) Holdning til aktiv volunteer eutanasi [Attitude towards active voluntary euthanasia]. Journal of the Norwegian Medical Association 111,460-463.

Ward, B. J. and Tate, P. A. (1994) Attitudes among NHS doctors to requests for euthanasia. British Medical Journal 308, 1332-1334.