euthanasia: why people want to die earlier

Pergamon

Sm. Sci. Med. Vol. 39, No. 5, 64-654, 1994 pp.

0277-9536(93)EW66-N Copyright 0 1994 Elsevier Science Ltd

Printed in Great Britain. All rights reserved 0277-9536/94 $7.00 + 0.00

EUTHANASIA: WHY PEOPLE WANT TO DIE EARLIER

CLIVE SEALE’ and JULIA ADDINGTON-HALL*

‘Department of Sociology, Goldsmith’s College, University of London, Lewisham Way, London SE146NW, U.K. and ‘Department of Epidemiology and Public Health, University College London,

66-72 Cower Street, London WClE6EA, U.K.

Abstract-The results from two surveys in England of relatives and others who knew people in samples drawn from death certificates are reported. The main focus is on a sample of 3696 people dying in 1990 in 20 health authorities, with supporting analysis from an earlier national sample of 639 people dying in 1987.

The incidence of people saying they wanted to die sooner, and of requests for euthanasia are reported. Excluding a proportion who did not wish to express a view, or did not know the answer, about a quarter of both respondents and the people who died expressed the view that an earlier death would be, or would have been, preferable. 3.6% of people in the 1990 study were said to have asked for euthanasia at some point in the last year of life. The extent to which such views were determined by the experience of pain, other distressing symptoms, dependency and social and cultural factors such as religious belief and social class is explored. The finding that dependency was important in causing the feeling that an earlier death would have been better, as well as requests for euthanasia, is related to the public debate about euthanasia, which often contains the assertion that fear of pain is a dominant factor. Pain was found to be a significant factor in death from cancer, but not as important for other causes of death. Social class, place of residence of the deceased, and strength and type of religious faith were found to be largely insignificant in influencing feelings about an earlier death and requests for euthanasia.

Key wordsAeath, euthanasia, hospice, cancer

INTRODUCTION

In recent years there have been periodical upsurges of

public and medical concern about euthanasia.

Currently, the position of the British Medical Associ-

ation is summarised as follows:

The law should not be changed and the deliberate taking of a human life should remain a crime. This rejection of a change in the law to permit doctors to intervene to end a person’s life is not just a subordination of individual wellbe- ing to social policy. It is, instead, an affirmation of the supreme value of the individual, no matter how worthless and hopeless that individual may feel [I].

Doubts about the wisdom of this position are

sometimes brought about by reports of particularly

harrowing cases, such as that in 1992 of Dr Nigel Cox in the United Kingdom. Dr Cox was convicted of attempted murder for administering a potentially lethal injection to his patient who was suffering unbearable pain. In this action he was following both the patient’s and the family’s wishes. In view of public concern over this conviction, Richard Smith, editor of the British Medical Journal wrote that:

It is time for the British to think deeply about euthanasia there is an urgency to the need for a (royal) commission as one effect of (the Cox) case is likely to be that patients will be terrified of talking to doctors about the possibility of their deaths being hastened. And dying alone afraid of talking to your carers may be much worse than being dead [2].

Much of the literature on this topic is written as moral philosophy, rehearsing the arguments for and

against euthanasia in different hypothetical situations [3]. Reports of single instances of euthanasia are also occasionally presented [4]. This paper reports an empirical contribution to this debate, describing the circumstances in which a representative sample of adults died, using the accounts given by relatives and others after the death. The analysis reported here assesses the influences on both respondents’ views and the reported views of the dying people themselves about the desirability of an earlier death. In addition, the causes of requests for euthanasia are reported.

Focus of investigation

This paper will assess the type of suffering that is most likely to influence the desire to die earlier. An analytic distinction is made between symptom distress and dependency, as well as between different groups of symptoms. This is important in assessing the claims of opponents of euthanasia that good care is an adequate substitute, a topic that is explored more fully in a subsequent paper.

In addition this paper will test the proposition that the desire to die earlier, and requests for euthanasia, are more common in urban dwellers, the younger dying, those in higher social classes and the non religious. By implication, the alternative proposition that these factors are unlikely to be important when people are very close to death (either themselves, or through having experienced bereavement) will be

SSM W/SD 647

648 CLIVE SEALE and JULIA ADDINGTON-HALL

tested. Thus the relative importance of cultural factors as against suffering will be assessed.

METHODS

This paper reports results from a survey of the relatives and others who knew 3696 people dying in 1990 in 20 district health authorities in England. This survey repeated the method of an earlier study of 639 deaths in 1987 based on a nationally representative sample [5]. Results from this earlier study are reported here in so far as they support or differ from the results from the larger study.

Data were gathered by means of structured interviews with people in the community who best knew the circumstances of the last year of life. Response rates for the two studies were 69% (1990) and 80% (1987). The 1990 study sampled deaths from cancer disproportionately, so in the analysis of the whole sample these were weighted by a factor of 0.27 to conform to the proportion of deaths due to cancer in the districts (26%). This gives an overall weighted sample of 2192. The effect of this on significance levels is conservative. Weighting is not used where cancer deaths are considered separately.

MEASURES

Dependent aariables

In both studies respondents were asked ‘Looking back now, and taking (the deceased’s) illness into account, do you think s/he died at the best time-or would it have been better if s/he had died earlier or later? In the 1990 study respondents were also asked ‘What about (the deceased)? Did s/he ever say that they wanted to die sooner?’ and ‘(If yes) did s/he ever say that s/he wanted euthanasia?

The questions were not asked for people who died suddenly with no illness or warning or time for care, and in a few cases where the respondent could not have been expected to know the answer (e.g. where a coroner was the only person who could be found to be interviewed). The frequencies for the responses to these questions are given in Table 1.

Table I. Views about dying earlier, and requests for euthanasia (percentages, with 1990 data weighted)

Respondents’ views Deceaseds’ reported wishes

1987 1990 I990

Best time 62 57 To die earlier 24 Better earlier 24 28 Asked for euthanasia 3.6 Better later 14 I5

Valid N (= 100%) 434 1720 1907” Don’t know/ other answer I7 I3 D/K 8 Missing I5 8 Miss. 5

N(= 100%) 639 2192 2192

“26 people who said that the person had wanted to die sooner said they did not know whether the person had asked for euthanasia. giving a base of 1881 for the euthanasia figure.

The table shows that about a quarter of respondents who answered the question felt that it would have been better if the person had died earlier, and that a similar proportion of the dying people were said to have felt this. A much smaller proportion were said to have asked for euthanasia.

In addition to the exclusion of sudden deaths, a proportion of respondents either did not know the answer to the question, or preferred to couch their answer in terms other than those offered. The response bias introduced by all of these exclusions was investigated in both data sets for the question concerning respondents’ views. In both surveys responders to the questions were more likely than those who did not reply to it to be reporting for people who suffered dependency and distress (as measured by the variables described in the section that follows), who were older, who suffered from cancer and who did not suffer from heart disease. In the 1990 survey respondents describing the deaths of women and of people who did not live in inner city areas were more likely to respond to the question. In both surveys responders were more likely to be sons or daughters of the deceased and less likely to be spouses than non responders. The strength or type of religious faith of either the deceased or the respondent was not associated with replying to the question, nor was the social class of the deceased. Response bias to the questions about the deceaseds’ views in the 1990 survey followed the same pattern [6].

Independent variubles

In addition to background variables such as age, gender and so on, and individual questions which will bc described as results are reported, some composite scales of symptom distress and dependency were created. Respondents were asked to report on the presence of 15 symptoms of illness, and to rate them, if present at all in the last year of life, on a three point scale ranging from ‘very’ (scoring 3) to ‘fairly’ (scoring 2) to ‘not very distressing’ (scoring I). A factor analysis of these symptoms (excluding pain which is treated separately in the analysis) revealed four factors in both data sets, which were used to create four different composite measures of symptom distress. In addition. dependency on others for activi- ties of daily living was assessed by seven items, where respondents also reported on the duration of dependency, ranging on a six point scale from ‘less than a week’ to ‘more than a year’ before death or final admission to a hospital or hospice. These dependency items were shown to be unidimensional by factor analysis in both data sets, and were combined to give a dependency score. Components of symptom distress and dependency scores are given in Tables 2 and 3.

In the analysis that follows, these scales were converted to give a three point scale (None, medium and high). Here, a zero means that no symptoms or dependency were reported, and the cut-off points for

Euthanasia: why people want to die earlier 649

Table 2. Components of four symptom distress scores

Appetite Control Mental Breath

Sickness/Vomiting Loss of bladder Mental Trouble with control confusion breathing

Dry mouth Loss of Depression or Persistent bowel control feeling miserable cough

Loss of appetite Unpleasant Sleeplessness smell

Difficulty swallowing Bedsore Constipation Reliability (alpha)

1987 = 0.7246 0.6454 0.4783 0.4850 1990 = 0.6558 0.5881 0.5208 0.5123

Range=&15 &I2 &9 I%6 Cut-off for high score = 3 3 3 3

‘high’ levels are given in the tables, other scores falling into the ‘medium’ category.

RESULTS

Distress and dependency: the role of suffering

The argument for euthanasia arises from the perception that illness near the end of life involves intolerable levels of suffering, such that the quality of peoples’ lives deteriorates to the point where it is no longer worth living. Respondents were asked to describe the dying person’s quality of life as either ‘good, fair or poor’ during their final year. The degree to which this was associated with views about dying earlier and requests for euthanasia is shown in Table 4. The table shows a moderate association between reported quality of life and all the measures of the desire to die earlier. However, the question about quality of life occurred directly after the question about whether it would have been better to die earlier. It might reasonably be argued that this might have biased respondents’ assessments of quality of life. Much earlier in the interview, before any question of wishes to die earlier had arisen, respondents were asked to describe the symptoms and dependency which the dying people experienced. The results show, firstly, that there is variation in these in the last year of life: some people suffer more than others, and different causes of death involve different types of distress and dependency. This is shown in Table 5.

Table 3. Components of dependency score

Dependency

Get in and out of bath or shower Dress/undress (inc. shoes/fastenings) Go to toilet (cope on own while there) Wash (and shave) Cut own toenails Make a hot drmk Need help at night Reliability (Alpha)

1987 = 0.9253 1990 = 0.9296

Range = &42 Cut-off for high score = 18

The table shows that very distressing pain was judged to be more common in cancer than other causes of death, and a high level of distress on the composite appetite measure was also reported for cancer. The other symptom distress measures were only marginally higher in cancer, but overall, the table suggests that cancer is a cause of death that involves higher levels of both physical and mental distress than other causes of death. Ischaemic heart disease, on the other hand, causes comparatively low levels of symptom distress. People dying from stroke (CVA) were somewhat less likely to suffer from very distressing pain, as well as scoring low on breath (Breathlessness and coughing). As might be expected, those dying from respiratory disease scored particularly highly on breath and people dying from mental conditions, including Alzheimers, scored particularly highly on mental. People dying from mental conditions and Alzheimer’s were also the most dependent. Respiratory disease and stroke involved dependency levels that were slightly above average; ischaemic heart disease was below average, and cancer about average [7].

If symptom distress leads to a desire to die earlier, one might expect such wishes to be more likely for people dying from cancer. If dependency is a causal factor, one might also expect such wishes to be more likely for those dying from mental conditions and Alzheimers, although due to mental confusion this would be a view more likely to be expressed by the

Table 4. Reported quality of life, views about dying earlier, and requests for euthanasia (percentages with base number in paren-

theses; 1990 data weighted)

1987 1990 Respondents’ Respondents’

views views

1990 Deceased’s

views

Better earlier

%

Better earlier

%

Wanted sooner

%

Wanted euthanasia

%

Good Fair Poor All Cramer’9

v= P=

12 (154) 12.5 (637) I3 (716) 1.0(712) I4 (132) 23(519) 23 (580) 3.3 (573) 48 (124) SO(5lO) 39 (553) 6.7 (546) 23 (410) 27 (1665) 24(1848) 3.5(1831)

0.38 0.35 0.25 0.13 < 0.00005 <0.00005 < 0.00005 < 0.00005


Table 5. Cause of death by distress and dependency (percent saying It was at the highest level; 1990 data

only)

Type of Other Mental distress Cancer IHD CVA circulatory Respiratory /Alzheimers Other

Pain 53 36 28 36 35 39 36 Appetite 83 42 45 4x 61 4x 57 Control 4s I7 38 26 41 39 37.5 Mental 63 45 41 46 58 70.5 57 Breath 35 72 I4 27 53 IX 2: Dependency 38 22 42 31 49 79 47

Total (100%) 2074 564 237 207 242 64 278

respondent than the person themselves. People dying from ischaemic heart disease should be the least likely to express this view. The results shown in Table 6 test these ideas. Table 6 shows that, in death from cancer, respondents were no more likely than average to say that it would have been better if the person had died earlier, in spite of this cause of death involving high levels of suffering [S]. This sentiment was most likely to be expressed for people dying from respiratory disease and, in particular, mental conditions. As expected, ischaemic heart disease is the least likely to be associated with respondents who felt an earlier death would have been better. IHD is also significantly less likely to be associated with requests for euthanasia from those who died.

An examination of the dying person’s reported views about dying sooner shows few differences, with the only exception being those dying from respiratory disease. In contrast to respondents’ views, it is death from cancer that is most likely to have led to requests for euthanasia. Dying from mental conditions is not associated with a high incidence of such requests, and this is in contrast to the views of respondents. This contrast might be expected in a group likely to be suffering from mental confusion.

Different diseases, then, are associated with different patterns of distress and dependency, and are also associated with differences in views about dying earlier and requests for euthanasia. The

Table 6. Cause of death, by views about dying earlier and reported requests for euthanasia (percentages; base numbers are given in

brackets: 1990 data o&1

Respondents’ views

Deceased’s views

Better earlier

Wanted SOD”IX

Wanted euthansia

CCl”CCX IHD CVA Other circulatory Respiratory Mental” Other

28 (1821) 23(1980) 5.2** (1952) l4** (396) 23 (443) 1.8’(441) 33 (195) 22 (217) 3.7 (216) 26(152) 23 (175) 2.9 (173) 40** (197) 30’ (222) 2.7 (219) 48” (54) 22 (58) I .8 (57) 32 (2101 27 (232) 4.3 (230) , I ~ I ~ I

*P < 0.05; l *P < 0.01; Significance tesls comparing each cause of death with all others.

‘Includes Alzheimers. % the 1987 survey, CVA was significantly more likely to be

associated with respondents who felt a” earlier death would have been better (P <0.05). Differences for IHD and respiratory death were not significant. Numbers were too low to test for ‘mental’ causes.

direct link between aspects of distress and dependency and views about dying earlier will now be explored. This is important in assessing what can be done by those who give care to relieve suffering, should one of the aims of this be to obviate the wish to die sooner. If pain, for example, is an important determinant, the argument that pain relief can reduce the likelihood of a wish to die sooner gains strength. Table 7 shows how the experience of pain affected this.

Pain is only weakly associated with views about an earlier death, as expressed both by respondents and as reported by the dying people themselves. Very distressing pain appears to be associated with more requests from the dying for euthanasia, but this is not statistically significant.

In fact other forms of distress and dependency show somewhat stronger associations with the desire to die sooner and requests for euthanasia, as Table 8 shows.

The measures of symptom distress are, for the most part, significantly associated with both respondents’ views and those of the dying people, with the exception of breath (in the 1987 survey appetite was also not significant). The associations for control and mental are somewhat stronger than the others. Dependency is also associated with these views. Reports of requests by the dying for euthanasia rise as levels of distress and dependency rise.

As the different types of distress and dependency were interrelated, multivariate analysis was used to assess which aspects of suffering had the strongest independent relationship with wishes to die earlier [9]. This was done separately for cancer and other causes of death and the results are shown in Table 9. Results

Table 7. Pain and views about dymg earlier (I990 weighted data; percentages, with base numbers in parentheses)

Respondents’ views

Deceased’s views

Level of pain

None Not very distressing Fairly Very All PC Cramer’s Y =

Better earlier

23 (341)

25 (183) 24 (409) 32 (688) 27 (1621)

0.01 0.09

Wanted Wanted Sooner euthanasia

17 (402) 2.6 (400)

2s (193) 2.8 (193) 25 (454) 2.8 (449) 29 (743) 5.0 (734) 25(1793) 3.7 (1775)

0.001 not significant 0.1 I 0.06


Table 8. Views about dying earlier and requests for euthanasia by level and type of suffering (1990 weighted data)

Appetite Control Mental Breath &pendency

Respondents’ views: % better if deceased had died ealier None 13 (216) 14 (663) I I (253) Medium 22 (267) 27.5 (247) 19 (373) High 30.5 (820) 37 (533) 34 (848) All 26(1304) 25 (1443) 26 (1474) P< 0.00005 0.00005 0.00005 Cramer’s v= 0.15 0.24 0.21

27.5 (691) 12.5 (404) 28 (409) 22 (421) 28 (548) 38 (539) 28 (1648) 25.5 (1364)

not significant 0.00005

0 0.25

Deceaseds’ views: % wanting to die ~mner None 13 (250) 16.5 (760) Medium 18 (299) 29 (266) High 30 (896) 32 (574) All 25(1445) 24 (1601) P< 0.00005 0.00005 Cramer’s v= 0.16 0.17

7 (299) 22.5 (762) I2 (468) 13(415) 23 (473) 24 (462) 34(913) 28 (590) 31 (585) 24 ( 1626) 24(1824) 23(1515) 0.00005 not significant 0.00005

0.28 0.06 0.19

Deceaseds’ views: % wanting euthanasia None 0.8 (250) I .7 (753) Medium 0.9 (299) 4.5 (264) High 5.7 (882) 5.7 (568) All 3.8 (1431) 3.6(1585) P< 0.00005 0.001 Cramer’s v= 0.12 0.1

I .2 (298) 2.5 (756) I .7 (465) I.1 (414) 3.9 (466) 2.8 (457) 5.3 (901) 5.2 (584) 5.9 (581) 3.5(1614) 3.7 (1806) 3.7(1503) 0.00005 0.05 0.001

0.11 0.06 0.1

are expressed as odds ratios, which can be interpreted as the increase in the odds of a person who has a high score for a particular measure of distress, compared

to a person who does not, scoring on the dependent variable (wanting to die sooner, for example). Thus

Table 9. Logistic regression showing the relative independent influence of high levels of symptom distress and dependency on views about dying earlier, and on reported requests for euthanasia (1990

datab)

High level Cancer Of: Odds

Respondents’ views Pain 2.8” Appetite 1.2 Control 1.7** Mental 3.2** Breath 1.0 Dependency 2.2” N= 934

Non cancer Odds

1.5 1.1 2.0’ 2.5’ I .o 3.9”

604

Deceaseds’ reported views Sooner Euthanasia Sooner Euthanasia

Pain 2.8” 1.1 1.2 Appetite a - 1.4 1.5 Control 1.6’ - 1.4 4.1 Mental 2.4* - 5.5** 2.6 Breath 1.0 - I.1 1.5 Dependency 1.4 - 2.6** 1.3 N= 1002 992 672 668

*P < 0.05. l *P < 0.01. ‘Could not be computed due to high standard error. Where a dash is shown goodness of fit for the model did not reach

an adequate significance level. The reference category for odds ratios is the absence of symptoms

or dependency. Thus only comparisons between high levels of distress and dependency and the reference category are reported here.

bFor respondents’ wishes in 1987 goodness of fit for cancer deaths did not reach an adequate significance level. For deaths from other causes 1990 significance levels for odds were replicated (conrrol odds = 3.8 (P < 0.05); mental odds = 4.7 (P < 0.05)) except in the case of dependency where the odds ratio (2.7) failed to reach significance.

one can say that of those dying from cancer in 1990, a high level of pain (the number in the top left hand comer of the first sub-table) increased the odds of respondents saying that it would have been better if they had died earlier by a factor of 2.8.

Table 9 shows that the presence of symptoms in the control and mental groups significantly increase the odds of respondents saying that it would have been better if the person had di.ed earlier. High levels of dependency also significantly increase the odds. Pain significantly increases the odds for cancer patients. Symptoms in the appetite and breath groups were not significantly associated with respondents’ views.

Considering the reported views of the person who died, a somewhat similar pattern is found, although there are important exceptions. Mental increases the odds of a person being said to have wanted to die sooner particularly strongly in the case of non-cancer deaths. As this symptom group includes depression, the expression of this feeling could be seen as a part and parcel of such a mental state. High levels of dependency increase the odds of wanting to die sooner for people not dying from cancer and the presence of pain increases the odds of this for those dying from cancer. The numbers reported as having asked for euthanasia are small, and this is why even quite high odds ratios fail to reach significance.

Social and cultural influences

A series of analyses was done to assess the hypoth- esis that social and cultural factors such as social class, strength and type of religious faith, and urban rather than rural residence influence peoples’ willing- ness to express a desire for an earlier death, or ask for euthanasia. Respondents were asked to assess both themselves and the deceased as having a ‘strong,’


Table 10. The influence of social and cultural factors on views about dying earlier, and requests for euthanasia (percentages with base

numbers in parentheses)

1987 1990 (weighted) Respondents’ Respondents‘

views views

1990 (weighted) Deceased’s

wews

Better earlier

Better earlier

Wanted Wanted sooner euthanasia

Social class of deceased I II IIInm 29(133) IIIm IV V 23 (265)

28 (520) 23 (5Xx) 3.5 (577) 25 (782) 24 (842) 3.5 (838)

Religion *

strong 15’(105) 25 (422) 25 (567) 3.5(560) SOfIle 28* (200) 27 (780) 23 (892) 3.6 (884) None 34’ (50) 31 (237) 27 (370) 4.6 (366) Place of residence Inner city 35’ (81) 24 (409) 23 (454) 3.2 (453) Other citv County ’

221 (I301 21* i223j

28 (894) 31 <417j

25 (993) 3.5(981) 24 {46Oj 4.3 ;455j

*P < 0.05. aRefers to religion of the deceased for deceased’s views. rehgmn of

respondent for respondents’ VIPWI.

‘some,’ or ‘no’ religious faith, as well as to report on the type of faith. In practice, analyses of the type of faith were restricted to Catholic, Church of England and other Protestants, as numbers reporting other types of faith were too low for analysis. The deceased’s social class was measured on the Registrar General’s scale, as recorded at death registration. Results are given in Table 10. The table shows only two significant associations, for strength of faith and place of residence in the 1987 survey. In view of the importance of religious views in the public debate about euthanasia, further analysis was conducted to discover any influence of type of faith on views about dying earlier and requests for euthanasia. The results of this are presented in Table Il. Significance tests (x ‘) were done for each category of religious faith, compared to ‘None’. There were no significant differences.

DISCUSSION

Two considerations should be borne in mind in assessing the significance of these results. Firstly,

Table I I. The influence type of religious faith on wishes to die earlier, and reauests for euthanasia

1987 1990 (weighted) Respondents’ Respondents‘

views views

1990 (weighted) Deceased’s

views

Faith’ Better earlier

Better earlier

Wanted Wanted scloner euthanasia

None AI/

Catholic C. of E. Protestant Strong faith Catholic C. of E. Protestant

34 (50) 31 (237) 27 (370) 4.6 (366)

21 (43) 25 (147) 22(184) 2.2(183) 26 (202) 27 (827) 25 (1000) 3.9 (989) I9 (32) 28 (132) 22(182) 3.4(180)

on/~ I5 (26) 24 (89) I9 (97) 3.7 (95)b lQ49) 25 (224) 27 (294) 4.6 (291) l5(13)b 3 I (75) 25 (128) l.3(127jb

‘Refers to religion of the deceased for deceased’s views, religion of respondent for respondents’ views.

bNumbers too low for valid test of significance.

wanting to die sooner may be a necessary precondi- tion for requests for euthanasia, but does not in itself constitute such a request. Respondents were asked to report on the dying peoples’ requests for euthanasia. but not on whether they themselves would have supported this. The numbers of people reported as requesting euthanasia were too low for some analyses, even with the fairly large data set available for 1990.

Secondly, the study did not involve interviews with dying people themselves. There are well known prob- lems in interviewing dying people; one has first to know that people are dying, for example, and this tends to restrict studies to those dying with cancer. However, this limitation probably means that reports of the incidence of requests for euthanasia are likely to be underestimates. This is because many respondents will not have known about every such request by the dying people during the year before death.

Cultural injiuences

The increasing proportion of elderly people in the population. with associated disability and distress, and the role played by modern medical technique in preserving life. combine to fuel public moral concern about euthanasia. The view that medical intervention at the end of life should not be excessive has often been conceded in medical circles [IO] on the grounds that there may be a point for some elderly or very ill people where life is no longer worth living. As well as the passive euthanasia of withholding treatment, active euthanasia may then be conceivable [1 I].

However, the argument for allowing euthanasia as a matter of social policy can also be understood as reflecting the emergence of the modern sense of self in public discourse. Armstrong has suggested that the call for euthanasia in modern conditions gives ‘the patient the right to speech and a claim thereby to be human’ [ 121. The modern sense of self is influenced by ideas about psychological fulfilment. achievable by exercising choice, asserting autonomy and maintain- ing control over the direction of one’s own life [13].

In the Netherlands, where active euthanasia under medical control is permitted, Fenigsen [I41 has identified the success of the pro-euthanasia movement as having depended on certain features of Dutch public life: a value placed on freedom of thought in a context of democratic liberalism, an encouragement given to the overthrow of taboos and opposition to authoritarian behaviour (which involves a degree of antimedical feeling), and a strong secular consciousness. It has been observed [15] that the discourse of consumerism and the language of rights is particularly strong in the U.S.A., thereby making that country more liable than Britain to accede to a social policy that allows active euthanasia.

Evidence suggests that the call for euthanasia is associated with modern, urban cultural conditions. Devins [16] in the U.S.A. found that urban dwellers


were more likely to support euthanasia, as did Jorgensen and Neubecker [17], also in the U.S.A. The latter study also found that whites, males and nonreligious people were more likely to be in favour. Kalish [ 181 found that Catholics were less likely to support euthaasia, younger people more likely. Lam [19] who surveyed members of EXIT, the British voluntary euthanasia society, found a disproportion- ate number of non-religious people from higher social classes.

The prominence of AIDS campaigners in calls for legalising euthanasia and providing for ‘living wills’ (these may express the wish for euthanasia should mental faculties deteriorate), can be understood in part as culturally determined. The living will is a gesture against future dependency and suffering [20]. AIDS campaigners represent urban, young, male and probably non-religious groups. Fenigsen [14] claims that 11.2% of Dutch AIDS patients die by active euthanasia-a much higher proportion than other disease groups.

Maas et nf. [21] investigated three classes of medical decisions concerning the end of life in large samples of deaths in the Netherlands. They estimate that 1.8% of all deaths in the country were caused by administering lethal drugs at the patient’s request (active euthanasia); 17.5% of all deaths involved such high dosages of opioids for symptom relief that in the opinion of the doctor they might have shortened life. In a further 17.5% of all deaths a decision not to treat (where treatment would probably have prolonged life) was made. Active euthanasia was more common amongst younger, male, urban dwellers. The other two classes of decision, though, were more common in women and in the elderly. Active euthanasia and the administration of high dosages of opioids were more common in cancer.

In spite of these indications in the literature that ideological and cultural factors play a part in influencing the public debate about euthanasia, on the whole the results (see Tables 10 and 11) did not confirm these as important influences on views about dying earlier or requests for euthanasia. City dwellers, those from higher social classes, and people from a particular faith, or who held particularly strongly to a faith, were not consistently more or less likely to say they wanted to die earlier, or want euthanasia. Social class and urban residence are only proxies for measures of cultural difference; it might be that if more direct measurement of attitudes on a wide range of topics had been done, significant associations would have been found. However, the absence of consistent association with the religious variables is more surprising, given the religious affili- ations of many who participate in the public debate about euthanasia. (It should be noted, however, that the results from the two data sets were contradictory on the matter of the strength of the respondents’ faith.) When nearing one’s own death, or recalling the death of a particular person to whom one may have

been close, it appears that religious considerations and cultural influences fade into insignificance in the face of the overwhelming physical and emotional experience of suffering.

Suffering

Recognition of the role of suffering in understand- ing the experience of illness has been urged by Stacey [22], who is concerned about an over-emphasis on cultural factors, and on the extent to which disease is socially constructed, in recent medical sociology. In this context, it is relevant to consider the assumptions about the role of suffering made by proponents and opponents of euthanasia.

Opposition to active euthanasia comes from several sources. On the one hand, there are specifi- cally religious objections [23]. On the other hand are the more secular arguments of many in the hospice movement, who propose an alternative solution to terminal suffering, based on the argument that good symptom control and emotional support obviate the desire to die earlier.

The hospice position has been outlined by Cicely Saunders. In an editorial in Palliative Medicine she argues that:

‘Kill me,’ a definite request for medically assisted suicide, though heard more often than it was 30 years ago, is still extremely uncommon. It may be voiced because of long unrelieved pain and is likely to fade away once this has been addressed as in almost all cases it can be (A)ttitudes change when a positive attitude and effective (pain) relief are introduced [24].

However, proponents of euthanasia refer to factors other than unpleasant physical symptoms. Thus Hurwitt, chairman of the Voluntary Euthanasia Society, has pointed out “the indignity (not merely the pain) of incurable illness” is at stake in the case for euthanasia [25].

The hospice movement is, perhaps, on firmer ground when speaking of pain and other symptom relief than the problem of extreme dependency. Saunders [24] recognises that images of unrelieved terminal cancer pain have become less prominent in calls for euthanasia in recent years. In their place have come fears of “extreme old age, brain failure and helplessness.” While she notes that “we cannot take all this away,” she feels that “we can ease and share it,” and at a practical level doctors can avoid the prolongation of such life by rejecting an “automatic commitment” to the preservation of life at all costs. In another article rejecting euthanasia [26] she argues that the deterioration of faculties, in particular “chronic brain failure” is rarer than many believe. The formation of “small community voiun- tary organisations to help the elderly” is cited as a method to help with the problem of dependency. The results of the study reported here show that certain forms of distress and dependency are more likely to lead to desires to die sooner, and to requests for euthanasia, than others. The importance of depen-

654 CLIVE SEALE and JIJLIA ADLXNGTO~-H~LI.

dency, and of the symptoms particularly associated with very old age, such as loss of bladder and bowel control and mental confusion, are demonstrated in Tables 8 and 9. Different causes of death involve different patterns of distress and dependency, with cancer tending to be very distressing, but with relatively low levels of dependency, and ischaemic heart disease involving relatively low levels of both (Table 5). Those dying with mental conditions experience relatively high levels of dependency. as well as certain types of distress. These patterns are then reflected in both respondents’ and the dying peoples’ views (Table 6). Pain is an important factor in desires to die earlier for those with terminal cancer. However, for those not dying from cancer, pain is not as important a factor as dependency in causing this state of mind (Table 9).

These findings have important implications for the public debate about euthanasia. particularly in relation to the position of the hospice movement. If good care is to obviate the desire to die sooner, it needs to address the problem of dependency as well as provide the symptom control in which hospice practitioners have developed such impressive exper- tise. Some environments are more suitable than others in helping disabled people lead independent lives, but there are inevitably limits to this. The issue of dependency in the elderly is a broad one, and may not be as amenable to remedy as certain symptoms have proved to be.

Acknowl~dRemmrs~The 1987 survey was carried out with financial support from the Medical Research Council in collaboration with Ann Cartwright. The 1990 survey was supported by the North East Thames Regional Health Authority, Bloomsbury and Islington District Health Authority, and other participating Health Authorities. and was done in collaboration with Mark McCarthy. The comments of Mark McCarthy, Michael Wadsworth, Sara Arber and Madeleine Simms on earlier drafts of this paper are gratefully acknowledged. The writing of this paper was made possible by a grant from the Leverhulme Trust.

I.

2.

3.

4.

5.

REFERENCES

British Medical Association. Euthanasia. BMA. London: 1988. Smith R. Euthanasia: time for a royal commission. Br. Med. J. 305, 7288729, 1992. For example: Singer P. A. and Siegler M. Euthanasia- a critique. N. Engl. J. Med. 322, 1881.-1883. 1990; Vaux K. L. The Theological Ethics of Euthanasia. Hastings Centre Report, special supplement, 1989; Reichel W. and Dyck A. J. Euthanasia: a contemporary moral quandary. The Lancer 1321-1323, 1989; Gillett G. Eu- thanasia, letting die and the pause. J. Med. Ethics 14. 61-68, 1988; Wilkinson J. The ethics of euthanasia. Palliat. Med. 4, 81-86, 1990. For example: Quill T. E. Death and dignity. N. Engl. J. Med. 324, 691-694, 1991; Annas G. J. Nancy Cruzan and the right to die. N. Engl. J. Med. 323,670.673 1990. Although this paper primarily reports the 1990 data, wherever possible analysis was done of 1987 data to assess whether it supported the picture given by 1990 data. Where the 1987 results do not support the 1990

6

7

x

9

I4

1s

16

17

IX

I9

20

21

22

23

24

25 26

results this is indicated either in the text or in footnotes. Note that the 1987 survey only contained one of the three dependent variables (concerning respondents’ own views). A full description of the methods used in the 1987 study is contained in: Cartwright A. and Seale C. The Natural History qf a Survey. Kings Fund, London, 1990: The 1990 survey is described in Addington-Hall J. and McCarthy M. Regional study of care for the dying: funding, methodology and sample characteristics. Submitted for publicatron. In all cases where a difference is reported in this paragraph, this involves a significance level based on chi-square at or below the 5% level. These findings about variation in distress and dependency by cause of death support those given in Seale C. F. Death from cancer and death from other causes: the relevance of the hospice approach. Palliariue Med. 5, I2 19, 1991; where it was found that cancer involved particularly intense levels of symptom distress, but that long term dependency was not characteristic, this being more likely in deaths from other causes. Cancer is the cause of death in a relatively young group of people. and is therefore more likely than other causes of death to have been reported on by a spouse rather than a child or other person. The independent influence of the type of respondent will be reported in another paper. but suffice to say that spouses tended to be less likely to wish that the person had died earlier. It is likely that this is why the relationship between cancer and respondents’ views is suppressed in Table 6. The data were checked for multicollinearity; a correlation matrix for the distress and dependency variables showed no correlations above 0.5. Smith T. Consensus on overtreating cancer. Br. Med. J. 297, 438, 1988. Euthanasia: an act of mercy or of murder’? The Guardian Education Supplement. October 6th. pp. 223. 1992. Armstrong D. Silence and truth in death and dying. Sot. .Sc,i. Med. 24, 651 -657, 1987. For an extended discussion of the role of psychology in creating this sense of self see Rose N. Gocw-ning the Socrl Routledge. London, 1989. Fenigsen R. The case against Dutch euthanasia. HUSI- ing.~ Centrc Rep-r Special supplement, pp. 22-30, 1989 Jennett B. Decisions to limit treatment. The Lancer 7x7 7X8 1987. Devins G. M. Contributions of health and demographic status to death anxiety and attitudes towards voluntary passive euthanasia. Omega 11, 2933302. 1980.-81. Jorgenson D. E. Neubecker R. C. Euthanasia: a national survey of attitudes toward voluntary termination of life. Omega II, 281-290, 1980~81. Kalish R A. Some variables in death attitudes. J. SOC. P.rycho/. 59, 137 145. 1963. Lam R. The queue for the EXIT. The Health Services .I. 15 October 1982. Higgs R. Living wills and treatment refusal. Br. Med. J. 295, 1221 1222, 1987. U.S.-style ‘living wills’ launched for Aids care. The Guardian 8th September 1992. Maas P. J., Delden J. J. M., Pijnenborg L. and Looman C. W. N. Euthanasia and other medical decisions concerning the end of life. The Lancer 338, 669%674, 1991. Stacey M. The SocYology qf Health and Heuling. Unwin Hyman. London, 1988. Whitfield S. Going gently into that good night. Nursing Times 83, 45. 30, 1987. Saunders C. Voluntary euthanasia. Palliative Med. 6, I ~5. I992 Hurwitt M. Letter to the Times, p. 17. 28 October 1991. Saunders C. Caring to the end. Nursing Mirror 4 September 1980.

euthanasia: why people want to die earlier

Documents