Relations between desire for early death, depressive symptoms and antidepressant prescribing in terminally ill patients with cancer

Abstract

Some patients with advanced cancer express the wish for an early death. This may be associated with depression. We examined the relations between depressive symptoms and desire for early death (natural or by euthanasia or physician-assisted suicide) in 142 terminally ill patients with cancer being cared for by a specialist palliative care team. They completed the Hospital Anxiety and Depression Scale questionnaire and answered four supplementary questions on desire for early death.

Only 2 patients expressed a strong wish for death by some form of suicide or euthanasia. 120 denied that they ever wished for early release. The desire for early death correlated with depression scores. Depressive symptoms were common in the whole group but few were on antidepressant therapy.

Better recognition and treatment of depression might improve the lives of people with terminal illness and so lessen desire for early death, whether natural or by suicide.

INTRODUCTION

The recent case involving a woman with motor neuron disease seeking immunity from prosecution for her husband should he assist her to die¹, along with legislative changes in the Netherlands decriminalizing euthanasia², has intensified an already lively debate^3,4. The prominence given to this issue might suggest it is encountered frequently in clinical practice. In particular, one might expect doctors caring for terminally ill patients with cancer to be dealing with numerous requests for physician-assisted suicide and euthanasia. There is very little research on the prevalence of requests for euthanasia in terminally ill cancer patients, and the few studies that exist from North America suggest that such requests are rare^5,6,7. In any event, supporters of physician-assisted suicide and euthanasia argue that patient autonomy is ultimate and that such requests should be heeded. This assumes full insight, informed consent, and competence in decision-making on the part of the patient, which may not always be the case. There may be many thoughts and emotions underlying such requests which should be explored with the patient⁸.

The psychological impact of a diagnosis of terminal cancer on an individual will be influenced by many factors. Depression, if present, is likely to affect end-of-life decisions. Beck has described a ‘cognitive triad’ in depressed patients consisting of negative views of self, the outside world and the future⁹. Other important factors affecting psychological adjustment to cancer are premorbid personality, ability to cope, social supports, and religious and cultural beliefs and values.

Some patients with advanced cancer may harbour the wish for an early or hastened death. This may manifest itself inter alia as hopelessness about the future, pleas for an end to their suffering, refusal of various treatments, suicidal ideation, direct requests for physician-assisted suicide or euthanasia, and suicidal acts. The frequency of suicide in the cancer population seems somewhat higher than in the general population¹⁰, with highest risk in the immediate period after diagnosis¹¹. Suicide is apparently rare in terminally ill patients¹². Suicidal behaviour or ideation is frequently associated with depressive illness. In patients with cancer, a strong link between desire for death and depression has been suggested⁶. At least 25% of patients with cancer, and as many as 77% of those with advanced cancer and serious impairment in performance status, have depressive symptoms¹³. Effective treatment reduces depressive symptoms in such patients¹⁴ and improves their quality of life. Yet depression in cancer patients seems to be underdiagnosed and undertreated¹⁵.

Various methodologies have been used to examine desire for death in patients with cancer. Chochinov et al.⁶, using the clinician-rated Desire for Death Rating Scale (DDRS), found desire for death to be rare among terminally ill cancer patients. Recently the same group¹⁶ measured changes in will to live over time among patients in palliative care. They found that patients had a strong will to live across all time intervals but that will to live fluctuated considerably for individuals. Rosenfeld and colleagues¹⁷ developed a self-report measure of desire for death, the Schedule of Attitudes toward Hastened Death (SAHD), which they validated in patients with terminal cancer¹⁸. Using this instrument, Breitbart et al.¹⁹ found that 17% of 92 such patients had a high desire for a hastened death.

In this study, we set out to establish the relations between depressive symptoms, desire for early death, and antidepressant medication in a population of patients with cancer receiving specialist palliative care.

PATIENTS AND METHODS

We studied patients referred to the palliative care team based at Our Lady's Hospice, Harold's Cross, Dublin, over four months. The sample consisted of consecutive patients referred for care at home and in the hospice inpatient unit, and those being treated at a nearby university teaching hospital. Patients who met the following inclusion criteria were recruited: intact cognitive status (i.e. a score of seven or greater on the Abbreviated Mental Test²⁰); full insight into the diagnosis and likely prognosis; willing and able to complete the questionnaires. Eligible patients were approached for consent and asked to complete the Hospital Anxiety and Depression Scale (HADS)²¹ and four supplementary questions relating to desire for death (see Table 3). The HADS consists of two subscales comprising seven items measuring anxiety and seven measuring depression, with each item scored from 0 to 3. Those who devised the scale suggest using a score of 8-10 on each subscale to identify ‘borderline’ cases and a score of 11 or more to identify ‘definite’ cases.

Table 3.

Frequencies of responses to questions on desire for death

Question	Score	Response options	No.	(%)
I go to sleep hoping that I won't wake up	0	Never	120	(84.5)
	1	Some of the time	13	(9.6)
	2	A lot of the time	1	(0.7)
	3	All of the time	1	(0.7)
		Missing	7	(4.9)
I think of ending my life, but I would not do it	0	Never	123	(86.6)
	1	Occasionally	11	(7.7)
	2	A lot of the time	2	(1.4)
	3	All of the time	1	(0.7)
		Missing	5	(3.5)
I would end my life if I had a chance	0	Definitely disagree	133	(93.7)
	1	Partly agree	2	(1.4)
	2	Mostly agree	0	(0)
	3	Definitely agree	1	(0.7)
		Missing	6	(4.2)
I wish the doctors would do something to end my life	0	Never	129	(90.8)
	1	Some of the time	6	(4.2)
	2	A lot of the time	0	(0)
	3	Definitely	2	(1.4)
		Missing	5	(3.5)

The four questions on desire for early death were designed to encompass the range of emotions from passive death-wish and suicidal ideation to potential requests for euthanasia. They were first presented to members of the palliative care team for comment on face validity and acceptability, and then piloted with 20 inpatients in the palliative care unit for acceptability and comprehensibility. Each question was scored 0-3, with significant death-wish considered present if the patient scored 2 or 3 on any question. The individual scores for the four questions on desire for early death were recorded, and were also summed to produce a total score for use in correlational analyses with HADS scores. Patient records and prescription charts were reviewed to identify patients taking antidepressant medication and the indication for it. Approval for the study was obtained from the St Vincent's University Hospital Ethics and Medical Research Committee.

RESULTS

142 patients completed the HADS and the four additional questions. 72 (50.7%) were male, and ages ranged from 33 to 90 years (mean 66.19, SD 12.44). The cancer types (Table 1) mirrored those seen in the general population. An additional 140 patients were ineligible for inclusion in the study—67 because they were too unwell, and 73 because they lacked insight into their diagnosis and prognosis. There were no significant differences in sociodemographic or medical characteristics between those patients who completed the questionnaires and those who were not eligible.

Table 1.

Sociodemographic and medical characteristics of patients (n=142)

Characteristic	No. (%)
Gender
M	72 (50.7%)
F	70 (49.3%)
Place of care
Home	38 (26.7%)
Hospice inpatient	59 (41.6%)
Acute hospital inpatient	45 (31.7%)
Cancer diagnosis
Lung	28 (19.7%)
Gastrointestinal	30 (21.1%)
Breast	25 (17.6%)
Prostate	12 (8.5%)
Ovary	6 (4.2%)
Genitourinary (other)	6 (4.2%)
Liver/pancreas	7 (4.9%)
Head and neck	8 (5.7%)
Brain	5 (3.5%)
Other	12 (8.5%)
Unknown primary	3 (2.1%)

Table 2 shows how patients scored on the HADS depression subscale. With the most conservative cut-off point, 17.6% were judged depressed; with inclusion of borderline cases the prevalence was 38%.

Table 2.

HADS depression subscale scores for 142 patients

HADS depression subscale score	No. of patients (%)
<8	88 (62.0%)
⩾8, < 11	29 (20.4%)
⩾11	25 (17.6%)

HADS=Hospital Anxiety and Depression Scale

Table 3 lists the four questions on desire for early death and the frequencies of responses to each. 2 patients expressed a strong wish for death by euthanasia, with one of them also expressing a strong desire for death by suicide. There was a significant correlation between desire for early death (the four questions combined) and the HADS total score (Table 4). Partial correlations were calculated for desire for early death and the anxiety subscale with depression controlled, and for the depression subscale with anxiety controlled. The depression subscale was seen to account for the correlation.

Table 4.

Pearson's correlation between desire for death and total HADS score, and partial correlation with the anxiety and depression subscales

	Coefficient	Significance
Pearson's correlation
HADS total score	0.4045	P<0.01
Partial correlations
Anxiety subscale	0.0785	NS
Depression subscale	0.4338	P=0.0005

HADS=Hospital Anxiety and Depression Scale

18 (12.7%) patients were taking antidepressant medication prescribed for the treatment of depression—a selective serotonin reuptake inhibitor in 13, a tricyclic antidepressant in 4 and a monoamine oxidase inhibitor in 1. A further 6 patients were taking tricyclic antidepressants as treatment for pain. Table 5 lists the characteristics of the patients receiving antidepressants for the treatment of depression.

Table 5.

Patients receiving antidepressants for the treatment of depression

	HADS depression subscale scores (n=142)
	<8 n=88 (62%)	8-10 n=29 (20.4%)	≥11 n=25 (17.6%)
Antidepressant
Yes	6 (7%)	4(14%)	8(32%)
No	82(93%)	25(86%)	17(68%)

Of the 25 patients who scored ≥11 on the HADS depression subscale, only 8 were taking antidepressants for depression. Of the 54 who scored ≥8 on the depression subscale, 12 were taking antidepressants for depression. 6 patients taking antidepressants for depression had scores of <8 on the HADS depression subscale. Only 2 depressed patients expressed a strong desire for early death, and one of these was not on an antidepressant.

DISCUSSION

Desire for early death (either natural or by suicide or euthanasia) was rare among the 142 terminally ill patients in this study. Only 2 patients clearly wanted a doctor to do something to end their lives. Similar results have been reported from the USA: despite the high profile of the issue in that country, requests for assisted death seem rare⁷. In our study the prevalence of desire for death was lower still. Chochinov et al.⁶ found that 17 of 200 terminally ill cancer patients (8.5%) had a sustained desire for death. Our results are even further at variance with those of Breitbart et al.¹⁹, who identified 17% of 92 terminally ill cancer patients as having a high desire for a hastened death. The discrepancy may be explained by the difference in questions used: the four questions in our study were focused specifically on early death, whereas in Breitbart's study (because of items in the questionnaire) there may have been difficulty in differentiating patients with a greater ‘acceptance’ of death from those who desired early death. Neither are our results directly comparable with Chochinov's more recent study looking at will to live¹⁶. Lack of a will to live, though clearly related to a desire for death, does not necessarily equate with a desire for early or assisted death.

In the Netherlands, 6-7% of patients with advanced cancer eventually choose to die with physician assistance²². Hendin has suggested that the legal sanctioning of assisted suicide and euthanasia in the Netherlands may be promoting a culture that encourages patients and doctors to view early death as a preferred way of dealing with suffering and terminal illness²³. Emanuel likewise suggests that legislation focusing on euthanasia and physician-assisted suicide diverts energy from the activities necessary to improve end-of-life care³. The low rate of desire for early death in our study may partly reflect cultural and legislative differences between Ireland and the Netherlands.

Though desire for early death was not common in our patients, when present it was associated with significant depressive symptoms. Two Canadian studies^5,6 in terminally ill patients with cancer have shown a similar association; and in patients with amyotrophic lateral sclerosis, Ganzini et al. identified hopelessness as a factor in consideration of assisted suicide²⁴. In a group of people with HIV-1 or AIDS²⁵, loss of community and disintegration were the two main factors influencing desire for euthanasia and assited suicide.

One of the 2 patients in our study with a strong death-wish and depressive symptoms was not on antidepressant medication. Kugaya et al.²⁶ reported on the retraction of requests for euthanasia in five cancer patients after successful treatment of depression. The association between depressive symptoms and death-wish highlights the importance of early recognition and effective treatment of depressive symptoms in patients with cancer. Depressive symptoms (present in at least 17.6% of our patient group) may exist in isolation or may be part of an adjustment disorder. They are not necessarily synonymous with depressive illness, although this diagnosis must always be considered. The distinction between depressive illness and adjustment disorder in a patient with depressed mood is important in guiding treatment²⁷. Although not a diagnostic instrument, the HADS has been shown to have high sensitivity and specificity in identifying major depression in patients with cancer²⁸. According to Hopwood et al., a cutoff score of 11 for the depression subscale of the HADS is optimal in terms of sensitivity, specificity and positive predictive value²⁹. Depression in patients with cancer does respond to antidepressant medication^30,31; so our finding that only 32% of patients scoring ≥11 on the HADS depression subscale were taking antidepressants probably reflects undertreatment. Formal psychological therapies are also effective in such patients³², but no patient was undergoing these.

Though the findings of this study may contribute to the debate on physician-assisted suicide and euthanasia, certain limitations must be acknowledged. The sample consisted of all patients attended by a specialist palliative care team. They may differ, therefore, from other patients dying from cancer who are receiving less adequate palliative care. Unlike the work of Chochinov^6,16 and Breitbart¹⁹, our study included patients at home and in a general hospital, as well as patients in a palliative care unit, so the findings may be more representative. The fact that only half the patients seen during the study period completed all questions may have introduced a bias, though this could have been towards under-representation of depression and desire for death: Chochinov has shown that patients without full insight into their prognosis (and thus excluded from our study) are more likely to be depressed than those with full understanding³³.

Finally, in view of the special cultural and religious traits of Irish people, we cannot say whether similar results would be obtained in other countries.

Though we demonstrated a correlation between desire for early death and depressive symptoms, clearly other factors are important, among them spiritual wellbeing, quality of life, physical symptoms, physical functioning, and the perception of being a burden to others¹⁹.

Hendin asserted that ‘the overwhelming majority of people who are terminally ill fight for life to the end’³⁴. Our findings indicate that this is true of patients terminally ill with advanced cancer receiving specialist palliative care in Ireland. Further research is needed to establish whether antidepressants are being under-used in this setting.