Article Text

Download PDFPDF

Prevalence, course and associations of desire for hastened death in a UK palliative population: a cross-sectional study
  1. Annabel Price1,
  2. William Lee1,
  3. Laura Goodwin1,
  4. Lauren Rayner1,2,
  5. Rosemary Humphreys3,
  6. Penny Hansford4,
  7. Nigel Sykes4,
  8. Barbara Monroe4,
  9. Irene Higginson2 and
  10. Matthew Hotopf1
  1. 1Department of Psychological Medicine, Institute of Psychiatry, King's College London, London, UK
  2. 2Department of Palliative Care, Policy and Rehabilitation, Cicely Saunders Institute, King's College London, London, UK
  3. 3South London and Maudsley NHS Foundation Trust, London, UK
  4. 4St Christopher's Hospice, London, UK
  1. Correspondence to Dr Annabel Price, Department of Psychological Medicine, Weston Education Centre, 10 Cutcombe Road, London SE5 9RJ, UK; Annabel.Price{at}


Objectives To determine the prevalence, severity and remission of desire for hastened death (DHD) in a UK representative sample of patients with advanced disease receiving palliative care and to examine the associations of desire for death.

Design A cross-sectional survey with 4-week follow-up.

Setting St Christopher's Hospice, Sydenham, South London, which is a large hospice with homecare, outpatient and inpatient facilities serving five London boroughs.

Participants 300 patients newly referred to the hospice for palliative care.

Main outcome measures The Desire for Death Rating Scale (DDRS).

Results At T1 33/300 (11%) reported DHD and 11/300 (3.7%) had more serious or pervasive DHD. Of those who expressed DHD at T1 and were interviewed at both time points, 35% no longer reported these thoughts. Of those who reported no DHD at T1, 8% reported DHD at T2. The majority of those who had more severe DHD at T1 had a reduced DHD score by T2. Factors associated with T1 DHD included presence of non-malignant disease, depression, more severe physical symptoms, hopelessness and perceived loss of dignity.

Conclusions The prevalence of DHD was at the lower end of that seen in previous studies using similar samples. More severe DHD was uncommon and for most part remitted to some extent during the study. The provision of symptom control and timely detection and intervention for depression coupled with a focus on optimising function, instilling hope and preserving dignity are likely to contribute to alleviation of DHD in patients with advanced illness.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Desire for death in patients with advanced disease has received particular attention in recent years, in part due to the ongoing debate over assisted dying for the terminally ill. At present, assisted dying for the terminally ill remains illegal throughout the UK, but the End of Life Assistance (Scotland) Bill1 was recently debated by the Scottish Parliament and it is likely that a new assisted dying bill will be introduced in the UK Parliament in the future, the last being overturned in 2006.2

Desire for hastened death (DHD) has been defined as (a) a passive wish (fleeting or persistent) for death without active plans, (b) a request for assistance in hastening death or (c) a plan to commit suicide.3 The experience of desire for death in advanced cancer may be categorised as (a) a theoretical exit plan, (b) an expression of despair or (c) a manifestation of ‘letting go’.4 Methods used for measuring DHD include the observer rated Desire for Death Rating Scale (DDRS)5 and the self-rated Schedule of Attitudes toward Hastened Death (SAHD).6

Studies of cancer patients have found varying prevalence rates for DHD. In a review of studies of DHD in cancer patients, Rodin et al found prevalence estimates ranging from 0% to 28% with sample sizes ranging from 60 to 256, with the wide range likely to be explained by heterogeneity in patient samples and also assessment methods and criteria for measuring DHD. Their study with a sample size of 326 patients with metastatic cancer using the SAHD found a prevalence of 53.1% for scoring ‘low’ on DHD, 8.3% for scoring ‘mild’ and 1.5% for scoring ‘high’.7 In a recent Korean prevalence study of 131 cancer patients, 13.7% scored ‘moderate’ on DHD and 1.7% scored ‘high’ using the SAHD.8

For cancer patients receiving palliative care, prevalence estimates range from 1% to 28% for ‘higher’ or ‘stronger’ DHD with sample sizes ranging from 92 to 256, again the range likely to be explained by heterogeneity in assessment criteria.5 9,,11 Two studies of DHD in palliative cancer patients using the DDRS both reported a prevalence of any desire for death of 45%.5 12

DHD has been examined less extensively in non-cancer patients with terminal or advanced progressive diseases, although in the Netherlands it has been reported that 20% of patients with motor neurone disease die as a result of euthanasia, compared with 5% of cancer patients and 0.5% of those with heart failure.13 In a sample of 47 patients with advanced AIDS receiving palliative care, 15% were rated as having a ‘serious and pervasive’ DHD, while 48% had any level of DHD.14

Associations of DHD have been consistently demonstrated with depression, hopelessness, high physical symptom burden, reduced physical functioning, low social support and lack of self-esteem.7 9 15,,17 Pre-existing psychiatric illness has also been identified as a risk factor for DHD.14 Personal factors also have an important role, with studies showing that patients without religious faith or with poor spiritual well-being have higher rates of DHD.16,,18 Perceived burden to others, loss of autonomy, wish to control death, existential concerns and fear of the future have also been identified as risk factors.3 15 Subjective loss of dignity is associated with DHD19 and is negatively correlated with will to live in the terminally ill.20 In studies investigating the attitudes of terminally ill individuals towards the legalisation of euthanasia or assisted suicide, desire for death was strongly associated with a current interest in pursuing one of these avenues.17 21

One potential issue in the legalisation debate is whether DHD is a stable phenomenon, however research on its longitudinal course is limited. In one longitudinal study, 64 patients with cancer pain were interviewed at two time points 4 weeks apart. Follow-up scores on a DHD measure were significantly higher than at baseline, but mean scores were low at both time points.16 Chochinov et al followed up six of an original sample of 200 terminally ill patients after 4 weeks. Of the six, four showed a decline in DHD and two reported a sustained wish to die.5

A recent US study interviewed 209 patients with advanced HIV at two time points 1 month apart and found that fluctuations in DHD were rare in patients who were not depressed. Also examined was the impact of antidepressant treatment on DHD in the 42 depressed patients who participated at both time points. A strong relationship was found between improvement in depression and reduced desire for death, and antidepressant medications were found to contribute to this relationship.22

DHD is likely to be influenced by values held within a culture such as religious beliefs. Most studies of DHD in patients with advanced disease have been carried out in American and Canadian samples, and while other studies have used Greek,23 Irish11 and Korean8 samples, there are no studies from UK patients with advanced disease.

Methodological approaches vary between studies, but the majority either do not report or give a limited description of the characteristics of participants versus non-participants, making it unclear how representative of the study population the samples are.

Information about the prevalence, remission and associations of DHD in a UK population with advanced disease will help clinicians to identify those at risk, and tailor assessment and intervention to the particular needs of this group. It will also allow for comparison between countries and provide data to inform the ongoing assisted dying debate in the UK.

This study aims to determine the prevalence, severity and remission rate of desire for death in a large representative UK sample of patients with advanced illness receiving palliative care and to examine the associations of desire for death.


Study design

This was a cross-sectional survey with subsequent follow-up of patients receiving palliative care who were newly referred to a hospice. This study was part of a wider survey of depression in palliative care, the results of which have been published elsewhere.24 Patients were interviewed within 1 week of referral to the service and then by telephone 4 weeks later.


Patients were recruited from St Christopher's Hospice, Sydenham, a large hospice in south east London with both a home care and inpatient service serving a large population across five boroughs. Recruitment commenced in January 2007 and all new patients accepted by the hospice were eligible for inclusion. The aim was to enrol 300 patients in the study. Patients were excluded if they lacked capacity to consent, had very poor functional status (scoring 4 on the Eastern Co-operative Oncology Group (ECOG) scale)25 or were too ill to participate, or had sensory impairment or language difficulties such that they would not be able to participate in the research interview.

Potential participants were initially assessed by a clinical nurse specialist who determined whether they were eligible for recruitment. Eligible patients were then contacted by telephone by study personnel to ask if they would be willing to participate. A face-to-face interview was then arranged by one of three researchers, usually at the patient's home, within a week of recruitment. Recruitment was completed in August 2008.


The first interview (T1) took place within 1 week of the patient being accepted by the hospice. After gaining written consent to participate, the interviewer administered the Abbreviated Mental Test Score.26 The interview was terminated for those who scored below 6/10, indicating cognitive impairment. The interviewer then administered several questionnaires including those described below, lasting approximately 1 h. The follow-up interview (T2) was conducted by telephone 4 weeks after the first interview; this interview took approximately 15 min to complete. Demographic and clinical data for participants and eligible non-participants were obtained from the patient notes. Ethical approval for this study was obtained from the Institute of Psychiatry NHS Research Ethics Committee (06/Q0706/93).


T1 and T2

Desire for Hastened Death

DHD was measured using the DDRS developed by Chochinov et al5 which consists of a semi-structured interview which begins with the question ‘Do you ever wish that your illness would progress more rapidly so that your suffering could be over sooner?’. If this question is answered in the affirmative, then further questions are asked to determine how serious and pervasive is the DHD. Following responses to the questions, the patient is assigned a score on an interviewer rated 7-point scale (see table 1). The DDRS has been used in terminally ill populations and has shown good concordance with visual analogue scale assessments.12 The presence of desire for death was defined as scoring >1 on the DDRS, producing a dichotomous variable which was used as the outcome for further analysis of associations. The DDRS was chosen over the SAHD because one objective of our study was to compare the prevalence of DHD between UK and non-UK palliative populations. Previous studies of directly comparable populations have used the same rating scale.5 12 We initially aimed to assess inter-rater concordance for scoring on the DDRS, but this was not possible because arranging several visits with different interviewers was felt to be inappropriate given the advanced stage of illness of the majority of participants.

Table 1

The Desire for Death Rating Scale (DDRS)

T1 only


Presence of depression was assessed using the PRIME-MD (Primary Care Evaluation of Mental Disorder), a subscale of the Primary Health Questionnaire27 used to detect the presence of psychiatric disorder in general populations, based on the DSM-IV diagnostic criteria for depression. Participants were assessed for the presence of ‘any depressive syndrome’ or ‘major depressive disorder’.

Loss of dignity

Subjective loss of dignity was assessed using a semi-structured interview and subsequent interviewer rating on a 7-point scale (0–6). This method of assessment developed by Wilson and colleagues has been shown to have good inter-rater reliability.21 A binary item was created for the analysis which compared ‘no loss of dignity’ and ‘minimal, not a problem’ (0–1) with ‘mild’ to ‘extreme loss of dignity’ (2–6).

Further measures

The European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire version 3 (EORTC-QLQ-C30)28 was used to examine health-related quality of life. The Brief Illness Perception Questionnaire (Brief-IPQ)29 was used to assess the cognitive and emotional representations of illness. The 40-item Mental Adjustment to Cancer Scale (MAC)30 was used to assess how people adjust to having a cancer diagnosis. Wording was slightly adapted for those without cancer in the sample. MAC scores were categorised into the original four factors: hopelessness-helplessness, fighting spirit, anxious preoccupation and fatalism. The Multidimensional Scale of Perceived Social Support31 was used to assess perceptions of social support adequacy from family, friends and a significant other. The total score for social support from all sources was used in the analyses.

For each scale, the score for each individual item was categorised by quartile splits into those scoring equal to or between the 25th and 75th percentiles (reference category), compared to those scoring below the 25th percentile, or those scoring above the 75th percentile. Some categories were combined if cell sizes became too small for meaningful analysis.


All data management and analysis was conducted using Stata v10.0.32

Multivariable logistic regression analyses were conducted to calculate ORs for the associations between each of the individual predictors and the outcome (no desire for death vs any desire for death). Analyses of the associations of DHD were adjusted for the presence of ‘any depressive syndrome’ except for demographic and clinical factors (table 3). Tests for trend were conducted to examine whether the goodness of fit of the models improved when the predictor of interest was added as a continuous variable (ordered from the lowest to the highest category) to the model.


Sample characteristics

During the study period 743 patients were eligible for inclusion into the study. Forty per cent of eligible patients entered the study. Four hundred and forty three eligible patients were not interviewed: 278 patients declined to take part, study personnel were unable to contact 83 patients, for 44 patients the treating clinician deemed the patient unsuitable for the study, the family declined on behalf of 13 patients and 25 patients did not participate for other reasons.

Three hundred participants were interviewed at T1. Just under half were female (49.3%) and the mean age was 68.5 years. The majority of participants were of white ethnicity and the distribution across occupational groups was fairly even. The most common diagnosis was lung cancer (27%) and the majority of the sample (59%) had metastatic disease. Most participants were restricted in physical activity but capable of self-care (80% ECOG 1 or 2). One hundred and nine participants (36.3%) met the criteria for ‘any depressive syndrome’ which includes both major depressive disorder and minor depression. Two hundred and thirteen patients were interviewed at T2, with the majority of T2 non-participants having died or become too unwell to be interviewed.

There was no difference in age, sex or diagnosis between participants and non-participants, however more participants had metastatic disease than non-participants. There were no significant differences in demographic or clinical characteristics between those who did and did not participate at both time points.

Prevalence and course of DHD

Figure 1 shows the prevalence of DHD at T1 and T2 and how DHD changed between the two time points. At T1, 33/300 (11%, 95% CI 7.5% to 14.5%) scored >1 indicating some DHD and 11/300 (3.7%, 95% CI 1.5% to 5.8%) scored 4 or more indicating a more serious or pervasive DHD. At T2, 30/213 (14%, 95% CI 9.4% to 18.8%) scored >1 indicating some DHD and 7/213 (3.3%, 95% CI 0.9% to 5.7%) scored 4 or more indicating a more serious or pervasive DHD.

Figure 1

Prevalence and course of desire for hastened death

Of those who expressed a DHD at T1 and were interviewed at both time points, 35% no longer reported these thoughts. Of those who reported no DHD at T1, 8% reported a DHD at T2.

Table 2 shows the DDRS scores for the 207 patients who participated at T1 and T2. Of the nine participants who scored 4 or above at T1, seven had a lower score at T2 and of the 198 who scored 2 or less at T1, 17 (8.6%) had a higher score at T2.

Table 2

Desire for Death Rating Scale scores, T1 (row) versus T2 (column)

Associations of DHD

Demographic and clinical factors

Table 3 compares demographic and clinical variables between those who did and did not express a DHD at T1.

Table 3

Demographic and clinical associations of desire for hastened death

The only factor which was significantly different for those with DHD was disease status: those with non-malignant disease had greater odds of DHD compared with the reference group with metastatic cancer. Whether participants had a history of depression or were referred for psychological support was not associated with a DHD, neither was time to death.

Assessment measures

Table 4 examines the association between depression, quality of life measures, symptoms, illness perceptions, social support, adjustment to illness, dignity, religious observance and DHD at T1.

Table 4

Associations of desire for hastened death

Depression and suicidality

The presence of both major depression and having any depressive disorder was associated with increased odds of DHD and 63.6% of the group with DHD met criteria for any depressive disorder. The presence of suicidal thoughts in the past 2 weeks was unsurprisingly associated with DHD, however 24.2% of those with DHD did not express suicidal ideas, and more than half of those with suicidal ideas did not express DHD.

Quality of life and function

In the unadjusted analysis, those with higher scores on measures of global quality of life, physical functioning, role functioning, cognitive functioning and social functioning had a reduction in odds for DHD. After adjusting for the presence of depression, the associations remained for global quality of life, physical functioning and cognitive functioning.

Social factors

Patients scoring in the mid category for financial difficulties had increased odds of DHD, compared to patients reporting low financial difficulties, which remained after adjusting for depression; however those in the highest category for financial difficulties did not have significantly increased odds for DHD. Level of social support was not found to be associated with DHD, although there was a trend towards the odds of DHD reducing as the level of social support increased.

Illness perceptions

Those with higher scores on identity (how much one experiences symptoms from the illness) and emotion (how much one is emotionally affected by the illness) had higher odds of DHD, which remained after adjustment for depression. There were also trends towards those who perceived higher consequences and concern, and lower personal control and treatment control having greater odds of DHD.

Adjustment to illness

Those who scored in the highest quartile for helpless-hopelessness and fatalism, and those who scored in the lowest quartile for fighting spirit had increased odds of DHD. These effects remained after adjusting for depression. The greatest odds were for those scoring in the highest quartile for helplessness-hopelessness.


Those who perceived more than a mild loss of dignity (score 3–6) had increased odds of DHD which remained after adjusting for depression.

Religious observance

Religious observance was not associated with DHD.


Our findings show that prevalence of DHD in our sample is at the lower end of the range shown in previous studies of similar populations using a variety of measures, and is substantially lower than in the two previous studies of similar samples using the DDRS.

More severe or persistent desire for death was uncommon in our sample and for most patients remitted to some extent over the study period. Over a third of those who reported any desire for death at T1 no longer did so at T2, and only one participant maintained a strong desire for death across time points. However, by the second interview a number of participants had reported DHD anew, with a slightly higher proportion (14%) expressing DHD at the second interview. These findings suggest that that DHD is not stable over time.

Patients with non-malignant disease had increased odds of DHD. This is the first study to show this. We suggest that the most likely explanation is selection into palliative care services. A much higher proportion of cancer patients have some contact with specialist palliative care than non-cancer patients, and the non-cancer patients may often be referred and accepted because of their level of distress.

In contrast to previous findings, we did not find that a previous history of depressive illness or a lack of religious faith were associated with DHD. Objective functional status and proximity to death were also not associated with DHD, showing that for our sample DHD was probably not closely linked with the experience of approaching death. Suicidal thoughts and DHD often coexisted, but a significant minority of those with DHD did not express suicidal ideas, indicating that the two are not necessarily synonymous.

Other findings reflect those of previous studies. The strongest positive associations of DHD were the presence of depression, suicidal thoughts, higher scores on measures of fatigue, pain, insomnia and dyspnoea, greater perception of symptoms and emotional effects, hopelessness and perceived loss of dignity. Negative associations were shown for global quality of life and physical, cognitive and emotional functioning. There was a trend towards a negative association between social support and DHD.

The provision of effective symptom control and timely detection and intervention for depression coupled with a focus on social support, optimising function and coping, instilling hope and preserving a sense of dignity, are likely to contribute to alleviation of DHD in patients with advanced illness; these are all interventions which are within the scope of palliative care services.

The argument for legalisation of assisted dying is based on provision for those with a persistent wish for death. People fulfilling this criterion are likely to be a small minority of those expressing a desire to die and there may be a number of remediable contributory factors for those with even the strongest DHD.

Clinical implications

Careful assessment of those expressing DHD in a palliative setting may reveal a number of biological, psychological or social contributory factors which are potentially remediable with targeted interventions. DHD should be enquired about for those with an identified depressive illness and depression should be managed proactively. Studies have shown that depression is treatable in patients with advanced disease. A recent systematic review provided evidence of the effectiveness of antidepressants for patients with life threatening illness33 and guidelines for the treatment of depression in palliative care have been published.34 A cluster randomised trial showed that cognitive behavioural therapy is effective in reducing both depression and anxiety in palliative patients35 and palliative care provision in itself may be an effective intervention for depression as shown in a recent trial comparing early palliative intervention versus usual oncological treatment.36

Strengths and limitations

Research using samples of patients with advanced disease presents methodological challenges. This study achieved a large representative sample in spite of these with an enrolment rate similar to those reported in previous surveys in palliative care.

Asking about DHD is a sensitive issue and its measurement and interpretation are complex. There is a potential risk of interpreting responses on assessment tools as equating to desire for death in themselves when it is likely that several factors are measured under the rubric of DHD – many healthy older people might express a desire for death which could be a manifestation of acceptance that life has run its course; at the other extreme are probably a group of patients who experience severe suffering and can see no relief from their distress other than by desiring a hastened death. Future work might further explore the phenomenology of DHD to draw out such distinctions. We were not able to explore the factors associated with a change in DHD across time points because the outcome was less frequent than we anticipated. Analysis of non-remission would have provided valuable information about those who experience more persistent DHD who may be more likely to seek hastened death or be considered for assisted dying were this legal. Our sample derived from a group of patients who were receiving specialist palliative care. This might mean that they were therefore more accepting of their disease and its prognosis, or that they were more distressed than other patients at the same point in their disease trajectory. Future work might be based on populations of patients with specific life threatening disease derived from primary care. Finally, we tested multiple associations without applying any adjustment to p values. We did this because many of the variables relating to quality of life, depression, symptoms and adaptation to illness were correlated. The effect sizes were generally large or very large, suggesting that the associations are unlikely to be due to type 1 error.


In our sample, the prevalence of DHD was at the lower end of that seen in previous studies using similar samples and considerably lower than that seen in previous studies using the same measure of DHD. More severe DHD was uncommon and for most DHD remitted to some extent across the study period. DHD was associated with a number of factors including depression, quality of life, physical symptom experience, illness perceptions, adaptive coping and dignity. Clinicians have an opportunity to address several potentially remediable factors when DHD is identified.


This study was supported by St Christopher's Hospice. The authors are grateful to the patients and their families who generously gave their time to participate in this study. The authors thank the staff at St Christopher's for making recruitment possible.



  • Funding LR is supported by the COMPASS research collaborative and the European Commission's Sixth Framework Programme (contract no. LSHT-CT-2006-037777, EPCRC). WL is a MRC Clinical Training Fellow. AP is supported by St Christopher's Hospice. IJH is a NIHR Senior Investigator. MH is supported by the NIHR Biomedical Research Centre for Mental Health at the South London and Maudsley NHS Foundation Trust and Institute of Psychiatry, King's College London and is a NIHR Senior Investigator.

  • Competing interests None.

  • Ethics approval The study was approved by Institute of Psychiatry NHS Research Ethics Committee (06/Q0706/93).

  • Provenance and peer review Not commissioned; externally peer reviewed.