General practitioners' report of continuous deep sedation until death for patients dying at home: A descriptive study from Belgium

Abstract

Background: Palliative sedation is increasingly used at the end of life by general practitioners (GPs). Objectives: To study the characteristics of one type of palliative sedation, ‘continuous deep sedation until death’, for patients dying at home in Belgium. Methods: SENTI-MELC, a large-scale mortality follow-back study of a representative surveillance network of Belgian GPs was conducted in 2005–2006. Out of 415 non-sudden home deaths registered, we identified all 31 cases of continuous deep sedation until death as reported by the GPs. GPs were interviewed face-to-face about patient characteristics, the decision-making process and characteristics of each case. Results: 28 interviews were conducted (response rate 28/31). 19 patients had cancer. 19 patients suffered persistently and unbearably. Pain was the main indication for continuous deep sedation (15 cases). In 6 cases, the patient was competent but was not involved in decision making. Relatives and care providers were involved in 23 cases and 18 cases, respectively. Benzodiazepines were used in 21 cases. During sedation, 11/28 of patients awoke, mostly due to insufficient medication. In 13 cases, the GP partially or explicitly intended to hasten the patient's death.

Conclusion: Continuous deep sedation until death, as practiced by Belgian GPs, is in most cases used for patients with unbearable suffering. Competent patients are not always involved in decision making while in most cases, the patient's family is.

Key words::

• Continuous deep sedation until death
• palliative sedation
• end-of-life care
• decision making
• general practitioners

Introduction

In the last phase of life, patients wish to be comfortable and free from suffering (1). When symptoms remain uncontrollable despite optimal palliative care, palliative sedation, i.e. reducing the patient's consciousness and thus the awareness of suffering, can be used as an option of last resort (2,3). Palliative sedation can vary from mild to deep sedation and can be used intermittently or continuously. The most far-reaching type of palliative sedation is continuous deep sedation until death.

European studies showed that the frequency of continuous deep sedation until death varies between 2.5% of all deaths in 2001–2002 (Denmark) and 16.5% of all deaths in 2007–2008 (UK) (3,4). In Belgium in 2007, its incidence was estimated to be 14.5% of all deaths (5). This is a striking increase compared with 2001, when 8.2% of all deaths followed the use of continuous deep sedation until death. Studies from the Netherlands demonstrated a similar increase (6).

The practice of continuous deep sedation until death has become the subject of much medical, ethical and societal debate focusing on the indications for and adequate performance of this practice (7–9). Although in a few countries, nationwide (10,11) or local (12–14) guidelines have been developed for the use of continuous deep sedation, no such guidelines are currently available in Belgium. Recently, the European Association for Palliative Care (EAPC) published a framework of recommendations for the use of sedation in palliative care comparable with earlier published international recommendations (7,15). The bottom-line of these recommendations is that sedation can be considered when the patient is suffering unbearably from refractory symptoms. The disease should be advanced and without prospect of improvement, with death expected within hours or days. When possible, the patient and/or their family should be actively involved in the decision making and benzodiazepines should be the drugs of first choice. The purpose of continuous deep sedation until death should be symptom relief and not the hastening of death.

Although most studies on the use of continuous deep sedation are performed in high care settings like hospitals or palliative care units (3,5,6,16), sedation is also carried out at home by the general practitioner (GP) (6,16–19). This is a care context, with fewer available technological treatments and a stronger focus on the continuity of care, especially with respect to the management of end-of-life care (20). A recent study has shown that the increase in the use of continuous deep sedation was most noticeable in home deaths in Belgium, from 3.7% in 2001 to 9.8% in 2007 (5).

This study aims to achieve more insight into the clinical characteristics of the patient, the decision-making process and the characteristics of continuous deep sedation until death at home, from the perspective of the GPs. We also investigated whether this practice is in conformity with existing recommendations (7,15).

Methods

Study design

We used data from the mortality follow-back SENTI-MELC study, conducted in 2005–2006 to monitor end-of-life care in general practice in Belgium (21,22). Data were gathered by the Belgian Sentinel Network of General Practitioners, a reliable surveillance system that monitors the health of the entire population. It provides information, representative of all Belgian GPs, and covers 1.75% of the total Belgian population (23). The Ethical Review Board of the Brussels University Hospital of the Vrije Universiteit Brussels approved the protocol of the study (21).

Selection of study subjects

This study focuses on one specific subtype of palliative sedation that is considered to be the most far-reaching subtype, continuous deep sedation until death. In the quantitative registration study, GPs reported weekly about all deaths of patients in their practice using a standardized registration form. Out of 2690 registered (sudden and non-sudden) deaths, we identified all deceased patients who (1) were aged one year or older at time of death; (2) died ‘non-suddenly and not totally unexpectedly’ as judged by the GP; (3) died at home; and (4) were continuously and deeply sedated until death. GPs involved in the cases meeting these inclusion criteria, were selected for an interview study about the end-of-life care they had provided at home. Details of the registration study are published elsewhere (21). Out of a sample of 415 non-sudden home deaths, 272 cases were eligible for the interview study. In 52 cases, no valid interview was held so a quasi-control group of 192 non-sedated patients could be obtained, consisting of all non-sudden home deaths about which a valid interview with the GP could be held.

Interview procedure

The selected GPs were contacted by telephone by an independent party and invited for a face-to-face structured interview. The interviews were face-to-face and structured, combining quantitative with qualitative elements. They contained mostly closed questions (in which the interviewer could tick boxes), and some open-ended questions for which the interviewer was instructed to literally write down the answer of the GP. Furthermore, the interviewer could note additional information or remarkable statements given by the GP; these comments were often paraphrased by the interviewer.

The interview took place within two months after the patient's death and lasted on average one hour. Three interviews could not be held for methodological reasons: one interview could not be held because the maximum of one interview per GP every two months was reached, one case was identified as a case of continuous deep sedation only after closing the interview period and in the last case, the interviewed GP was not the attending physician in the last week of the patient's life.

Patient anonymity and GP confidentiality were preserved by the use of strict procedures. Before each interview, the interviewers gave the GPs a closed envelope containing patient information, prepared by an independent party, to make sure their answers related to the correct patient. GPs also used anonymous codes to refer to the patient in the registration form so patient names were never identifiable to the interviewers or other research group members. The interviews were not audio taped and in all data files, the identity of the GP was permanently deleted.

Measurements

For this study, we used the information obtained from questions about the patient's clinical characteristics, the decision-making process, the performance and the potential life-shortening effect of the sedation in each interview, from the GPs’ perspective. The patient's socio-demographic information was obtained from the registration form. Where possible, quantitative data was supplemented with additional (qualitative) paraphrases from the GPs.

Two researchers, (LA & JACR) then compared this information with five key recommendations made in different guidelines on palliative sedation (7,15), i.e. (1) whether the situation was medically hopeless; (2) whether the suffering was unbearable and persistent; (3) whether there was active involvement in the decision making of patient and/or family; (4) whether benzodiazepines were used; and (5) whether there was an intention to hasten death.

Outcomes and analysis

All closed-ended questions were analysed using SPSS 17.0 and described in tables. χ²-tests, Fisher's Exact tests and Mann–Whitney U-tests were used to identify differences between sedated and non-sedated patients. For the variable ‘treatment goal’, when there were multiple answers, we classified the additional goal under the main goal. The main goal of treatment always focused on the most ‘curative’ treatment. So ‘cure and life-prolonging’ was recoded into ‘cure’ and ‘life-prolonging and comfort/palliation’ into ‘life-prolonging.’

Answers to the open-ended questions and the additional paraphrases of GPs remarks were entered in Excel and were possible, coded into relevant categories by two researchers, (LA & JACR). The tables were then scrutinized to identify results that could be further explained and strengthened by the paraphrases. Wherever information was available, two researchers selected appropriate paraphrases (LA & JACR).

Results

Study population characteristics

220 registered non-sudden home deaths were included in the interview study. In respect of the 31 cases of continuous deep sedation until death at home, 28 interviews were held (response rate 28/31).

16 of 28 patients who were sedated continuously and deeply until death were men. 21 patients were aged <80 years. 19 had cancer, 2 had a cardiovascular disease, 2 had a pulmonary disease, 2 had a nervous system disease and 3 patients had another disease. 13 patients had been ill for more than six months; 24 had been diagnosed more than six months before death. The most frequently reported physical symptoms distressing the patient to a large extent in his/her last week of life were lack of energy (n = 11) and pain (n = 11). As for psychological symptoms, these were sadness (n = 13), worrying (n = 10) and agitation (n = 10).

Compared with non-sedated patients (n = 192), sedated patients were more often younger than 65 years, had been ill for less than one month, were less distressed by lack of energy, and were more distressed by pain (P-values <0.05).

Symptoms and prognosis

In all 28 cases in which the GP decided to sedate the patient, the GP judged that there was no chance of improvement of the patient's medical situation. GPs considered the suffering of the patient to be persistent and unbearable in 19 cases. Of the 9 remaining patients, 6 patients suffered ‘quite a lot’ to ‘very much’ from a combination of physical and psychological symptoms such as pain, dry mouth, lack of energy, difficulty breathing, drowsiness, feeling sad worrying and feeling nervous in their last week of life. In the 3 other cases, the physical and psychological symptoms were less present.

The number of patients receiving treatment aimed at comfort/palliation increased when they were nearing death: from 9 patients two to three months before death to 27 patients in the last week of life.

Decision-making process

In 2 cases, data on the decision-making process were missing (Figure 1). The decision to sedate the patient deeply was discussed with the patient in 12/26 cases and in 23 cases with the patient's relatives (Figure 1). 10 patients had requested continuous deep sedation according to the GP. In 16 cases, there was an explicit request by the patient's relatives e.g.: case 8: ‘the family asked to calm the patient down.’

Figure 1. Continuous deep sedation until death: the decision-making process (n = 28). ¹Reason(s) for not discussing: the proposed act was obviously the best for the patient (n = 2), discussion would have done more harm than good (n = 1), other (n = 8). ²Request for continuous deep sedation until death from the patient: n = 10. ³Request for continuous deep sedation until death from the patient's relatives: n = 7. ⁴Request for continuous deep sedation until death from the patient's relatives: n = 4. ⁵Request for continuous deep sedation until death from the patient's relatives: n = 4. ⁶Reason(s) for not discussing: patient was incompetent (n = 2), patient was unconscious or subcomatose (n = 4), patient had dementia (n = 3).

In 6 cases, the patient was competent, but not involved in the decision-making process. There was discussion with relatives in all but one of these cases. Reasons for not discussing the decision to sedate with a competent patient were in 3 cases the physician thought that it was better for the patient not to discuss, e.g. case 13: ‘the patient did not realize that the end was near, it was not easy for the patient to reconcile himself/herself to the situation’; that there was no explicit request from the patient, that the patient preferred to leave the decisions up to the physician, that the patient's wishes had been clear for years and that the physician had to react quickly to the distress of the patient.

Physicians discussed the decision to sedate the patient in 2 cases with a colleague physician, in 15 with a nurse and in 1 case with both. Pain was the most frequently reported indication for continuous deep sedation (n = 15) (Table I), e.g. case 3: ‘the pain was unmanageable.’

Table I. Characteristics of continuous deep sedation until death.

Variable	n = 28	%
Indications for continuous deep sedation until death^a
Pain	15	54
Agitation	7	25
Dyspnoea	4	14
Other^b	13	46
Types of drugs used for inducing continuous deep sedation until death
Benzodiazepines only	9	32
Benzodiazepines in combination with opioids and/or other drugs	12	43
Opioids only	4	14
Opioids and other drugs (excl. benzodiazepines)	3	11
Administration of artificial nutrition or hydration during continuous deep sedation until death
Limited amount of artificial nutrition and/or hydration^c	8	29
No artificial nutrition and hydration	20	71
Duration of continuous deep sedation until death
0–24 h	13	46
1–7 days	13	46
>1 week	1	4
Unknown	1	4
Patient woke up during continuous deep sedation until death
Yes	11	39
No	17	61
Reason(s) for waking up^a
Insufficient effect of medications	7	25
Possibility of contact with family	4	14
Other	3	11
Physicians’ appraisal of patient’s quality of dying
(Very) gentle death	20	71
Neutral	1	4
Not (very) gentle death	7	25

^aMultiple answers possible.

^bOther reported indications for deep sedation are: fear (of pain or death) n = 3 (11), (nocturnal) restlessness n = 2 (7), conventional treatment/preventive n = 1 (4), epilepsy n = 1 (4), existential reasons n = 1 (4), hopeless situation n = 1 (4), to do something n = 1 (4), nervousness n = 1 (4), patient was already somnolent n = 1 (4).

^cPurpose of limited administration of nutrition and/or hydration n = 7 (25)^a: Comfort of the patient n = 2 (7), symptom control n = 1 (4), infusion n = 1 (4), other n = 5 (18), unknown n = 3 (11).

Performance and evaluation of continuous deep sedation

Continuous deep sedation was induced in 21 of the 28 cases with benzodiazepines. In 8 cases, artificial nutrition and/or hydration were administered during sedation. The most frequently reported reason for limited administration of nutrition and/or hydration was patient comfort (n = 5), e.g. case 8: ‘to avoid a dry mouth.’

During continuous deep sedation, 11 out of 28 patients woke up, mostly due to insufficient medication (n = 8) or to communicate with the family (n = 3). Other explanations were ‘to administer food and fluid’ (case 4) and that the family did not want a deep sedation (case 8).

According to the GP's appraisal, 7 patients did not die a (very) gentle death: ‘it was a painful struggle; the patient could not really be sedated deeply, woke up and had oxygen deficiency’ (case 1). The patients who did not die a gentle death suffered more frequently persistently and unbearably, had a higher likelihood to wake up after the onset of the sedation, had less often discussed the use of sedation with the GP, and had less often a wish for euthanasia than patients who died a gentle death. However, these differences are not statistically significant.

Patients’ wishes for euthanasia and physicians’ intentions

8 patients had previously expressed a wish for euthanasia (Table II), ‘you are going to help me if it does not go well, aren't you? I want life to remain valuable, if not, then I want you to put an end to it’ (case 1). In 2 out of these 8 cases, the GP had partly the intention to hasten the patient's death. In 3 cases, the GP had the explicit intention to shorten the patient's life. 6 GPs perceived the use of continuous deep sedation to be similar to euthanasia (n = 2) or life-ending without explicit patient request (n = 4): ‘Depending on what is asked each time, you end up in a sort of euthanasia situation. You want it to go well’ (case 28).

Table II. The potential life-shortening effect of continuous deep sedation until death.

	Patient’s request for euthanasia
	Alln = 28^a (%)	Yesn = 8 (%)	Non = 19 (%)
Intention of hastening death
No intention to hasten death	14 (50)	3 (38)	11 (58)
Partly with the intention to hasten death	7 (25)	2 (25)	5 (26)
Explicit intention to hasten death	6 (21)	3 (38)	3 (16)
Use of sedation also perceived as life ending act by GPs^b
Perceived as euthanasia	2 (7)	2 (25)	0
Perceived as ending life without an explicit patient request	4 (14)	0	4 (21)

^aPatient's request for euthanasia: n = 1 (4) unknown.

^b22 cases (79) were not perceived as life-ending acts.

Comparison with recommendations

Finally, we compared our GPs’ practices as found in our study with five key recommendations for the use of continuous deep sedation: (1) the medical situation should be without prospect of improvement; (2) the suffering should be unbearable and persistent; (3) the decision to deep sedation should be discussed with a competent patient or an incompetent patient's family; (4) benzodiazepines should be administered; (5) the GP should have no intention to hasten the patient's death (7,15). In 5 cases, the GP acted in accordance with all five recommendations for the use of continuous deep sedation, in 9 cases four recommendations were met, in 13 patients, three criteria were met and in 1 case only two recommendations were met.

Discussion

Summary of main findings

Our study describes the practice of continuous deep sedation until death at home in Belgium in 2005–2006 from the perspective of the general practitioner. For all patients, the medical situation was without prospect and 19/28 patients suffered persistently and unbearably. 6 GPs had not consulted a competent patient but 23/26 had consulted the patient's family. Other care-providers were involved in 18/26 cases in the decision making. In three-quarters of the cases, benzodiazepines were used to induce sedation and in 8 cases artificial nutrition and/or hydration were administered. 11 of the 28 patients awoke after the onset of continuous deep sedation. 8 patients had previously expressed a wish for euthanasia. 13 out of 28 GPs partially or explicitly intended to hasten the patient's death.

Strengths and limitations of the study

Our study is the first to provide detailed insight into the practice of sedation at home in Belgium. A major strength is the case-specific information obtained from 28 sedation cases, selected from a large representative sample of non-sudden deaths and gathered by a reliable nationwide surveillance network of GPs. The research procedures were of high quality due to quality control measures. Furthermore, anonymity was preserved for the patient and confidentiality for the GP (22). However, our sample size is small and results have to be interpreted carefully. Possible recall and an interviewer bias could not be excluded entirely, but were limited as much as possible by weekly registrations and interviews within two months of the patient's death.

Indications

In our study, continuous deep sedation was used in 11/28 cases for patients younger than 65 years, which conforms with other studies (6,16,17,24,25). The most frequently mentioned indication for sedation in our study was pain. While this finding is in line with some other studies (8,16,17,19), a review of 17 studies, conducted in several settings, reported that a syndrome of delirium and agitation was the most frequently mentioned indication for sedative use (25). Most GPs have adequate knowledge of pain control for common problems in general practice, but might not always be fully aware of good management techniques for specific cancer pain (26). This may explain to some extent the ongoing presence of pain at the end of life in several of these cases.

In 9/28 cases, the GP considered the suffering of the patient not to be persistent and unbearable. However, suffering is a long-debated and complex concept. A Dutch study found that patients and physicians not always agree on the nature and extent of unbearable suffering (27). Unbearable suffering is considered to be subjective and thus difficult to assess by the physician (28). This points to the need for more research about the concept of suffering in relation to the practice of continuous deep sedation until death.

Decision-making process

While it is generally recommended to include competent patients in the decision making (11,15), 6 GPs in our study did not discuss the use of sedation with competent patients, only with their family members. GPs gave several reasons why they did so. Some thought that it was better for the patient, others indicated that they thought that the patient preferred to leave the decisions up to the physician, or that they felt that the patient's wishes had been clear for years to the GP. This is in line with studies showing that some physicians think that sharing full information is too heavy a burden for the patient (29) or that not all patients want detailed information about their situation (30). Nevertheless, adequate patient-centred communication entails timely discussion of how and to what extent the patient wishes to be involved in the decision making (31), and it is unclear to what extent the GPs in our study have explored patients’ wishes for involvement sufficiently and initiated advance care planning timely. Further research should address this in more detail. Further, in our study, GPs consulted other healthcare providers in 18/26 cases, possibly because they are less accessible in the home situation (16).

Medication

GPs in our study used benzodiazepines in three-quarters of all cases. It is commonly advised that benzodiazepines should be the drugs of first choice and that using only opioids may be counterproductive (9,25). Our findings are comparable with the literature, although two recent Dutch studies reported higher levels of benzodiazepines (6,18). Although the GPs in our study indicated that they had performed continuous deep sedation until death in all cases, 11/28 patients awoke after the onset of continuous deep sedation, something also found in a study in Japanese palliative care units (32). In 7 cases, this was due to insufficient use of medications, which suggests that GPs’ medication policy for the use of sedation could be further enhanced.

Artificial nutrition and hydration

Further, in recommendations, the administration of artificial nutrition and hydration during sedation is not encouraged unless the benefits outweigh the harm (7). In our study, artificial nutrition and hydration was forgone in 20/28 cases (71%). Other studies in home settings report frequencies from 51% (17) to 95% (6); studies in nursing-homes usually report higher frequencies and studies in hospitals lower frequencies (6,17). This points to the need for further evidence regarding the benefits and disadvantages of the administration of artificial nutrition and hydration in patients nearing death.

Continuous deep sedation or euthanasia?

Several recommendations and studies state that continuous deep sedation and euthanasia represent two different clinical practices with different techniques and intentions (7,10,12–15,24). However, our findings suggest that some GPs do not always strictly distinguish between continuous deep sedation and euthanasia and its respective techniques. In 13 cases, the GP indicated to have had a partly or explicit intention to hasten the patient's death. In 5 of these cases, the patient had requested euthanasia. Other studies mention similar intentions of physicians (5,16,24).

In 2 cases, the GP perceived the use of continuous deep sedation as euthanasia or in 4 cases as life-ending without explicit patient request. This shows that for some GPs, there is a lack of clarity about how continuous deep sedation for dying patients and euthanasia are recommended to be used, resulting in a grey zone between the two practices. However, whether GPs use sedation as an alternative to euthanasia as sometimes has been suggested (33), e.g. because they consider it better care, less distressing to perform or ‘easier’ to do because it lacks a formal and time-consuming procedure, deserves further study.

Implications for clinical practice and future research

Given the high incidence of continuous deep sedation until death in Belgium and the fact that several GPs did not act in accordance with one or more recommendations, our study showed the need for monitoring the quality of this end-of-life practice, as well as the potential for improvement.

Since the introduction of a nationwide palliative sedation guideline has actually improved the practice of sedation in the Netherlands, implementation of a nationwide guideline in Belgium seems to be advisable (8). Timely communication of the GP with the patient, the patient's family, as well as professional caregivers, might contribute to higher GP awareness of patient's wishes at the end of life. Furthermore, medical training and education about the differences and similarities between the use of sedation and euthanasia in end-of-life care and attention how they should be used in GPs practices seem needed.

In general, further in-depth and international research from multiple perspectives (physicians, nurses and relatives) is needed that adds to our understanding of this practice.

Acknowledgements

The authors thank all sentinel GPs for their contribution to this study; the ‘Vlaamse Liga tegen Kanker’ for the research grant ‘Emmanuel van der Schueren’ and project support and Koen Meeussen, MD (Vrije Universiteit Brussel) and Tinne Smets, MD (Vrije Universiteit Brussel) for the advice.

The research on which this paper is based is linked to a larger project, the ‘UNBIASED’ study (UK—Netherlands—Belgium InternAtional SEDation study), which is a collaboration between research teams in UK, Belgium and the Netherlands with funding from the Economic and Social Research Council (UK) (grant no: RES-062-23-2078), Research Foundation Flanders (BE), the Flemish Cancer League (BE), the Special Research Funds of the University of Ghent (BE), Netherlands Organisation for Scientific Research (NL) and the Netherlands Organisation for Health Research and Development (NL). For information, contact Professor Jane Seymour: e-mail: [email protected].

Financial resources

Support for the study came from the Research Council of the Vrije Universiteit Brussels in Belgium (project GOA27 2003-2007) and the Institute for the Promotion of Innovation by Science and Technology in Flanders as a Strategic Basic Research project (SBO) (contract SBO IWT 050158) (2006–2010), as part of the ‘Monitoring Quality of End-of-Life Care (MELC) Study’. The Belgian Sentinel Network of GPs is supported by the Flemish Ministry for Welfare, Public Health and Family and the Direction générale de la Santé du Ministère de la Communauté Française. This study was also realized by a research grant Emmanuel van der Schueren from the Vlaamse Liga tegen Kanker.

Ethics committee

The study was approved by the Ethical Review Board of the University Hospital of the Vrije Universiteit Brussels.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.