rss
J Med Ethics 40:458-462 doi:10.1136/medethics-2013-101459
  • Clinical ethics
  • Paper

Estimating the potential life-shortening effect of continuous sedation until death: a comparison between two approaches

  1. Agnes van der Heide1
  1. 1Department of Public Health, Erasmus MC, Rotterdam, The Netherlands
  2. 2Department of Anesthesiology, EMGO+ Institute for Health and Care Research, VU University Medical Center, and Hospice Kuria, Amsterdam, The Netherlands
  3. 3Julius Centre for Health Sciences and Primary Care, University Medical Center, Utrecht, The Netherlands
  1. Correspondence to Sophie M Bruinsma, Department of Public Health, Erasmus MC, P.O. Box 2040, Rotterdam 3000 CA, The Netherlands; s.m.bruinsma{at}erasmusmc.nl
  • Received 8 March 2013
  • Revised 27 May 2013
  • Accepted 6 June 2013
  • Published Online First 11 July 2013

Abstract

Context In some cases, physicians estimate that continuous sedation until death may have a life-shortening effect. The accuracy of these estimations can be questioned.

Aim The aim of this study is to compare two approaches to estimate the potential life-shortening effect of continuous sedation until death.

Methods In 2008, 370 Dutch physicians filled out a questionnaire and reported on their last patient who received continuous sedation until death. The potential life-shortening effect of continuous sedation was estimated through a direct approach (question: Did continuous sedation, according to your estimation, hasten the patient’s death? If yes: by how much time?) and an indirect approach (estimated life expectancy minus duration of sedation). The intrarater agreement between both approaches was determined with a weighted κ.

Results According to the direct approach, sedation might have had a life-shortening effect in 51% of the cases and according to the indirect approach in 84%. The intrarater agreement between both approaches was fair (weighted κ=0.38). In 10% of all cases, the direct approach yielded higher estimates of the extent to which life had been shortened; in 58% of the cases, the indirect approach yielded higher estimates.

Conclusions The results show a discrepancy between different approaches to estimate the potential life-shortening effect of continuous sedation until death.

Introduction

A significant minority of dying people experience serious symptoms that are unresponsive to conventional therapies. In such circumstances, palliative sedation may be considered. Palliative sedation is the deliberate lowering of a patient's level of consciousness in the last stages of life.1 Sedation can be used intermittently or continuously until death, and the degree of sedation necessary to relieve suffering may vary from superficial to deep.2 Continuous sedation is most frequently used in patients suffering from physical symptoms such as delirium, dyspnoea, pain and nausea.3 ,4 The moral status of continuous sedation until death has been the subject of fierce ethical debate.5 This debate mostly focuses on whether continuous sedation may hasten death. It is stated in guidelines that physicians should use sedation with the intention to relieve suffering, and not with the intention to shorten the patient's life.1 ,6 Further, to preclude a potential life-shortening effect, it is recommended to restrict the use of continuous sedation to patients with an estimated life expectancy of at most 2 weeks.1 ,7 ,8 When the patient's life expectancy is within this limit and when sedatives are properly dosed, continuous sedation until death has presumably no life-shortening effect.1 ,3 ,8

The underlying assumption in guidelines is that physicians can estimate a patient's life expectancy with sufficient accuracy. However, estimating life expectancy of patients with advanced disease is known to be very difficult.7 ,9 ,10 Physicians tend to overestimate survival: it has been shown that survival of patients is typically 30% shorter than predicted by physicians, but that the accuracy of physicians’ predictions increases when death approaches.10

Although it is assumed in guidelines that continuous sedation until death has no life-shortening effect when used proportionally,1 physicians may have a different perspective. In a large-scale nationwide follow-up study, performed in 2005 in the Netherlands, physicians estimated that continuous sedation until death might have had a life-shortening effect in 26% of the cases.3 In 20% of these cases, it was estimated that sedation had shortened the patient's life by less than 1 week, in 4% of the cases by less than 1 month and in 2% by more than 1 month. These estimates should be interpreted cautiously. On the one hand, physicians might be cautious when they are directly asked to estimate the potential life-shortening effect of continuous sedation until death. The sensitivity of the topic might have led to socially desirable answers and physicians might have been reluctant to state their true opinions.11 For instance, because they might feel that a life-shortening effect is undesirable and morally complex, and because acknowledging such an effect may suggest that a patient's death was actively hastened. On the other hand, physicians are known to be inclined to overestimate the life expectancy of patients with advanced disease.9 It has been shown that survival of patients is typically 30% shorter than predicted by physicians, but that the accuracy of physicians’ predictions increases when death approaches.9 The fact that physicians commonly have difficulty in estimating life expectancy and often are inclined to overestimation might also hold true for the life-shortening effects of sedative medication in patients with a limited life expectancy.12 Obviously, estimating the true life-shortening effect of continuous sedation would require an experimental study, which is, however, not an option for this patient group.

The aim of this study therefore is to get insight in the accuracy of estimates of the life-shortening effect of continuous sedation until death by comparing two different approaches. We compare a direct approach, where we ask physicians to estimate the life-shortening effect of continuous sedation until death, and an indirect approach, where we ask the physicians to estimate the patient's life expectancy and relate that to the duration of the sedation.

Methods

Study design and data collection

A secondary analysis was performed of data that were collected among physicians in a study that evaluated the practice of palliative sedation after the introduction of the Royal Dutch Medical Association guideline.13 ,14 Data collection took place between February 2008 and September 2008. For this study, a structured questionnaire was sent to a random sample of 1580 physicians: 1128 in the north-western and south-western regions of the Netherlands (general practice, n=466; nursing home, n=195; and hospital, n=467) and 452 general practitioners in the north-eastern region.13 Physicians were asked to report on the last patient for whom they had been responsible for providing continuous sedation until death. One of the issues addressed was the potential life-shortening effect of continuous sedation until death. Physicians were asked whether or not the use of continuous sedation until death might have had a life-shortening effect (‘Did continuous sedation, in your estimation, hasten the patients’ death?’); to estimate the patient's life expectancy at the start of sedation (‘What, in your estimation, was the patient's life expectancy, at the time continuous sedation was started?’); and to assess the actual duration of the sedation (question: ‘How long after the start of continuous sedation did the patient die?’) (table 1).

Analysis

We used a direct and an indirect approach to estimate the potential life-shortening effect of continuous sedation until death. The direct approach was based on the direct question about the life-shortening effect of sedation. Response categories ranged between ‘no life-shortening’ and ‘more than one month shortened’. With the indirect approach, the duration of the sedation until death was subtracted from the estimated life expectancy at the start of sedation. The estimated life expectancy consisted of categories varying from ‘less than one day’ to ‘more than one month’. The question about the duration of the sedation was answered in weeks, days and/or hours. To make statistical calculations possible, life expectancy was recoded into hours by taking the middle of categories: for example, ‘less than one day’ was recoded as ‘12 h’. The difference between the recoded estimated life expectancy and the duration of the sedation was calculated (table 1).

Table 1

Two approaches of estimating the life-shortening effect of continuous sedation until death

A total of 606 physicians (38%) filled out the questionnaire. Response rates were 43% for general practitioners, 50% for nursing homes physicians and 24% for clinical specialists. Of the responding physicians, 370 (61%) reported about their last case.11 Physicians were on average 49-years-old (range 28–64). A majority of the physicians were men (64%). Of all physicians, 15% worked in a hospital, 17% in a nursing home or hospice and 68% were general practitioners. Physicians had on average 19 years work experience (range 1–38). The patients physicians reported on were on average 70 years old (range 3–99) at the time of their death. The majority of the patients (71%) had cancer as their main diagnosis. The majority of the patients were male (52%) (table 2).

Table 2

Characteristics of responding physicians and patients and sedation characteristics (N=370)

Physicians were excluded from the analyses if the answer to one of the questions was ‘unknown’, ‘don't know’ or uninterpretable (more than one answer). The intra-rater agreement between the direct and indirect approaches was determined by calculating a weighted κ: quadratic weights were used to take into account the size of differences between both approaches. The scores were interpreted using the Landis and Koch criteria: <0.00=poor agreement, 0.00–0.20=slight agreement, 0.21–0.40=fair agreement, 0.41–0.60=moderate agreement, 0.61–0.80=substantial agreement and 0.81–1.00=almost perfect agreement.13 We also calculated weighted κ values for two subgroups to see whether the level of agreement was different for patients with a relatively long estimated life expectancy and patients with a relatively short estimated life expectancy. A cut-off point of 1 week was used for this analysis, based on the median value of the estimated life expectancy. The significance level was set at 5%. For the analysis, Excel and SPSS V.22.0 were used.

Results

In 269 cases (74%), physicians estimated that the patient had a life expectancy of less than 1 week, in 84 cases (23%) between 1 and 2 weeks, and in 10 cases (3%) more than 2 weeks at the start of sedation. The duration of the sedation varied between less than 1 day, and 1 and 2 weeks (table 2).

When asked directly, 148 physicians (41%) stated that continuous sedation did not have a life-shortening effect, 151 physicians (42%) stated that there might have been a life-shortening effect and 64 physicians (17%) did not know. Using the indirect approach, the use of sedation might have had a life-shortening effect in 84% and no such effect in 16% of the cases (table 3).

Table 3

Estimated life-shortening effect (direct approach), life expectancy, duration of sedation and the estimated life-shortening effect (indirect approach) (N=370)

After excluding cases where the answer to one of the questions was ‘don't know’, uninterpretable or missing, 289 cases remained available for analysis of the level of agreement between the direct and indirect approaches of assessing the estimated life-shortening effect of sedation. When asked directly, 147 physicians (51%) estimated a potential life-shortening effect of continuous sedation (table 4). In 13% of these cases, it was estimated that sedation might have shortened the patient's life by less than 1 day; in 28% of the cases, between 1 day and 1 week; in 8% between 1 and 2 weeks; and in 2% by more than 1 month. The duration of the sedation was shorter than the estimated life expectancy at the start of sedation according to 242 physicians (84%). So, following this indirect approach, it was estimated that the use of sedation might have had a life-shortening effect in 84% of the cases (table 4). In 22% of these cases, it was estimated that sedation might have shortened the patient's life by less than 1 day; in 43% of the cases, between 1 day and 1 week; in 17% between 1 and 2 weeks; and in 2% by more than 1 month. In 10% of these cases, the direct approach yielded higher estimates of the extent to which life had been shortened than the indirect approach; in 58% of the cases, the indirect approach yielded higher estimates and there was no difference in 31% of the cases. The level of agreement between the direct and indirect approaches as assessed by weighted κ was 0.38, indicating ‘fair’ agreement (table 4). Further analysis showed that the level of agreement was somewhat higher for patients with an estimated life expectancy of less than a week (weighted κ=0.26) as compared with patients with a life expectancy of more than 1 week (weighted κ=0.10).

Table 4

The estimated life-shortening effect of continuous sedation (direct and indirect approach) (n=289*†)

Discussion

In this study, we found that in 51% of the cases Dutch physicians estimate that the use of continuous sedation until death might have had a life-shortening effect when they are directly asked about such an effect. In contrast, on the basis of physicians’ estimations of patients’ life expectancy at the start of sedation and the duration of sedation until death, it can be estimated that such an effect occurs in 84% of cases. The finding of this substantial discrepancy between the two approaches to estimate the life-shortening effect of continuous sedation until death confirms the difficulty of predicting the life expectancy of patients with advanced disease, and of estimating the potential life-shortening effect of end-of-life interventions. However, our finding that there is ‘fair’ agreement between both approaches suggests that this discrepancy is not merely the result of random inaccuracy. In general, physicians might be relatively cautious when they are directly asked to estimate the potential life-shortening effect of continuous sedation until death, because they feel that such an effect is undesirable and morally complex, and because admitting such an effect may suggest that a patient's death was actively hastened. On the other hand, physicians are known to be inclined to overestimate the life expectancy of patients with advanced disease,10 which may extend to patients who are provided with continuous sedation until death. Such inclination could explain the high proportion of cases in which sedation might have shortened life, when the estimate is based on our ‘indirect’ approach.

Our finding that the agreement between the two approaches to estimate the potential life-shortening effect of continuous sedation was higher for patients with a life expectancy of less than 1 week confirms findings from previous studies that estimations of life expectancy become more accurate when death approaches.10 An implication could be, if physicians indeed consistently overestimate their patients’ life expectancy, that in some cases where sedation would be a beneficial intervention, it is started too late or not at all, which would involve unnecessary suffering.

It is striking that, according to both approaches, physicians often think that continuous sedation until death can have a (mostly limited) life-shortening effect. Several empirical studies have suggested that sedation as used in clinical practice has no significant life-shortening effect.15–20 It can, in either of the approaches, be questioned if physicians tend to overestimate the life-shortening potential of sedation, but our data allow no firm conclusion here. The importance of a potential life-shortening effect of continuous sedation until death can also be questioned. The Royal Dutch Medical Association guideline argues that the life expectancy of a maximum of 2 weeks is conditional for palliative sedation.1 The authors of the European Association of Palliative Care framework even stated that sedation should only be contemplated if the patient is hours or days from death.6 However, one could hold the position that palliation with life-shortening side effects is morally justified, as long as proportionately consequential reasons are present. The indication for sedation originates from the presence of one or more refractory symptoms that lead to severe and unbearable suffering. In such circumstances, the life expectancy criterion can be weighed against the severity of refractory symptoms.21 If the benefit of palliation outweighs the harm of an earlier death, if there are no other alternatives and if dosages are titrated according to the patient's need, palliative sedation may be indicated, even if death is not imminent.22 In some cases where the conditions for continuous sedation until death are not met, brief or intermittent sedation may be a possible alternative.1

Our study had some limitations. First, the physicians in our study provided information retrospectively. The data could therefore be influenced by recall bias. Further, we asked about the physician's most recent case, which may not always represent physicians’ usual practices or approaches.13 Third, the risk that physicians matched their answers and estimations in the questionnaire cannot be ruled out. The true differences between the two approaches might therefore be larger than suggested in our study. To improve accuracy, it might useful to include patients prospectively in a future study: life expectancy could then be estimated before the start of sedation, and the duration of sedation could be timed. For this study, a secondary analysis was performed of data that were collected in 2008. Therefore, this approach could not be followed in the present study. It can be concluded that estimating the life expectancy of patients who are provided with continuous sedation until death is difficult. Recommendations in guidelines that continuous sedation until death should only be used for patients with a life expectancy of less than 1 or 2 weeks may therefore be difficult to translate to clinical practice. In research, the type of question that is used to estimate the life-shortening effect of sedation and other end-of-life interventions has to be taken into account when interpreting the results. Based on this study, we cannot conclude whether a direct question or a more veiled approach is preferable.

Acknowledgments

A secondary analysis was performed, of the data that was collected among physicians in a study that evaluated the practice of palliative sedation after the introduction of the Royal Dutch Medical Association guideline (the Amsterdam Rotterdam Sedation project). We would like to thank all the respondents for filling out the questionnaires.

Footnotes

  • Contributors SB analysed the data. The results were discussed with JR and AH. SB wrote the paper, which was critically read by all the authors. All authors approved the final version. AH is the guarantor.

  • Funding The original study was funded by the Netherlands Organisation for Health Research and Development (ZonMw) and received additional funding from Sint Laurens Fonds Rotterdam and Stichting Palliatieve Zorg Dirksland-Calando.

  • Competing interests None.

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

References

Free sample
This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of JME.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.