The statistical significance of differences between physicians and nurses was assessed with ² and Kruskal-Wallis tests. Logistic or linear regression analysis was performed to adjust for setting and working experience. For all tests, P < .05 was considered statistically significant.

Results
The questionnaire was completed by 606 physicians (response rate, 38%) and 278 nurses (response rate, 48%). Cases were reported by 370 physicians (61%), mainly in general practice, and 185 nurses (67%), mainly hospital nurses. Most patients had cancer (Table). The severe symptoms that were most commonly reported before the start of continuous sedation by physicians and nurses were fatigue, pain, and longing for death. Nurses reported anxiety significantly more often than physicians. The decisive indications for starting continuous sedation that were most reported by both physicians and nurses were dyspnea and pain (data not shown). Nurses specified pain as the decisive indication more often than physicians, who reported physical exhaustion and delirium more often than nurses. Psychological exhaustion and existential suffering were reported mainly when there was more than 1 decisive indication.

View this table: [in this window] [in a new window] [as a PowerPoint slide]   Table. Respondents, Patient Characteristics, Decision Making, and Effect of Continuous Sedationa

In most cases, life expectancy at the start of continuous sedation was estimated to be less than 2 weeks. In nurses' cases, competent patients and relatives had been involved in the decision making significantly more often. While physicians reported more often than nurses that they felt they were put under pressure to start continuous sedation, mostly by patients and relatives, they less often reported that continuous sedation had been provided with the full or partial intention of hastening the patient's death. In most cases, respondents reported that symptoms had been properly relieved, that relatives had seemed satisfied with the course of continuous sedation, and that the quality of dying had been good.

Comment
Pain, dyspnea, and delirium are commonly reported indications for palliative sedation.^5-6 Our study confirmed these findings and pointed to physical exhaustion as another decisive indication. This is consistent with fatigue being the most common severe symptom before the start of the sedation, apparently to the extent that it can become refractory to treatment and thus become an indication for continuous sedation until death. Psychological exhaustion and existential suffering were also mentioned as indications, mostly in combination with physical symptoms. The use of continuous sedation until death, therefore, often follows from a clinical picture in which a combination of physical and nonphysical symptoms results in a refractory state.

Nurses more often than physicians reported that patients and relatives were involved in decision making, they less often felt pressure from patients or relatives to start continuous sedation, and they more often thought that the physicians' intention for using sedation was to hasten the patient's death.

Although these physicians and nurses did not report on the same cases, we think we can interpret these differences because the practice on which they reported seems to be comparable. First, virtually all these respondents' cases involved both a physician and a nurse. Second, our respondents were selected in the same regions of the Netherlands and in the same period. Lastly, the patients' characteristics did not differ notably, and physicians and nurses assessed the frequency of severe symptoms similarly.

We therefore think that the differences in decision making, feelings of pressure, and intention might at least be partly related to the different roles of physicians and nurses in continuous sedation practice.

First, nurses usually have more frequent contacts with patients. Such frequent contacts might explain why more nurses indicated that patients and relatives were actively involved in the decision making. Nurses' greater involvement in patient handling might explain why more nurses mentioned pain as a decisive indication for starting continuous sedation and why they also reported more feelings of anxiety. Both pain and anxiety might be more easily recognized during daily care and nursing activities.

Second, the differences might also reflect the physicians' and nurses' different responsibilities. Because physicians are responsible for the decision making on the use of continuous palliative sedation, patients and relatives might put particular pressure on them to start sedation.⁷ In studies of euthanasia, physicians have also reported to feel subjected to pressure.⁸ Thus, taking account of the varying perspectives and emotions of all those involved in end-of-life decision making may be complex.

Finally, it has been shown elsewhere that nurses sometimes worry about the potential use of continuous sedation for accelerating death.⁹ Taken together with our finding that more nurses than physicians think that continuous sedation was used with the intention to hasten death, this may reflect that physicians are not always clear about their intentions. This stresses the paramount importance of proper communication between physicians and nurses.¹⁰

In conclusion, while continuous sedation until death is usually provided because of severe and refractory physical symptoms, nonphysical symptoms may also contribute to the clinical picture. At several points, physicians and nurses experience the decision-making process differently. End-of-life care would benefit from fuller communication between physicians and nurses about all relevant aspects of the patients' situation and the care provided.

AUTHOR INFORMATION
Correspondence: Dr Swart, Department of Public Health, Erasmus Medical Center, PO Box 2040, 3000 CA Rotterdam, the Netherlands (s.swart{at}erasmusmc.nl).

Author Contributions: Dr Swart and Mr Brinkkemper contributed equally to this work and share first authorship. Study concept and design: Swart, Rietjens, Blanker, van Zuylen, Ribbe, Zuurmond, van der Heide, and Perez. Acquisition of data: Swart, Brinkkemper, Rietjens, and Perez. Analysis and interpretation of data: Swart, Brinkkemper, Rietjens, Blanker, van Zuylen, van der Heide, and Perez. Drafting of the manuscript: Swart, Brinkkemper, Rietjens, and Perez. Critical revision of the manuscript for important intellectual content: Swart, Rietjens, Blanker, van Zuylen, Ribbe, Zuurmond, van der Heide, and Perez. Statistical analysis: Brinkkemper, Rietjens, Blanker, and Perez. Obtained funding: Swart, Rietjens, Zuurmond, van der Heide, and Perez. Administrative, technical, and material support: Swart, Brinkkemper, Rietjens, and Perez. Study supervision: Rietjens, van Zuylen, Ribbe, Zuurmond, van der Heide, and Perez. All authors formed part of the project group for this research and approved this article.

Financial Disclosure: None reported.

Funding/Support: This study was supported by a grant from ZonMw and received additional funding of Sint Laurens Fonds Rotterdam and Stichting Palliatieve Zorg Dirksland Calando.

Additional Contributions: Wijnanda van Burg-Verhage, MD, worked on the pilot. Paul van der Maas, PhD, MD, and Johannes van Delden, PhD, MD, provided critical comments on earlier versions of this article. Furthermore, Anneke Tooten, MSc, Claudette Promes, RN, Jaap Talsma, MD, and the Comprehensive Cancer Centres of Amsterdam and Rotterdam assisted with the data collection, and we thank all the respondents for filling out the questionnaires.