We received questionnaires for 3623 of the 6927 initial cases. From nonresponse analyses, we found that for 725 cases response was not possible owing to issues of access to the patient's medical file or patient identification. These cases were removed from the sample, and the response rate was 58.4%. Cases were weighted to be representative of all deaths in Flanders in 2007. The response rate in 2001 was 58.9%.

Results
The overall prevalence of continuous deep sedation until death increased significantly between 2001 and 2007 from 8.2% to 14.5%, and this increase occurred in all care settings, among both sexes, in all age groups, and in patients with various causes of death (Table 1). Opioids were used for sedation in 83%, often as the sole drug, especially in care homes. Sedation rarely lasted longer than 1 week. Artificial nutrition or hydration was withheld in most cases at home and in care homes, while 63% of sedated hospital patients received artificial nutrition and hydration until death. In one-fifth of all sedated patients (and 27% of sedated hospital patients), neither patient nor family had given consent for sedation. The patient had requested or consented to sedation in 53% of sedation cases at home, while the family had at least given consent in 78% of cases in care homes. There was a (co)intention to hasten death in 17% of cases, and the physician indicated the lack of alternatives to sedation in 82% (Table 2).

View this table: [in this window] [in a new window] [as a PowerPoint slide]   Table 1. Prevalence of Continuous Deep Sedation Between 2001 and 2007a

View this table: [in this window] [in a new window] [as a PowerPoint slide]   Table 2. Characteristics of Performing Continuous Deep Sedation (CDS) Until Death in 2007 by Place of Deatha

Comment
The increase of continuous deep sedation across all care settings and patient groups indicates its generally rising acceptance as a medical end-of-life practice. This is likely related to recent developments in the implementation and organization of palliative care in Belgium, partly instigated by a law on palliative care in 2002.⁸ Still, the increase in Flanders is striking and raises further questions. Other factors presumably influenced the increase. Although euthanasia is legal in Belgium since 2002, it is possible that some physicians and patients view continuous deep sedation as a psychologically and medically preferable alternative to euthanasia. The effects of institutional policy can also be taken into account, since many Belgian hospitals introduced additional safeguards to the legal requirements for euthanasia,⁹ possibly causing continuous deep sedation to be favored above euthanasia as a last resort decision. Or perhaps physicians are now simply more willing to report the performance of deep sedation because of the legalization of euthanasia. Lastly, it could also be that physicians prefer to sedate a patient until death rather than deal with a multitude of persistent but nonetheless nonrefractory symptoms.¹⁰ Further (qualitative) research is needed to investigate these hypotheses.

Clinical characteristics of continuous deep sedation differ between settings and show aberrations from internationally proposed guidelines and recommendations: opioids are frequently used as the sole drug (especially in care homes), patient or family consent is often lacking (especially in hospital), and sedation is often performed with an intention to hasten death (especially at home). Furthermore, in some cases alternatives to sedation had been possible for the treatment of symptoms. These results suggest that continuous deep sedation may sometimes be inadequately performed and ethically questionable and lead us to conclude that the formulation of an official clinical guideline is recommended for Belgium, most likely for other countries as well. The advantages of such a guideline have been demonstrated in the Netherlands, where recent research showed that the national guideline's instructions had been increasingly applied since its introduction in 2005.¹¹ It is, however, also clear from this research in the Netherlands that more is needed than mere implementation of a guideline to raise physicians' awareness for adequate sedation: training and knowledge dissemination—with specific focus points for each setting—are equally necessary.

In conclusion, the prevalence of continuous deep sedation increased considerably in Flanders, Belgium, between 2001 and 2007, across care settings and among various patient groups. More research is needed to assess the plausibility of the uttered explanatory hypotheses. Our findings further point to the need for the implementation of a clinical practice guideline in Belgium, since Belgian physicians do not seem to be familiar with existing international recommendations.

AUTHOR INFORMATION
Correspondence: Mr Chambaere, End-of-Life Care Research Group, Vrije Universiteit Brussel, Laarbeeklaan 103, 1090 Brussel (Jette), Belgium (kchambae{at}vub.ac.be).

Author Contributions: Study concept and design: Chambaere, Bilsen, Cohen, Onwuteaka-Philipsen, Mortier, and Deliens. Acquisition of data: Chambaere and Cohen. Analysis and interpretation of data: Chambaere, Bilsen, Cohen, Rietjens, Onwuteaka-Philipsen, Mortier, and Deliens. Drafting of the manuscript: Chambaere, Bilsen, and Cohen. Critical revision of the manuscript for important intellectual content: Chambaere, Bilsen, Cohen, Rietjens, Onwuteaka-Philipsen, Mortier, and Deliens. Statistical analysis: Chambaere and Cohen. Obtained funding: Cohen, Mortier, and Deliens. Study supervision: Bilsen, Cohen, Onwuteaka-Philipsen, Mortier, and Deliens.

Financial Disclosure: None reported.

Additional Contributions: The Flemish Agency for Care and Health provided the information in the death certificates and coordinated the mailing of the questionnaires; Wim De Brock (lawyer) processed and made anonymous the received questionnaires (financial compensation received from project researchers), and Geert Pousset, MA, assisted in conducting the data collection. We are deeply indebted to all physicians who provided the data for this study.