End-of-life practices in 11 German intensive care units

Background End-of-life care is common in German intensive care units (ICUs) but little is known about daily practice. Objectives To study the practice of end-of-life care. Methods Prospectively planned, secondary analysis comprising the German subset of the worldwide Ethicus‑2 Study (2015–2016) including consecutive ICU patients with limitation of life-sustaining therapy or who died. Results Among 1092 (13.7%) of 7966 patients from 11 multidisciplinary ICUs, 967 (88.6%) had treatment limitations, 92 (8.4%) died with failed CPR, and 33 (3%) with brain death. Among patients with treatment limitations, 22.3% (216/967) patients were discharged alive from the ICU. More patients had treatments withdrawn than withheld (556 [57.5%] vs. 411 [42.5%], p < 0.001). Patients with treatment limitations were older (median 73 years [interquartile range (IQR) 61–80] vs. 68 years [IQR 54–77]) and more had mental decision-making capacity (12.9 vs. 0.8%), advance directives (28.6 vs. 11.2%), and information about treatment wishes (82.7 vs 33.3%, all p < 0.001). Physicians reported discussing treatment limitations with patients with mental decision-making capacity and families (91.3 and 82.6%, respectively). Patient wishes were unknown in 41.3% of patients. The major reason for decision-making was unresponsiveness to maximal therapy (34.6%). Conclusions Treatment limitations are common, based on information about patients’ wishes and discussion between stakeholders, patients and families. However, our findings suggest that treatment preferences of nearly half the patients remain unknown which affects guidance for treatment decisions. Supplementary Information The online version of this article (10.1007/s00063-022-00961-1) contains supplementary material, which is available to authorized users.


Introduction
In Germany, the use of intensive care services during terminal hospitalizations has increased steadily in recent years, particularly in the older age groups [10].The culture of medicine has moved from a more paternalistic model to taking a patient's autonomy into increasing consideration in order to ensure that patient care is aligned with patient goals [11].While some consider end-of-life decision-making a growing challenge in German intensive care units (ICUs), others see a positive development towards humanized care.
End-of-life decisions are made when goals of care shift from curative to palliative care because of patient's treatment preferences or prolonged life-sustaining treatment that is no longer beneficial for the patient.Decision-making regarding end-of-life treatments, however, is complex and requires an active process of deliberation and communication among clinicians, the patient and family members [7].German intensivists perceive a considerable discrepancy between current endof-life practice and desired practice [28].The German Civil Code stipulates that the wishes of patients without decision-making capacity are to be determined hierarchically from (1) a written advance directive, (2) prior verbal statements about the preferred type, duration, and circumstances of the treatment in question, or (3) patient's general statements and values.However, in practice decision-making is fraught with difficulties due to uncertainty of prognosis and ambiguous patient wishes [20].The objective of this large multicenter study was to observe and characterize end-of-life practices in multidisciplinary German ICUs in 2015-2016 as a subgroup analysis of a world-wide study [3].

Setting
This is a prospectively planned, secondary analysis of the Ethicus-2 database specifically describing end-of-life practices in German ICUs.The Ethicus-2 study was a prospective, observational study of 199 ICUs in 36 countries evaluating consecutive adult ICU patients who died or had a limitation of life-sustaining treatment during a 6-month period [3].German centers were invited to participate through the German SepNet Critical Care Trials Group, a consortium of over 100 physicians and 50 academic and nonacademic hospitals in Germany.Institutional ethics committee approval, with a waiver of informed consent, was obtained from each participating center.The study was registered in the German Clinical Trials Register (DRKS-ID: DRKS00010044).

Patients
Consecutive adult patients admitted to participating ICUs who died or had any limitation of life-saving treatments over a 6month period were recruited in each ICU between September 1, 2015, and September 30, 2016 and were prospectively included.Patients were followed up until discharge from the ICU, death, or 2 months from the first decision to limit life-sustaining therapies.

Study procedure and data collection
Questionnaires and study material were translated into German.A data study form describing practice and communication of end-of-life decisions was completed for each patient by the senior intensivist in each participating ICU who was responsible for the respective end-oflife decisions.Mutually exclusive end-oflife categories were defined previously [3]: withholding (WH) and withdrawing (WD) treatment, shortening of the dying process (SDP), failed cardiopulmonary resuscitation (CPR), and brain death (BD) (Supplementary Table 1).Other data included patient age, gender, clinical characteristics, type and time of treatment limitations, whether discussed with patients or families, information about patient wishes (meaning any kind of statement about what the patient may want), concurrence with known patient wishes, and reasons as well as obstacles for treatment decisions.
To describe ethical practice, 12 variables were assessed post hoc as described previously [24].Items represent structured ethical practice, guidelines and legislation (end-of-life practice score [EPS]) ( [19]; Sup-plementary Table 2).Each positive answer received 1 point.The sum was operationalized as anICU-specific ethical practicescore with a range of 0 to 12 points.

Statistical analysis
Treatment limitations were categorized hierarchically according to the most active limitation (WD > WH).Since there was only one patient in the SDP category, this patient was included in the withdrawing treatment category.
For categorical variables, we report numbers and proportions within end-oflife groups.For continuous variables, we report medians and interquartile ranges.Differences between groups were tested with the Wilcoxon-Mann-Whitney test or the chi 2 (χ 2 ) test.All analyses are performed using the statistical software R [27].

Centers
Eleven ICUs participated in this study.Nine were mixed medical/surgical, one was medical, and one was a neurosurgical ICU.Nine centers were in academic hospitals (Appendix).
Treatment limitations at the time of the first decision to withhold or withdraw lifesupporting therapy are shown in .Fig.The proportion of advance directives was not different between patients with withdrawing or withholding treatments (.Table 2).

Practice of decision-making
Information about presumed treatment desires was available in 816 patients (78.3%) and more often for patients with treatment limitations than those without (785 [82.7%] vs 31 [33.3%],p < 0.0001; . Table 1).If patient desires were known, they were usually followed (628 [98%]), but they were actually known in only 641 (58.7%) of patients, suggesting uncertainty about the patient wishes in the remaining 451 patients (41.3%; .Table 1).In patients with treatment limitations, physicians obtained information mostly from the families (766 [93.9%]) and only in 143 cases (17.5%) from the patients themselves.
Supplemental table 3 shows the reasons, considerations, and difficulties of end-of-life decision-making.Physicians responsible for decision-making stated that the primary reason for limiting treatment was unresponsiveness to maximal therapy (335 [34.6%]).Patient or family requests were named in 140 (14.5%) and 27 (2.8%),respectively.Poor quality of life was the primary reason in 39 (4%) patients, while age was rarely the primary reason (4 [0.4%]).Primary considerations for decision-making were mostly based on the ethical principles of good medical practice (505 [52.3%]), best interest of the patient [25] (258 [26.7%]), or autonomous patient decision/advance directive (180 [18.6%]).Economic, religious, social, or legal concerns were not mentioned.Almost all physicians (945 [97.7%]) reported that they had no difficulty about either withholding or withdrawing treatment.

Time intervals
The median time interval between ICU admission and first treatment limitation was 2 [IQR 0-8] days, and between the first treatment limitation and death 1 [IQR 0-3] day.The time between ICU admission and first end-of-life decision was significantly shorter for patients with advance directives (median: 1 day 21 h [0 day 5 h to 5 days 21 h]) compared to patients without advance directives (median: 2 days 13 h [0 day 10 h to 9 days 23 h]; p < 0.05).In contrast, the time between first treatmentlimitationand deathwas significantly longer for patients with advance directives (median: 1 day 3 h [0 day 6 h to 3 days 1 h]) than for patients without advance directives (median: 0 day 16 h [0 day 2 h to 2 day 6 h]; p < 0.001).Decisions to withhold treatments (1 day [0 h to 4 days 20 h]) in the ICU were made in median 2 days earlier than decisions to withdraw treatments (3 days 2 h [15 h to 10 days 22 h]; p < 0.001; .Table 2).

Discussion
The results of this study showed that among 1092 patients in German ICUs who died or had limitations of life-sustaining therapy, decisions to limit life support preceded 9 out of 10 deaths, and only 8% of deaths occurred after full cardiopulmonary resuscitation (failed CPR).It is noteworthy that 22% of patients with a limitation of life-sustaining therapy were discharged alive from the ICU.
Treatment limitations occurred more often in older patients, in patients with advance directives, or decision-making capacity.Physicians commonly sought information about patient wishes, based their decision-making on shared decision-making with patient and families and ethical considerations.They reported having no difficulties with either withholding or withdrawing life support.Limitations also occurred more often in ICUs with a higher Ethical Practice score, suggesting the importance of palliative structures like local ethical standards and written practice guidelines to improve decisionmaking confidence for practitioners in the palliative situation.
Our findings illustrate the growing importance of palliative care in German ICUs.It has become an everyday occurrence, but there is lack of recognition of its importance.Out data point to some opportunities for improvement.Treatment limitations in German ICUs occur more frequently and failed CPR less frequently than before.Previously, a retrospective German study from 2002-2006 found that only 29% of deaths were preceded by end- of-life decisions and only 3.5% of patients survived end-of-life decisions [18].This is in contrast to 88.6% of decisions to withdraw and withhold and a hospital survival rate of 13.8% in the study we presented here.Compared to worldwide data from Ethicus-2, failed CPR occurred in a similar range in North American ICUs (8.5%) but less often than in ICUs in Australia/New Zealand (4.3%) or Northern Europe (3.7%)[3].
Palliative care in the ICU is increasingly provided through interdisciplinary team meetings, integration of palliative care specialists, ethics consultation and family conferences [22].Ethical principles and practices of palliative care in the ICU have been outlined by national medical societies [20].However, although most German intensivists practice palliative care, only a minority feels confident doing so [2].This may be due to a perceived lack of structures and standards which support the change from curative to palliative care, namely lack of interdisciplinary or ethics case reviews, palliative care training or standard operating procedures for endof-life care.Indeed, the EPS (end-of-life practice score) which assesses end-of-life protocols and palliative care consultations seems to suggest that treatment limitations occurred more often in ICUs with a higher EPS.However, this association needs to be treated with caution since more research is needed to understand the validity of this novel score.
Our data suggest a perceived gap between available information about patient wishes and unambiguous directives for decision-making.Physicians had information about patient wishes and discussed treatment limitations with families in over 80% of patients.They perceived that the patient will was followed in about 60% of patients.However, the question about the patient desires remained unanswered in about 40%, leading to the assumption that the patient's will remained unclear despite discussion with families.In our study, 27% of patients had advance directives, which are legally binding in Germany.This is similar to recent findings from the University of Hamburg [8] but lower than the prevalence in North America (49%) reported in the worldwide Ethicus-2 study [9].However, the prevalence of advance directives may not be high enough to support patient-oriented end-of-life decisions in most patients.Furthermore, the advance directives that are in use in Germany often contain unspecific wording which makes them unsuitable for many acute situations [15].More and better advance directives Medizinische Klinik -Intensivmedizin und Notfallmedizin 8 • 2023 669 are urgently needed in an ageing society.We speculate that the patient's will remained unclear in a considerable proportion of ICU patients.Given Germany's aging population and the increasing proportion of elderly patients who receive intensive care treatment at the end-of-life [10], avoiding inappropriate intensive care is a growing challenge.
If physicians did not speak with the family, this was often because the family was not available.Data also suggest that ICU physicians involve most patients and families in the decision-making.How-ever, the most common primary reasons for treatment limitations were unresponsiveness to maximal therapy, severity of disease, or underlying comorbidity.This suggests a more physician-centered approach with the intent to avoid nonbeneficial treatments regarding ethical principles such as good medical practice or best interest of the patient as primary consideration.This discrepancy can be explained with the uncertainty of patient wishes and family needs.According to a recent survey, 11% of families felt overwhelmed and wanted less participation in decision-making [12].It would be desirable to have regular meetings between treating physicians and nurses, patients and relatives and other doctors engaged in the patient's care like family doctors, so that the patient's wishes can be evaluated according to the disease course and achievable therapeutic options.
Our study cannot answer whether end-of-life decisions were timely or delayed.Physicians declined a delay due to disagreements between health care providers.The first treatment limitation occurred 2 days after ICU admission and death occurred 1 day after the first limitations.These intervals are comparable to findings from other studies [26,29].
End-of-life decision-making is considered a team effort, but in our study most decisions were reported to be initiated by physicians-except in one case, nurses were never reported to bring up the topic first.Nurse initiation was even lower than in the 1999-2000 (Ethicus-1) study [6] and declined across all European ICUs [5].On the other hand, two-thirds of decisions were discussed with nurses and nearly total agreement with decisions was reported.
Physicians in the present study reported not having difficulties with withholding or withdrawing therapy.This is notable because withholding therapy is sometimes considered to be psychologically easier and more passive than withdrawing treatment [17].However, a recent prospective multicenter study in 43 French ICUs in 2013 showed a similar pattern [16].Given that only a decade ago in Germany, the issue of limiting life support was discussed in a controversial manner with concerns that limiting life support could be illegal [4], this finding indicates that the practice of limiting nonbeneficial treatment has become more accepted and reflective of national recommendations [13].The answers were given by senior physicians in the present study.Thus, this answer should not be transferred on younger colleagues, who should not be left alone with these decisions unless adequately trained.
In France, Quenot et al. found similar physicians' perceptions of nonbeneficial therapy, including exhaustion of therapeutic options and terminal status of chronic disease.In French ICUs, physicians also addressed age as a factor which determines nonbeneficence [21], whereas physicians in our study rarely reported age to be the primary reason for an end-of-life decision.Physicians mostly stated ethical principles such as good medical practice or best interest of the patient as primary consideration in decision making, and not economic or social obligations.
We can only speculate on the surprisingly high ICU and hospital survival rate (22% and 16%, respectively) after treatment limitation.A similar phenomenon in the Ethicus-2 comparison study in European ICUs was discussed as result of decisions made before or during hospitalization due to the patients' wishes [24].A study in Finnish ICUs found that one in four patients survived 1 year, depending on housing type, prehospital fitness, and the need of postoperative care in an ICU [1].
Our study has strengths and limitations.To our knowledge, this is the largest prospective and patient-based study of end-of-life decisions in the ICU in Germany.Data were collected centrally, submitted to quality controls, and used in previous and international studies [23,24], thus, enabling comparison.The study also has limitations.Participating ICUs were predominantly academic and self-selected on account of their ethical interest which introduces selection bias.Thus, findings may not be generalizable.Moreover, the collected data did not elicit the perceptions of nurses.This may have introduced bias since nurses perceive end-of-life decisionmaking more negatively than physicians [14].Answers were given by senior physicians; thus, they may not reflect the uncertainty experienced by younger physicians.Moreover, self-reported answers to ethical questions underlie social desirability bias.We left the classification of whether decision discussions were shared to the responsible physician.Thus, we cannot rule out misclassification bias.Finally, our findings cannot be extrapolated to other countries where there is less limitation of life-supporting therapies due to different cultures, healthcare systems, and population demographics.

Conclusions
In German ICUs, decisions to limit life support precede nine out of 10 deaths, and 22% of patients with a limitation of lifesustaining therapy survive the ICU.Physicians often seek information about patient wishes, base their decision-making on discussions about prognosis and ethical considerations, and have no difficulties with either withholding or withdrawing life support.However, our findings suggest that treatment preferences of nearly half of the patients remain unknown and fail to guide treatment decisions.Further work should investigate structured approaches to implement palliative care, validate the ethical practice score, and explore timing and nature of discussions.More efforts are needed to increase the appropriateness and prevalence of advance directives.
TrialsGroup-Ethicus2 study group declare that they have no competing interests.C. S. Hartog reports funding grants paid to her institution from the Innovations Funds of the German Federal Joint Committee (FKZ 01VSF19020, FKZ 01VSF17010) and the European Society of Intensive Care Medicine (ESICM; 2018-2021).
For this article no studies with human participants or animals were performed by any of the authors.All studies mentioned were in accordance with the ethical standards indicated in each case.
Open Access.This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.24.Sprung

Fig. 2 8
Fig. 2 8 Treatment limitations at the time of first decision to withhold or withdraw life-supporting therapy.DNR Do not resuscitate; IV intravenous, DNR Withhold means to withhold cardiopulmononary resuscitation.DNR Withdraw means that the order for DNR is removed

Table 1
Study population c number of patients with available data Medizinische Klinik -Intensivmedizin und Notfallmedizin 8 • 2023 667

Table 2
Patients by end-of-life categories Values were calculated for the comparison of patients with WD and with only WH decisions CPR cardiopulmonary resuscitation, WD and WH withdrawal and withholding of life-sustaining therapy, respectively, IQR interquartile range, ICU Intensive care unit

Table 3
Practice of decision making for patients with treatment limitations (n =967) DNR do not resuscitate, CPR cardiopulmonary resuscitation, ICU intensive care unit CL, Ricou B, Hartog CS et al (2019) Changes in end-of-life practices in European intensive care units from 1999 to 2016.JAMA322:1692-1704 25.Sprung CL, Truog RD, Curtis JR et al (2014) Seeking worldwide professional consensus on the principles of end-of-life care for the critically ill: the Welpicus study.Am J Respir Crit Care Med 190:855-866 26.Stachulski F, Siegerink B, Bösel J (2021) Dying in the neurointensive care unit after withdrawal of life-sustaining therapy: associations of advance directives and health-care proxies with timing and treatment intensity.J Intensive Care Med 36:451-458 27.Team RC (2020) R: a language and environment for statistical computing.R Foundation for Statistical Computing, Vienna (https://www.R-project.org)28.WeissM, MichalsenA, ToenjesAetal(2017)End-oflife perceptions among physicians in intensive care units managed by anesthesiologists in Germany: a survey about structure, current implementation and deficits.BMC Anesthesiol 17:93 29.Wunsch H, Harrison DA, Harvey S et al (2005) Endof-life decisions: a cohort study of the withdrawal of all active treatment in intensive care units in the United Kingdom.Intensive Care Med 31:823-831Hier steht eine Anzeige.