The primary objective of the current survey study is to obtain detailed information about European intensive care unit (ICU) clinicians' awareness of palliative care and relevant, local end-of-life (EoL) legislation.

INTRODUCTION In countries with patient/family autonomy-oriented end-of-life (EOL) legislation, published data indicate variation of healthcare professionals' awareness of the specific legal provisions and/or regulations enabling limitation of life-sustaining treatments. Furthermore, communication difficulties and/or uncertainties about the validity of previously recorded patient preferences (e.g. in the form of advance directives) may result in heterogenous application of pertinent EOL laws/regulations. This may include cases of poor compliance with, or even opposition to new laws.

The absence of accurate and in-depth EOL legal knowledge of clinicians may be associated with perceptions that applicable laws/regulations are either too restrictive or excessively "liberal" for the patient. Furthermore, critical care physicians and/or nurses may lack specific knowledge about palliative care. Accordingly, physicians/nurses may also not be aware of the palliative nature of their everyday practice or have a clear understanding of primary and specialist palliative care.

Given the above-mentioned knowledge gaps of healthcare professionals caring for critically ill patients, the investigators propose to conduct a Europewide, descriptive survey for intensive care unit (ICU) clinicians, primarily aimed at addressing the following questions: 1) "How much of ICU everyday practice is palliative care and how is it practiced?"; and 2) "How do clinicians perceive the medico-legal framework around therapy limitations, patient autonomy and prognostication discussions with patients and families?" In addition, this survey will enable the investigators to comparatively assess the perceived extent of ICU palliative care practice in the presence vs. absence of EOL legislation.

METHODS Study participants and protocol approval The target population of this descriptive, open survey will include ICU physicians and nurses working in at least 100 ICUs across 50 European countries and Israel. Each one of participating ICUs will be represented by one or more healthcare professionals, and the target convenience-sample size will amount to at least 500 respondents (see also below-provided statistical analysis plan). This survey study is part of the European Union-funded Enhancing Palliative Care in ICU (EPIC) project (Proposal number: 101137221; HORIZON-HLTH-2023-DISEASE-03-01). The study protocol has been approved by the Scientific Council of Evaggelismos Hospital, Athens, Greece.

Informed consent and data protection A similar methodology has been recently described. According to the Helsinki declaration, participation in research requires informed consent of the participant. According to Regulation 679/2016 {or General Data Protection regulation (GDPR) of the European Parliament and of the Council, "consent" of the data subject means any freely given, specific, informed and unambiguous indication of the data subject's wishes by which he or she, by a statement or by a clear affirmative action, signifies agreement to the processing of personal data relating to him or her." Participation in and completion of the current survey means that the participant accepts to share their knowledge on palliative care practice, EOL legislation and EOL practice in their country. The estimated time for survey completion is between 30 and 60 min.

The procedure for "pseudonymisation" is detailed below. A set of 20 consecutive random numbers (range, 10-29) will be generated using "Research Randomizer (www.randomizer.org). Each one of the aforementioned numbers will be assigned to a European country (country code) drawn according to alphabetic order from the list of participating countries. Next to the 2 digits of the country code, a one-to-three-digit number will be placed to represent each participant's temporal order of participation relative to other participants from the same country. The country-specific, temporal order of participants will be determined according to the exact time of submission of the completed study questionnaire. This will form the personal, three-to-five digit code for each study participant.

The survey database will be hosted by the CHARITE Berlin Simulation and Training Centre. Data on questionnaire domains will be stored next to the participant's personal code in Microsoft Excel format. This electronic Masterfile will be stored onto the password-protected hard disk of a personal computer. In addition, next to the participant's personal code, investigators will enter the following personal data 1) age; 2) gender; 3) religion; 4 and 5) country and center/city of professional activity; 6) characteristics of the participant's ICU (e.g. general, medical, surgical, etc.); 7) the participant's professional focus (e.g. clinical, research, etc.); and 8 and 9) the participant's professional grade (e.g. consultant, professor, etc.) and ICU experience (in years of work). There will be no collection/recording of participants' names or electronic mail addresses. Therefore, the Masterfile will fulfill the criterion of "pseudonymization." Development and testing of the questionnaire The final version of the questionnaire is presented in the uploaded official and approved version of the current protocol. The questionnaire was developed based on published, relevant literature, investigator consensus and comments collected by preliminary testing; the latter comprised administration of the questionnaire to a total of 30 Danish, British and Greek ICU clinicians (10 from each country). Danish and British clinicians were asked to qualitatively assess the readability and understandability of the survey questions and provide comments for improvement as regards content and clarity. Greek clinicians were asked to first respond to the questionnaire, in order to obtain an estimate of the actual time required for survey completion, and subsequently also provide their comments on content and clarity.

In its finalized form, the survey consists of an introduction and 4 sections. The introduction includes core definitions concerning advance directives and shared decision making, palliative care (https://www.who.int/europe/news-room/fact-sheets/item/palliative-care), treatment limitation, active shortening of the dying process (or euthanasia), and family. The first section aims at collecting data on participant characteristics, including age, gender, religion, location of professional activity (including hospital name), professional focus, grade and experience. The second section includes questions on patient-level, family-level and healthcare-level domains of palliative care, as well as participant-rating (5-point scale) of the importance of domain elements; these elements pertain to physical and psychosocial aspects of care, communication, family support and system-level support. The third section includes questions (and rating by 5-point scale wherever appropriate) about EOL legislation and practices such as applicability of and clinician-compliance with advance directives, treatment limitation (i.e. withholding or withdrawing of life-sustaining treatments), euthanasia, terminal analgesia/sedation, EOL decision-making and family presence during cardiopulmonary resuscitation. Lastly, the fourth section includes 12 questions with two-choice answers (i.e. yes/no) in the context of a recently introduced EOL practice score (with definitions of score subcomponents appended), and an additional 3 general questions about palliative care practice in the participant's ICU.

Website, survey promotion and access to the questionnaire The electronic version of the questionnaire will be set up at a dedicated website using LimeSurvey. Invitation links for this mandatory survey (meaning: all website visitors are expected to fill the survey) will be sent Europe-wide by EPIC partner European Society of Anaesthesiology and Intensive Care (ESAIC) with the use of social media and email channels; ESAIC has approximately 18.000 followers. Additional invitation links will be emailed by the investigators to their networks of research associates; the latter will also be encouraged to share the invitation links with their own networks of colleagues/associates.

Additional protocol features There are no planned incentives for participants. Data collection will be performed once from each participating center, within February 10th to June 30, 2024.

The questionnaire (appended to the current text) will be administered over a total of 22 electronic pages. The items (i.e. subsection questions) per page will range from 1 (open-ended questions) to 12 (multiple-choice questions). Options for responses such as "do not know" are provided for all questions of sections 2 and 3 and question 25.1 of section 4. In addition, adaptive questioning (i.e. "do not know" or "No" responses to the first question of a subsection triggering cancellation of all subsequent questions of that particular subsection) is used from question 15.1 through question 21.1 of section 3; this is aimed at reducing the number of questions in cases of 1) lack of participants' knowledge about whether certain EOL practices are legally allowed; or 2) participant awareness of absence of any preceding/current, local legalization of an EOL practice. In section 4, EOL practice score questions have by original protocol only 2 possible answers, whereas questions 25.2, 25.3 and 25.4 are open ended.

The completeness of participants' responsiveness (i.e. "all survey questions answered") will be checked before questionnaire submission, and participants will be prompted through a "pop-up" dialog box to provide any missing responses, or indicate that they do not wish to respond to a specific question. Furthermore, just prior to questionnaire submission, another "pop-up" dialog box will display a pre-submission prompt to review (and revise if needed) the originally provided responses.

Internet protocol (IP) address data will be used to ensure the uniqueness of survey participants. More specifically, unlimited, repetitive access to the questionnaire will be allowed from an IP address of a client computer until survey submission. However, following submission of the filled questionnaire through an IP address, will result in discontinuation of survey accessibility through that particular IP address.

Due to the use of snowballing dissemination and sharing of invitation links, it will not be possible to calculate the survey response rate. However, survey completion rate will actually be calculated by dividing the total number of submitted questionnaires to the number of first-survey page visitors. Furthermore, questionnaire completeness rate will be calculated by dividing the number of questionnaires without any missing response to the total number of submitted questionnaires. Missing responses will correspond to questions to which participants will have indicated that they do not wish to answer (see also above).

Mutually contradictory responses from participants originating from the same study center, or the same city, or the same country constitute a possibility, especially for the third and fourth section of the survey. For example, two or more participants from the same center/city/country may respond differently as regards the legal bindingness of advance directives. In such cases, mutually contradictory responses will be used to estimate the extent of country-specific variation in the perception of law and/or local palliative care practice.

Outcome Measures Study outcomes are detailed in the dedicated section. Statistical analysis plan Statistical analyses will include 1) descriptive statistics; 2) assessment of distribution normality (by Kolmogorov Smirnov test); 3) assessment of European country-/region-level heterogeneity in the responses; and 4) region-level comparisons of responses.

The investigators expect to determine substantial heterogeneity in the responses, primarily between northern and southern Europe and central and southern Europe, and generally between countries with high vs low palliative care service integration.

High palliative care score will be defined as presence of at least 23 positive responses (out of 35 possible positive responses) to question groups 10 through 14 of survey section 2.

Generalized estimating equations (GEE) models with robust standard errors and exchangeable working correlation structure accounting for the factor center will be applied to examine associations between a high palliative care score, the 9 clinician demographic variables of survey section 1 [with "country" stratified according to "northern, central or southern European region"], and the EOL practice score and/or its subcomponent EOL practice variables listed in question group 24 of survey section 4. The dependent variable in this analysis will be high palliative care score, yes or no. Reference category for region will be "southern Europe". The maximum number of GEE explanatory variables will be 21. Consequently, a target sample size of 500 participants is expected to result in approximately 24 observations (i.e. response data-points) per variable; this exceeds the recommended threshold of 20 observations per variable by 20% and establishes an adequate safety margin, aimed at addressing the possibility of missing responses.

Significance will be set at two-sided P <0.05. Analyses will be conducted using the latest versions of SSPS or R software.

The study questionnaire is included in the uploaded and approved, official version of the study protocol.

Minor modification in the study questionnaire: To ensure data protection, original question No. 5 (in section "clinician demographics") has been changed from "5. Name of hospital/healthcare facility and city of your professional activity (free text)" to "5. Total number of intensive care unit (ICU) beds at your hospital (number)".

NOTE: The current version of the registered protocol and the original ClinicalTrials.gov record contain a minor inconsistency in the definition of the palliative care score and the third primary outcome, arising from an initial, naive error in the calculation of the maximum possible score. In the finalized analysis, the palliative care score correctly ranges from 0 to 29 (Section 2, question groups 10-14), and a "high" palliative care score is defined as ≥20/29 (i.e., ≥ two thirds of the maximum score). This represents a refinement of the statistical analysis plan rather than a substantive change in study objectives, design, or data collection.

Similarly, the third primary outcome has been refined to more accurately reflect the structure of Section 4. Instead of a single 0-13 sum, it is now specified as a composite outcome comprising: (a) the End-of-Life Practice Score (0-12), defined as the number of "yes" responses to the 12 binary items in question group 24; and (b) the Perceived Palliative Practice Quality score (0-4), derived from Question 25.1, coded ordinally from 0 ("not at all") to 4 ("very high degree"), with "do not know" treated as missing. This refinement does not alter the underlying data collected or the conceptual focus of the outcome; it only clarifies and separates the organizational practice component from the clinicians' perceived palliative care focus.

No changes have been made to the survey instrument, recruitment, or ethical aspects of the study. The submitted working version of the ClinicalTrials.gov record is intended solely to correct these scoring details and align the registry with the final, prespecified analysis used in the manuscript.

Inclusion Criteria:

• European and Iraeli intensive care unit physicians or nurses.

Exclusion Criteria:

• Absence of informed consent for study participation.

Physicians or nurses working in European intensive care units (ICUs), or in ICUs in Israel.