BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] http://bmjopen.bmj.com/ on September 29, 2021 by guest. Protected copyright. BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from

End-of-life Management Protocol Offered Within Emergency Room (EMPOWER) – study protocol for a multicentre study

Journal: BMJ Open ManuscriptFor ID peerbmjopen-2019-036598 review only Article Type: Protocol

Date Submitted by the 20-Dec-2019 Author:

Complete List of Authors: Yash Pal, Rakhee; National University Hospital Kuan, Win Sen; National University Hospital , Singapore, Emergency Medicine; National University Hospital Tiah, Ling; Changi General Hospital Kumar, Ranjeev; Khoo Teck Puat Hospital Wong, Yoko; Singapore Clinical Research Institute Shi, Luming; Singapore Clinical Research Institute Zheng, Charles; Singapore Clinical Research Institute Lin, Jingping; National University Hospital Liang, Sufang; National University Hospital Segara, Uma; National University Hospital Yong, Woon Chai; National University Hospital; NUS Yong Loo Lin School of Medicine http://bmjopen.bmj.com/ Chan, Noreen; National University Hospital; NUS Yong Loo Lin School of Medicine Chua, Mui Teng; National University Hospital; NUS Yong Loo Lin School of Medicine Ibrahim, Irwani; National University Hospital; NUS Yong Loo Lin School of Medicine

PALLIATIVE CARE, ACCIDENT & EMERGENCY MEDICINE, Quality in Keywords: health care < HEALTH SERVICES ADMINISTRATION & MANAGEMENT on September 29, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 17 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 5 6 7 8 9 I, the Submitting Author has the right to grant and does grant on behalf of all authors of the Work (as defined 10 in the below author licence), an exclusive licence and/or a non-exclusive licence for contributions from authors 11 who are: i) UK Crown employees; ii) where BMJ has agreed a CC-BY licence shall apply, and/or iii) in accordance 12 with the terms applicable for US Federal Government officers or employees acting as part of their official 13 duties; on a worldwide, perpetual, irrevocable, royalty-free basis to BMJ Publishing Group Ltd (“BMJ”) its 14 licensees and where the relevant Journal is co-owned by BMJ to the co-owners of the Journal, to publish the 15 Work in this journal and any other BMJ products and to exploit all rights, as set out in our licence. 16 17 The Submitting Author accepts and understands that any supply made under these terms is made by BMJ to 18 the Submitting Author Forunless you peer are acting as review an employee on behalf only of your employer or a postgraduate 19 student of an affiliated institution which is paying any applicable article publishing charge (“APC”) for Open 20 Access articles. Where the Submitting Author wishes to make the Work available on an Open Access basis (and 21 intends to pay the relevant APC), the terms of reuse of such Open Access shall be governed by a Creative 22 Commons licence – details of these licences and which Creative Commons licence will apply to this Work are set 23 out in our licence referred to above. 24 25 Other than as permitted in any relevant BMJ Author’s Self Archiving Policies, I confirm this Work has not been 26 accepted for publication elsewhere, is not being considered for publication elsewhere and does not duplicate 27 material already published. I confirm all authors consent to publication of this Work and authorise the granting 28 of this licence. 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 29, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 17 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 TITLE PAGE 4 5 6 End-of-life Management Protocol Offered Within Emergency Room 7 8 (EMPOWER) – study protocol for a multicentre study 9 10 Rakhee Yash Pal1, Win Sen Kuan1,2, Ling Tiah3, Ranjeev Kumar4, Yoko Kin Yoke Wong5, 11 Luming Shi5, Charles Qishi Zheng5, Jingping Lin1, Sufang Liang1, Uma Chandra Segara1, Woon 12 Chai Yong6,7, Noreen Guek Cheng Chan6,7, Mui Teng Chua1,2*, Irwani Ibrahim1,2 13 14 1 Emergency Medicine Department, National University Hospital, National University Health System, Singapore 15 2 Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 16 3 Accident & EmergencyFor Medicine, peer Changi General review Hospital, Singapore only 17 4 Acute and Emergency Care Centre, Khoo Teck Puat Hospital, Singapore 18 5 Singapore Clinical Research Institute, Singapore 19 6 Division of Palliative Care, National University Cancer Institute, Singapore 20 7 Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 21 22 *Corresponding author: 23 Dr Mui Teng Chua 24 Address: Emergency Medicine Department, National University Hospital, 9 Lower Kent Ridge Road, Level 25 4, Singapore 119085 26 Email: [email protected] 27 Telephone number: +65 67723024 28 29 Addresses of all authors: 30 1 Emergency Medicine Department, National University Hospital, National University Health System, 31 Singapore. 9 Lower Kent Ridge Road, Level 4, Singapore 119085

32 http://bmjopen.bmj.com/ 33 2,7 Departments of Surgery and Medicine, Yong Loo Lin School of Medicine, National University of 34 Singapore, Singapore. 10 Medical Dr, Singapore 117597 35

36 3 37 Accident & Emergency Medicine, Changi General Hospital, Singapore. 2 Simei Street 3, Singapore 38 529889 39 4 Acute and Emergency Care Centre, Khoo Teck Puat Hospital, Singapore. 90 Yishun Central, Singapore 40 on September 29, 2021 by guest. Protected copyright. 41 768828 42 43 5 Singapore Clinical Research Institute, Singapore. 31 Biopolis Way, Singapore 138669 44 45 6 Division of Palliative Care, National University Cancer Institute, Singapore. 5 Lower Kent Ridge Road, 46 Singapore 119074 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 17 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 End-of-life Management Protocol Offered Within Emergency Room 4 5 (EMPOWER) – study protocol for a multicentre study 6 7 Rakhee Yash Pal1, Win Sen Kuan1,2, Ling Tiah3, Ranjeev Kumar4, Yoko Kin Yoke Wong5, 8 Luming Shi5, Charles Qishi Zheng5, Jingping Lin1, Sufang Liang1, Uma Chandra Segara1, Woon 9 Chai Yong6,7, Noreen Guek Cheng Chan6,7, Mui Teng Chua1,2*, Irwani Ibrahim1,2 10 11 1 Emergency Medicine Department, National University Hospital, National University Health System, Singapore 12 2 Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 13 3 Accident & Emergency Medicine, Changi General Hospital, Singapore 14 4 Acute and Emergency Care Centre, Khoo Teck Puat Hospital, Singapore 15 5 Singapore Clinical Research Institute, Singapore 16 6 Division of Palliative Care,For National peer University Cancerreview Institute, Singapore only 17 7 Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 18 19 ABSTRACT 20 21 Background 22 23 Patients at their end-of-life (EOL) phase frequently visit the emergency department (ED) due to their symptoms, yet 24 the environment and physicians in ED are not traditionally equipped or trained to provide palliative care. This 25 multicentre study aims to measure the current quality of EOL care in ED to identify gaps, formulate improvements 26 and implement the improved EOL care protocol. We shall also evaluate healthcare resource utilisation and its 27 associated costs. 28 29 Methods and analysis 30 31 This study employs a quasi-experimental interrupted time series design using both qualitative and quantitative methods, involving the EDs of 3 tertiary hospitals in Singapore, over a period of 3 years. There are 5 phases in this

32 http://bmjopen.bmj.com/ study: (1) retrospective chart reviews of patients who died within 5 days of ED attendance; (2) pilot phase to 33 validate the CODETM questionnaire in the local context; (3) pre-implementation phase; (4) focused group 34 discussions (FGDs); and (5) post-implementation phase. In the prospective cohort, patients who are actively dying 35 or have high likelihood of mortality this admission, and whose goal of care is palliation, will be eligible for 36 inclusion. At least 140 patients will be recruited for each pre- and post-implementation phase. There will be face-to- 37 face interviews with patients’ family members, review of medical records and self-administered staff survey to 38 evaluate existing knowledge and confidence. The FGDs will involve hospital and community healthcare providers. 39 Data obtained from the retrospective cohort, pre-implementation phase and FGDs will be used to guide prospective

40 improvement and protocol changes. Patient, family and staff relevant outcomes from these changes will be measured on September 29, 2021 by guest. Protected copyright. 41 using time series regression. 42 43 Ethics and dissemination 44 45 The study protocol has been reviewed and ethics approval obtained from the National Healthcare Group Domain 46 Specific Review Board, Singapore. The results from this study will be actively disseminated through manuscript 47 publications and conference presentations. 48 49 (290 words, excluding subheadings) 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 17 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 ARTICLE SUMMARY 4 5 Strengths and limitations: 6  This novel multicentre study evaluating EOL care provided in the ED will be able to evaluate the quality of 7 clinical management rendered to EOL patients via a multi-prong approach, from clinical charts and 8 perceptions of family members who were at the bedside 9  By involving our community partners, we shall also be able to identify difficulties faced in caring for these 10 patients in a non-institutional setting 11  Results of this study have the potential to identify current gaps and barriers, and inform emergency 12 physicians globally on best practices in providing EOL care in the ED 13  Some limitations to the study design include response rates to questionnaires, selection bias due to non- 14 probability sampling and information bias from retrospective chart reviews 15 16 For peer review only 17 TRIAL REGISTRATION 18 19 Ethics approval for this study was obtained from the National Healthcare Group Domain Specific Review Board 20 (DSRB reference no: 2018/00838) for waiver of informed consent for the retrospective cohort in pre-Phase 1. 21 Written informed consent will be obtained from all study participants in all other phases. The study protocol has 22 been registered in ClinicalTrials.gov (NCT03906747) on April 8, 2019. 23 Enrolment for the study is currently ongoing. 24 25 KEY WORDS: 26 27 Palliative care, emergency medical services, cost effectiveness, terminally ill 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 17 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 4 BACKGROUND 5 6 In a report by the Institute of Medicine in 1997, a decent or good death, is one that should be free of unnecessary 7 distress and suffering for both patients and their caregivers, and should be in general accord with their wishes, 8 within reasonable consistency with clinical, cultural and ethical standards.1 The World Health Organization (2016) 9 defines palliative care as a multi-dimensional approach to improve the quality of life for dying patients and their 10 families by caring for the physical, psychosocial and spiritual needs. Emergency physicians, being at the forefront of 11 medicine and healthcare delivery, frequently encounter death as a daily occurrence. Globally, it is estimated that at 12 least 35 million (60%) out of 58 million people who are dying annually suffer from advanced illnesses and would 13 benefit from palliative care services for better quality of life.2 In addition, up to 80% of terminally ill patients prefer 3,4 14 to die at home and yet only a small percentage are able to achieve this. 15 16 As the first point of contactFor in the healthcarepeer system, review the emergency department only (ED) is particularly important for 17 provision of appropriate end-of-life (EOL) care, especially when a majority of EOL patients utilise emergency services in the last month of their lives.5 Many patients present to the ED due to symptoms of distress, poorly- 18 controlled symptoms at home, or families who are unable to provide care as a result of physical and mental distress.6 19 Traditionally, the ED has not been regarded as an optimal place for EOL care, due to its inherent nature of 20 overcrowding and chaos, resulting in lack of tranquil infrastructure for families to spend meaningful time with their 21 loved ones during their last moments.4,7 ED physicians are also perceived to lack knowledge and confidence in 22 managing such patients as palliative care contradicts their conventional training on aggressive resuscitation 23 measures. The initiation of palliative care from ED by emergency physicians was found to be at a dismal rate of 24 18% in one study.8 This highlights the need for more specialised training for emergency physicians to adequately 25 manage the increasing number of patients facing progressive illnesses and disabilities towards the end of their lives. 26 27 Palliative care services in Singapore have grown and developed over the past 30 years and is acknowledged as an 28 essential specialty in the healthcare system.9 However, palliative care development in ED is limited10 and caring for 29 patients’ psychosocial and spiritual needs is not the usual goal in the ED. On the other hand, studies have shown that 30 implementation of EOL pathways not only increases the knowledge and confidence of nursing personnel in 31 managing dying patients, but also leads to consistent and improved patient care.11–13 Moreover, palliative care in 14

32 terminally ill patients has been shown to reduce hospitalisation costs. Novel initiatives for EOL care education for http://bmjopen.bmj.com/ 33 ED healthcare staff and clinical guidelines for delivery of EOL care have been implemented in several Singapore 34 public hospitals. These initiatives serve to enhance the quality of EOL care provision for imminently dying patients 35 and their grieving family members. Effective EOL programmes improve patient-provider communication, provide 36 comfort and relief of burdensome symptoms in terminally-ill patients, pay attention to spiritual needs, maintain 15 37 patients’ dignity, and respect family members’ social concerns, during the last hours or days of life. 38 39 Since the progressive initiation of the ED EOL programme in Singapore since 2014, its impact on quality improvement has yet to be determined.

40 on September 29, 2021 by guest. Protected copyright. 41 We propose this multicentre interrupted time series study to improve the quality of EOL care for actively dying 42 patients in the ED. The specific objectives in this study are to: 43  Systematically measure the current quality of EOL care in three Singapore EDs to identify gaps in 44 management 45  Formulate improvements needed to address these gaps and implement the improved EOL care protocol 46  Measure (a) the quality of care post-implementation of the improved EOL care, and (b) the impact on 47 health services utilisation and associated costs 48 49 METHODS/DESIGN 50 51 Study design 52 53 This is a quasi-experimental interrupted time series study conducted in the EDs of 3 public hospitals (National 54 University Hospital [NUH], Changi General Hospital [CGH] and Khoo Teck Puat Hospital [KTPH]) in Singapore to 55 evaluate the quality of EOL care rendered with measures conducted post-implementation of interventions to assess 56 improvement in EOL care. The project is overseen by biostatisticians and epidemiologists from the Singapore 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 17 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 Clinical Research Institute (SCRI) and site investigators from participating hospitals. The study comprises 5 phases: 4 pre-Phase 1, pilot phase, and Phases 1 to 3. Apart from pre-phase 1 which is retrospective, all other phases are 5 conducted prospectively. Details of each phases with variables and data to be collected are detailed below. 6 7 Study setting and sites 8 9 The public hospitals included in this study, namely NUH, CGH and KTPH, belong to the 3 main healthcare clusters 10 in Singapore – the National University Health System, Singapore Health Services and National Healthcare Group, 11 which serve the country’s western, eastern and northern populations, respectively.16 Each of the 3 hospitals are 12 tertiary centres with annual ED census of more than 100,000 attendances. 13 14 Pre-Phase 1 (retrospective) 15 16 In pre-Phase 1, we aim Forto determine peer the baseline data review prior to any implementation only of EOL protocols in the 17 respective hospitals to facilitate the assessment of the impact on health services utilisation. Using the hospital database, patients who died within 5 days of ED attendance will be identified. Chart reviews will be performed to 18 determine if the patient fulfilled the criteria for EOL management. Information on the health services utilisation and 19 associated costs will be collected. The period of the chart reviews are as follows: 20  NUH: 01 January 2013 to 31 December 2013 21  CGH: 01 July 2016 to 30 June 2017 22  KTPH: 01 July 2016 to 30 June 2017 23 24 These periods of chart reviews correspond to the 12 months that preceded initiation of any EOL initiative in the EDs 25 of the respective hospitals. 26 27 Pilot phase (January to April 2019) 28 29 The pilot phase tests the face and construct validity, and reliability of a newly developed questionnaire for 30 measuring the quality of EOL care in EDs in the Asian context. The questionnaire development takes reference from 31 the Care of the Dying Evaluation (CODETM).17 Participants, who are family members of EOL patients, will be

32 requested to complete the newly developed questionnaire (re-named Care of the Dying Evaluation - Emergency http://bmjopen.bmj.com/ 33 Medicine, CODE-EM), and then interviewed about their experience of completing the questionnaire and to 34 participate in a retest of the questionnaire one month later either by phone or mail. The post-questionnaire interview 35 aims to examine the validity of the questionnaire in terms of language, length, timing and relevance. 36 37 Phase 1 (April 2019 to April 2020 projected) 38 39 At the start of this study, all 3 hospitals have EOL workflows in place that was developed within each department. A brief summary and comparison of the 3 workflows are illustrated in Figure 1. In Phase 1 of the study, quality 40 on September 29, 2021 by guest. Protected copyright. 41 indicators will be measured prospectively to evaluate the aspects of EOL care rendered to patients under existing 42 workflows. Data of recruited patients from the start of study enrolment until demise or terminal discharge from the 43 hospital will be collected. Their electronic medical records and ED consult notes will be reviewed from time of ED 44 attendance until terminal discharge or death to assess their inpatient progress and discharge outcomes. Apart from chart reviews of the EOL patients, the validated CODE-EM questionnaire will also be conducted with the nominated 45 caregivers of these patients. The quality indicators will be divided into patient-, family- and staff-centred outcomes. 46 47 48 Figure 1. EOL care workflow in the Emergency Departments of National University Hospital (NUH), Changi General Hospital 49 (CGH) and Khoo Teck Puat Hospital (KTPH) 50 51 52 The patient-centred outcomes to be evaluated include proportion of patients who fits EOL criteria and died within 5 53 days, proportion of patients on EOL pathway who have documentation that patient and/or family are given 54 opportunities to discuss an individualised care plan, and proportion of patients with symptoms who are prescribed 55 with medicines with individualised indications for use, dosage and route of administration. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 17 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 In addition, the following data will be collected for all recruited EOL patients: demographic information, clinical 4 presentation, comorbidities, healthcare utilisation, interventions post-EOL, length of hospitalisation, 5 communications with the family, comfort care measures, cost of hospitalisation, prior do-not-resuscitate decisions, 6 disposition (death, terminal discharge home or hospice, admission to inpatient unit) and cause of death. 7 8 Family-centred indicators are evaluated using the validated CODE-EM questionnaire (see Supplementary Material 9 1) to assess the quality of care and the level of support provided to patients and their families in the ED. The key 10 indicators include proportion of participants who perceived that their family member was treated with dignity and 11 respect by doctors and nurses, and the proportion of participants who perceived that they were adequately supported. 12 13 For staff-centred indicators, we aim to determine the level of healthcare providers’ knowledge and perception of the 14 EOL care using a locally-designed anonymous survey (see Supplementary Material 2) conducted monthly in the first 15 3-months, followed by 6 and 12 months later. This survey will be self-administered. The main components to be 16 evaluated are level of confidenceFor in peer communications review and providing emotional only support, knowledge, clarity of nurses’ 17 and doctors’ roles, conduciveness of physical environment and satisfaction of overall care quality rendered in managing EOL patients in the ED. 18 19 Phase 2 (March to April 2020 projected) 20 21 Phase 2 will commence at the end of Phase 1 based on the results obtained and will encompass focus group 22 discussions (FGDs), revision of EOL care components and implementation of improved EOL care in the 3 23 institutions. There will be separate FGDs for healthcare professionals, family members and primary care providers 24 (such as general practitioners and home hospice healthcare workers) conducted to identify the gaps to the current 25 EOL care. Recommendations to improve the quality of the existing EOL care will also be sought. 26 27 By combining the qualitative and quantitative results of Phases 1 and 2, the study investigators from the 3 hospitals 28 will assimilate all data and revise the current protocol accordingly. Combined and in-house staff training will ensue 29 thereafter. 30 31 Phase 3 (May 2020 to May 2021 projected)

32 http://bmjopen.bmj.com/ 33 In Phase 3, new and improved EOL workflows will be rolled out in all 3 hospitals. The quality indicators and 34 outcomes to be evaluated will be the same as in Phase 1 (as detailed above). The results obtained in Phase 3 will 35 then be compared to that in Phase 1. 36 37 Project materials 38

39 Permission was obtained from the original developers of the CODE-EM questionnaire to adapt it for local and ED contexts (see Supplementary Material 1), which was validated during the pilot phase. To evaluate staff-centred

40 on September 29, 2021 by guest. Protected copyright. indicators, a locally designed staff survey was used (see Supplementary Material 2). The staff survey was developed 41 by ED attendings and nurses trained in palliative care for the purpose of evaluating quality of care rendered during 42 the initiation of the EOL workflow in its early stages in NUH in 2013. The staff survey was piloted and improved 43 upon during those initial stages, not within the study period. 44 45 Patient selection 46 47 All the following inclusion criteria must be fulfilled: 48 • Actively dying patient or high likelihood of mortality within this admission 49 • Family accepts that the goals of care are provision of comfort, symptom relief and respect of dignity 50 • Patient is not a candidate for cardiopulmonary resuscitation, endotracheal intubation or transfer to the 51 intensive care unit 52 • Any of the life-limiting conditions: chronic frailty with poor functional state and limited reversibility 53 (Karnofsky Performance Scale [KPS] <40%)18; chronic severe illness with poor prognosis (terminal cancer, 54 end-stage renal failure [refusal or withdrawal of dialysis], end-stage respiratory, heart or liver disease, 55 advanced neurological disease including dementia, Parkinson’s disease and severely disabling stroke); or, 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 17 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 acute severe catastrophic conditions and at risk of dying with complications that are not reversible, as 4 subject to the treating clinician’s judgment 5 6 We excluded the following subjects: vulnerable population (e.g. prisoners, pregnant women); refusal to participate; 7 patients who have been recruited, or had declined participation, during the previous ED attendance(s); patients in 8 peri-arrest state; and/or family members who are not available or present at the patient’s bedside. 9 10 Sample size calculation 11 12 Preliminary data from 2014 to 2017 in NUH for calculation indicated an average of 7 patients who fulfilled the EOL 13 criteria in NUH every month. We expect to receive approximately 15 eligible EOL patients per month from the 3 14 centres and about 195 eligible patients in Phase 1 (13 months) and Phase 3 (13 months), respectively. 15 16 Based on NUH’s existingFor data, the currentpeer correct identificationreview rate is 24.4%. only We performed a power analysis to 17 estimate the expected power with various combinations of recruited subjects and observed improvement. We could achieve an 80% power with an alpha of 0.05 for a two-sided test, if we were to: 18 1) Recruit 200 subjects for each phase and expected to observe an absolute difference ≥ 13%; 19 2) Recruit 180 subjects for each phase and expected to observe an absolute difference ≥ 14%; 20 3) Recruit 160 subjects for each phase and expected to observe an absolute difference ≥ 15%; 21 4) Recruit 140 subjects for each phase and expected to observe an absolute difference ≥ 16%. 22 23 Outcomes of interest and statistical analyses 24 25 (i) Pilot phase analysis planning 26 27 The post-questionnaire interviews about the experience of completing the questionnaire will be transcribed verbatim 28 and analysed using a content analysis framework. Data will be refined into specific categories, with words and 29 phrases of shared meaning. A randomised selection of interview transcripts will be independently reviewed by a 30 second investigator, not directly involved in the data collection, to check for coding, and any discrepancies will be 31 discussed with a third investigator. The stability of the developed questionnaire over time will be assessed using the

32 following measures: percentage agreement; κ statistic (Cohen’s for nominal response options and weighted for http://bmjopen.bmj.com/ 33 ordinal response options); and Spearman’s correlation coefficient (for ordinal data). The criteria for good stability 34 over time is defined as percentage agreement >70%; κ>0.6; and r>0.7 and moderate stability over time as percentage 35 agreement >30%; κ>0.40; and r>0.3. Confirmatory factor analysis will be used to assess construct validity. The 36 suitability of questions will be examined by inspection of the correlation matrix and the Goodness of Fit Index. 37 38 (ii) Statistical analysis for other phases 39 Descriptive statistics will be obtained about the baseline characteristics of participants. Continuous variables will be

40 on September 29, 2021 by guest. Protected copyright. presented as mean ± standard deviation while categorical variables will be reported as absolute numbers and 41 percentages. Differences regarding outcomes will be analysed using independent-samples t-test or Mann-Whitney U 42 test, or paired t-test for pre- and post-EOL measurements. Symptom burden (e.g. differences in frequency of 43 symptoms) and CODE-EM questionnaire will be calculated. 44 45 Time series regression will be used to analyse the results from staff survey, using 7 data points (i.e. collected at 46 every 2 months) from each phase. The best fit pre- and post-programme revision lines will be estimated by using 47 linear regression, and auto-correction will be adjusted for by using maximum likelihood methods where appropriate. 48 The change in the slopes of the regression lines will be estimated. Data will be analysed using Stata version 15 49 (Stata Corp, College Station, Texas, USA). 50 51 (iii) Health economic impact analysis 52 53 This study will also evaluate the financial viability of EOL care by conducting a health economic evaluation, from a 54 hospital system perspective. Information on resource utilisation associated with patient management will be 55 recorded for each participant, including initial management in ED, services associated with protocol implementation 56 and follow-up. The potential impact of protocol implementation on medical and administrative management 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 17 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 resources will be studied by comparing with historical data from chart reviews. After identifying the extra care 4 components associated with study protocol implementation (e.g. accessing the on-call palliative care specialist), for 5 base case analysis, private rates of each service will be applied to calculate the average cost and savings. The health 6 economic analysis will consist of 2 components: (1) a cost-benefit analysis to assess the cost-benefit ratio, which 7 will be presented as “for every 1 extra dollar spent on the protocol implementation, how much savings can be 8 realised by the potential reduction in hospital admissions and length of hospital stay”; and (2) a cost-effectiveness 9 analysis to estimate the incremental cost-effectiveness ratio. 10 11 To address possible variations in patient medical conditions and the outcomes, such as variations in compliance, and 12 length of hospital stay, a series of one-way sensitivity analyses will be conducted to address the impact of each 13 parameter's uncertainty and assess the robustness of study estimations for generalisability. 14 15 ETHICS AND DISSEMINATION 16 For peer review only 17 Ethics approval for this study was obtained from the National Healthcare Group Domain Specific Review Board (DSRB reference no: 2018/00838) for waiver of informed consent for the retrospective cohort in pre-Phase 1. 18 Written informed consent will be obtained from all study participants in all other phases. The study protocol has also 19 been registered in ClinicalTrials.gov (NCT03906747). The findings from this study will be disseminated locally and 20 internationally through manuscript publications in peer-reviewed journals and conference presentations at national 21 and international platforms. 22 23 DISCUSSION 24 25 To the best of our knowledge, this is the first multicentre study using both qualitative and quantitative methods and 26 focusing on evaluation of EOL care provided in the ED. Additionally, apart from focusing on patients with chronic 27 irreversible illnesses such as advanced cancer, our study cohort includes other death trajectories such as sudden 28 death due to an acute catastrophic event. This is an area where information on palliative care is lacking. 29 30 Our study design has distinct advantages over the traditional epidemiological cohort (before-and-after) and case- 31 control designs that are limited by confounding and lack of usefulness in intervention studies. The interrupted time

32 series study design would be able to detect changes that are delayed or intermittent and can assess if the change is http://bmjopen.bmj.com/ 33 permanent or temporary. The design is simpler without the need for randomisation, which is not possible in EOL 34 research. There is also the ability to control for confounding variables and regression to the mean. 35 36 Furthermore, our study design allows us to obtain both qualitative and quantitative knowledge with regard to current 37 quality of care and identify areas for improvement. The information obtained from intimate FGDs with the family 38 members and our community healthcare partners would be consequential in providing in-depth information on the 39 current quality of care and difficulties faced in order to guide changes in protocols and policies. The unique setting of a multicultural Singapore society may also provide some insights into previously under-investigated cultural

40 on September 29, 2021 by guest. Protected copyright. beliefs and values that can be a focus of future qualitative studies in the region. 41 42 Despite the strengths of our study, the design presents some inherent limitations. First, the chart reviews of the 43 retrospective cohort in Pre-Phase 1 may have information bias. We aim to mitigate this by looking through all 44 available paper and electronic medical records to reduce missing information. Second, the response rates from 45 questionnaires and staff surveys may also pose a challenge. In order to encourage participation, small tokens of 46 appreciation and incentives will be given out to study participants. Third, the use of non-probability sampling and 47 enrolment when research assistants are around during office hours may also introduce a degree of selection bias. 48 This is however necessary due to economic constraints that preclude round-the-clock staffing of research personnel. 49 However, we have mitigated this by screening through ED attendances beyond office hours and over the weekend, 50 and contacting eligible family members in the wards whenever appropriate for invitation to participate. 51 52 In conclusion, the results of our study would have the potential to identify current barriers and inform emergency 53 physicians globally on the best practices in providing EOL care in the ED for patients who are at the last stages of 54 their lives from a myriad of illnesses. 55 56 (3245 words) 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 17 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 4 DECLARATIONS 5 6 Ethics approval and consent to participate 7 Ethics approval for this study was obtained from the National Healthcare Group Domain Specific Review Board 8 (DSRB reference no: 2018/00838) for waiver of informed consent for the retrospective cohort in pre-Phase 1. 9 Written informed consent will be obtained from all study participants in all other phases. The study protocol has also 10 been registered in ClinicalTrials.gov (NCT03906747). 11 12 Consent for publication 13 Not applicable. 14 15 Availability of data and materials 16 Not applicable. For peer review only 17 Competing interests 18 The authors declare that they have no competing interests. This study is funded by a major government funding 19 body (National Medical Research Council, Ministry of Health, Singapore) and there is no funding or assistance from 20 any commercial organisation. 21 22 Funding 23 This work is supported by a grant from the National Medical Research Council Health Services Research Grant 24 (grant number: MOH-HSRG-EoL17Ju-0001). The funding body provides funds for employment of research 25 assistants, engagement of statistical support from Singapore Clinical Research Institute and pays for all study-related 26 expenses such as reimbursements of study participants and manuscripts processing fees. It is not involved in the 27 design of the study, and collection, analysis, and interpretation of data, and in writing of the manuscript. 28 29 Author contributions 30 RYP conceived the study, designed the final study protocol, provided the domain knowledge expertise and is the 31 principal investigator of the grant; WSK helped in the design of the final study protocol, contributed to the technical

32 design and revised the initial manuscript draft; LT and RK helped in the design of the final study protocol and are http://bmjopen.bmj.com/ 33 coordinating site investigators; YKYW, LS and CQZ contributed to the technical design and provided biostatistical 34 and epidemiological support; JL, SL, UCS, WCY and NGCC provided clinical input and perspectives to the 35 qualitative aspects of the study; MTC helped in the design of the final study protocol, drafted the initial manuscript 36 and coordinated ethics approval; II helped in the design of the final study protocol, had a part in the technical design 37 and coordinated collaboration among investigators from all institutions. All authors read and approved the final 38 manuscript. 39 Acknowledgements

40 on September 29, 2021 by guest. Protected copyright. Not applicable. 41 42 Author details 43 1 Emergency Medicine Department, National University Hospital, National University Health System, Singapore 44 2 Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 45 3 Accident & Emergency Medicine, Changi General Hospital, Singapore 46 4 Acute and Emergency Care Centre, Khoo Teck Puat Hospital, Singapore 47 5 Singapore Clinical Research Institute, Singapore 48 6 Division of Palliative Care, National University Cancer Institute, Singapore 49 7 Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 17 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 4 REFERENCES 5 6 1. Field MJ, Cassel CK. Approaching death: improving care at the end of life. N Engl J Med 1998; 339: 274– 7 275. 8 2. Stjernswärd J, Foley KM, Ferris FD. The Public Health Strategy for Palliative Care. J Pain Symptom 9 Manage 2007; 33: 486–493. 10 3. Tang ST. When death is imminent: where terminally ill patients with cancer prefer to die and why. Cancer 11 Nurs 2003; 26: 245–251. 12 4. Gloss K. End of life care in emergency departments: a review of the literature. Emerg Nurse 2017; 25: 29– 13 38. 14 5. Smith A, McCarthy E, Weber E, et al. Half of older Americans seen in emergency department in last month 15 of life; most admitted to Hospital, and many die there. Health Aff 2012; 31: 1277–1285. 16 6. Krikorian A, LimoneroFor JT, peerMaté J. Suffering review and distress at the end-of-life.only Psychooncology 2012; 21: 799– 17 808. 7. Beckstrand RL, Wood RD, Callister LC, et al. Emergency nurses’ suggestions for improving end-of-life care 18 obstacles. J Emerg Nurs 2012; 38: e7. 19 8. Kistler EA, Sean Morrison R, Richardson LD, et al. Emergency department-triggered palliative care in 20 advanced cancer: proof of concept. Acad Emerg Med 2015; 22: 237–239. 21 9. Lien Centre for Palliative Care, Duke-NUS Graduate Medical School. Report on the National Strategy for 22 Palliative Care., https://www.singaporehospice.org.sg/en/wp- 23 content/uploads/2018/08/Report_on_National_Strategy_for_Palliative_Care-5Jan2012.pdf (2011, accessed 24 13 June 2017). 25 10. Chan GK. End-of-life models and emergency department care. Acad Emerg Med 2004; 11: 79–86. 26 11. Selman L, Robinson V, Klass L, et al. Improving confidence and competence of healthcare professionals in 27 end-of-life care: An evaluation of the ‘Transforming End of Life Care’ course at an acute hospital trust. BMJ 28 Support Palliat Care 2016; 6: 231–236. 29 12. Luhrs CA, Penrod JD. End-of-life care pathways. Curr Opin Support Palliat Care 2007; 1: 198–201. 30 13. Weng TC, Yang YC, Chen PJ, et al. Implementing a novel model for hospice and palliative care in the 31 emergency department. Med (United States) 2017; 96: e6943.

32 14. Aldridge MD, Kelley AS. The myth regarding the high cost of end-of-life care. Am J Public Health 2015; http://bmjopen.bmj.com/ 33 105: 2411–2415. 34 15. Threapleton DE, Chung RY, Wong SY, et al. Care towards the end of life in older populations and its 35 implementation facilitators and barriers: a scoping review. J Am Med Dir Assoc 2017; 18: 1000-1009.e4. 36 16. Poon CH. Public healthcare sector to be reorganised into 3 integrated clusters, new polyclinic group to be 37 formed. The Straits Times, https://www.straitstimes.com/singapore/health/public-healthcare-sector-to-be- 38 reorganised-into-3-integrated-clusters-new (2017, accessed 16 October 2019). 39 17. Mayland CR, Lees C, Germain A, et al. Caring for those who die at home: The use and validation of ‘Care Of the Dying Evaluation’ (CODE) with bereaved relatives. BMJ Support Palliat Care 2014; 4: 167–174.

40 on September 29, 2021 by guest. Protected copyright. 18. Karnofsky D, Burchenal J. The clinical evaluation of chemotherapeutic agents in cancer. Eval Chemother 41 agents 1949; 191–205. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 17 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Figure 1. EOL care workflow in the Emergency Departments of National University Hospital (NUH), Changi General Hospital (CGH) and Khoo Teck Puat Hospital (KTPH) 31

32 168x119mm (120 x 120 DPI) http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

1 2 Supplementary Material 1 3 CODE-EM Family Survey Tool 4 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 5 CODE-EM Questionnaire 6 7 8 There was enough help with nursing care in the ED, such as Strongly Strongly 9 1. giving medicines, changing diapers and helping the patient Agree Neutral Disagree agree disagree 10 find a comfortable position in bed. 11 12 The bed area in the ED and surrounding environment was Strongly Strongly 13 2. Agree Neutral Disagree comfortable for the patient. agree disagree 14 15 The bed area in the ED and surrounding environment was Strongly Strongly 16 3. Agree Neutral Disagree 17 comfortable for the family. agree disagree 18 For peer review only The bed area in the ED and surrounding environment had Strongly Strongly 19 4. Agree Neutral Disagree 20 adequate privacy for the patient agree disagree 21 22 23 Did you have confidence and trust in the ED nurses who Yes, in Yes, in No, not in any of 24 5. were caring for the patient? all of them some of them the nurses 25 26 Did you have confidence and trust in the ED doctors who Yes, in Yes, in No, not in any of 6. 27 were caring for the patient? all of them some of them the doctors 28 29 30 In your opinion, during the patient’s stay in the ED, did the Yes, all of Yes, some No, s/he did not appear to 31 7. 32 patient appear to be in pain? the time of the time be in pain 33 In your view, did the doctors and nurses in the ED do enough Yes, all of Yes, some No, not Not applicable, 34 8. s/he was not in to help relieve the pain? the time of the time at all 35 pain 36 http://bmjopen.bmj.com/ 37 38 In your opinion, during the stay in the ED, did the patient Yes, all of Yes, some No, s/he did not appear 9. 39 appear to be restless? the time of the time to be restless 40 41 In your view, did the doctors and nurses in the emergency Yes, all of Yes, some No, not Not applicable, 10. s/he was not 42 department do enough to help relieve the restlessness? the time of the time at all restless 43 44 on September 29, 2021 by guest. Protected copyright. 45 No, s/he did not have a In your opinion, during the stay in the ED, did the patient Yes, all of Yes, some 46 11. noisy rattle to the the time of the time 47 appear to have a “noisy rattle” when breathing? breathing 48 No, Not applicable, 49 In your view, did the doctors and nurses in the ED do enough Yes, all of Yes, some s/he did not have 12. not at the time of the time a noisy rattle to 50 to help relieve the “noisy rattle” when breathing? all 51 the breathing 52 53 In your opinion, during the stay in the emergency 54 Yes, all of Yes, some No, s/he did not have 13. department, did the patient appear to have difficulty 55 the time of the time difficulty breathing breathing? 56 In your view, did the doctors and nurses in the emergency Not applicable, 57 Yes, all of Yes, some No, not s/he did not 14. department do enough to help relieve the breathing 58 the time of the time at all have difficulty 59 difficulty? breathing 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 17

1 2 During the time in the ED, the patient's care and treatment Strongly Strongly 3 15. Agree Neutral Disagree agree disagree 4 was discussed with you and/or the family. BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 5 The healthcare team at the ED explained the patient's Very Very 6 16. condition and treatment in a way you found easy or difficult Easy Neutral Difficult easy Difficult 7 to understand. 8 9 How would you assess the overall level of emotional support Very 17. Poor Fair Good Excellent 10 given to you by the ED healthcare team? poor 11 12 The ED healthcare team discussed the patient's religious or Strongly Strongly 18. Agree Neutral Disagree 13 spiritual needs. agree disagree 14 15 16 17 19. Were you told that the patient would be likely to die soon? Yes No 18 For peer review only 19 19a) If yes, who told you s/he was likely to pass away soon? 20 21 22 23 Did a member of the ED healthcare team talk to you about 24 20. what to expect during the dying process (e.g. what Yes No 25 symptoms may arise)? 26 If yes  Was the discussion about what to expect during the 27 20a) Yes No 28 dying process helpful? 29 If no  Would a discussion about what to expect during the 30 20b) Yes No 31 dying process have been helpful? 32 33 34 In your view, the patient was treated with respect and Strongly Strongly 35 21. Agree Neutral Disagree agree disagree 36 dignity by the ED doctors. http://bmjopen.bmj.com/ 37 In your view, the patient treated with respect and dignity by Strongly Strongly 22. Agree Neutral Disagree 38 the ED nurses. agree disagree 39 40 41 42 Overall, in your opinion, were you adequately supported 23. Yes No 43 during the patient's stay in ED? 44 on September 29, 2021 by guest. Protected copyright. 45 46 47 48 ------THE END ------49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 17 BMJ Open

1 2 3 Supplementary Material 2 4 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 5 6 7 Staff Survey on Palliative Care in the Emergency Department 8 9 This is a survey on staff perceptions of the care provided to dying patients in the Emergency 10 Department. We want to know more about your clinical practices, confidence in managing dying 11 patients, and views on the experiences on dying patients and their families in our department. 12 13 Please answer these questions as they apply to you right now. We appreciate that your responses 14 have likely changed over time, and we seek to capture your answers at this current point. 15 16 17 18 For peer review only 19 Section A: Demographics 20 Please put a tick the box below to indicate your choice. 21 22 1. Please tell us your gender 23  Male 24  Female 25 26 27 2. Please tell us your profession 28  Medical (if medical, please answer questions 3 and 4) 29  Nursing (if nursing, skip questions 3 and 4) 30 31 3. If you are a medical doctor, please select your designation 32  Associate consultant and above 33  Senior resident 34 35  Resident physician with supervisory role 36  MOPEX MO http://bmjopen.bmj.com/ 37  Clinical associate/ resident physician with MO role 38  EM resident 39  Non-EM resident (e.g. Internal Medicine, Radiology) 40  Locum doctor 41 42 43 4. If you are doing a rotation in the Emergency Department, please indicate the duration of your 44 posting on September 29, 2021 by guest. Protected copyright. 45  1 month 46  2 months 47  3 months 48  6 months 49 50  For other duration, please state in months: ______51 52 5. Your years of clinical experience from graduation 53  < 2 54  2-5 55  5-10 56 57  10-20 58  > 20 59 60 6. Your palliative care training (Please select the most appropriate option)

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 17

1 2 3  Graduate diploma or higher 4 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from  Module/elective from medical/nursing school 5 6  In-service lectures/clinical attachment/course 7  On the job training only 8  No training 9 10 7. How would you rate your knowledge of palliative care? 11  Extensive knowledge 12 13  Adequate knowledge 14  Limited knowledge 15  No knowledge 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 29, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 17 BMJ Open

1 2 3 The following questions utilise a Likert scale from 1 to 5 (where 1 is strongly disagree, 5 is strongly 4 agree). Please circle the number that most reflects your opinion now. BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 5 6 7 Section B: Care provided by the emergency department 8 Strongly Strongly 9 Disagree Neutral Agree Disagree Agree 10 11 8. I am confident of managing a dying patient in the emergency 12 1 2 3 4 5 13 department. 14 15 9. I am satisfied by the overall 16 standard of palliative care 17 provided in the emergency 1 2 3 4 5 18 department. For peer review only 19 20 10. I am clear on the roles and 21 responsibilities of the 22 emergency department nurses 23 1 2 3 4 5 24 and doctors in the care of a 25 dying patient. 26 27 11. The physical environment in the 28 emergency department is 29 conducive for the provision of 1 2 3 4 5 30 care in a dying patient. 31 32 12. The overall experience of a 33 dying patient and their family in 34 the emergency department is 1 2 3 4 5 35 36 satisfactory. http://bmjopen.bmj.com/ 37 38 39 Section C: Communication with patients and relatives 40 Strongly Strongly 41 Disagree Neutral Agree Disagree Agree 42 43 13. I am able to give emotional 44 support to a dying patient and 1 2 3 4 5 on September 29, 2021 by guest. Protected copyright. 45 their family. 46 47 14. I routinely discuss a patient’s 48 religious or spiritual needs with 1 2 3 4 5 49 the family. 50 51 52 53 54 55 56 57 58 59 60

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 17

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 5 Strongly Strongly 6 Disagree Neutral Agree Disagree Agree 7 8 15. I am able to explain a dying 9 patient’s condition and/or 10 treatment in a way the family 1 2 3 4 5 11 can easily understand. 12 13 16. I routinely engage the family in 14 decisions about a patient’s care 15 1 2 3 4 5 16 and treatment. 17 18 17. I routinely tellFor a patient’s peer family review only 19 what to expect when a patient 20 is dying (e.g. symptoms like 21 restlessness, drowsiness, 1 2 3 4 5 22 incontinence or changes in 23 breathing patterns). 24 25 26 27 28 Please feel free to comment, if you wish to, on any aspect of the care you provide to dying patient 29 and their families, the environment in which this care is delivered or additional aspects of care such 30 as the availability of equipment or the timeliness of support: 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 29, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Thank you for your participation 60

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from

End-of-life Management Protocol Offered Within Emergency Room (EMPOWER) – study protocol for a multicentre study

Journal: BMJ Open ManuscriptFor ID peerbmjopen-2019-036598.R1 review only Article Type: Protocol

Date Submitted by the 20-Mar-2020 Author:

Primary Subject Palliative care Heading: on September 29, 2021 by guest. Protected copyright. Secondary Subject Heading: Emergency medicine

PALLIATIVE CARE, ACCIDENT & EMERGENCY MEDICINE, Quality in Keywords: health care < HEALTH SERVICES ADMINISTRATION & MANAGEMENT

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 18 BMJ Open

1 2 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 18 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 TITLE PAGE 4 5 6 End-of-life Management Protocol Offered Within Emergency Room 7 8 (EMPOWER) – study protocol for a multicentre study 9 10 Rakhee Yash Pal1, Win Sen Kuan1,2, Ling Tiah3, Ranjeev Kumar4, Yoko Kin Yoke Wong5, 11 Luming Shi5, Charles Qishi Zheng5, Jingping Lin1, Sufang Liang1, Uma Chandra Segara1, Woon 12 Chai Yong6,7, Noreen Guek Cheng Chan6,7, Mui Teng Chua1,2*, Irwani Ibrahim1,2 13 14 1 Emergency Medicine Department, National University Hospital, National University Health System, Singapore 15 2 Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 16 3 Accident & EmergencyFor Medicine, peer Changi General review Hospital, Singapore only 17 4 Acute and Emergency Care Centre, Khoo Teck Puat Hospital, Singapore 18 5 Singapore Clinical Research Institute, Singapore 19 6 Division of Palliative Care, National University Cancer Institute, Singapore 20 7 Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 21 22 *Corresponding author: 23 Dr Mui Teng Chua 24 Address: Emergency Medicine Department, National University Hospital, 9 Lower Kent Ridge Road, Level 25 4, Singapore 119085 26 Email: [email protected] 27 Telephone number: +65 67723024 28 29 Addresses of all authors: 30 1 Emergency Medicine Department, National University Hospital, National University Health System, 31 Singapore. 9 Lower Kent Ridge Road, Level 4, Singapore 119085

32 http://bmjopen.bmj.com/ 33 2,7 Departments of Surgery and Medicine, Yong Loo Lin School of Medicine, National University of 34 Singapore, Singapore. 10 Medical Dr, Singapore 117597 35

36 3 37 Accident & Emergency Medicine, Changi General Hospital, Singapore. 2 Simei Street 3, Singapore 38 529889 39 4 Acute and Emergency Care Centre, Khoo Teck Puat Hospital, Singapore. 90 Yishun Central, Singapore 40 on September 29, 2021 by guest. Protected copyright. 41 768828 42 43 5 Singapore Clinical Research Institute, Singapore. 31 Biopolis Way, Singapore 138669 44 45 6 Division of Palliative Care, National University Cancer Institute, Singapore. 5 Lower Kent Ridge Road, 46 Singapore 119074 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 End-of-life Management Protocol Offered Within Emergency Room 4 5 (EMPOWER) – study protocol for a multicentre study 6 7 Rakhee Yash Pal1, Win Sen Kuan1,2, Ling Tiah3, Ranjeev Kumar4, Yoko Kin Yoke Wong5, 8 Luming Shi5, Charles Qishi Zheng5, Jingping Lin1, Sufang Liang1, Uma Chandra Segara1, Woon 9 Chai Yong6,7, Noreen Guek Cheng Chan6,7, Mui Teng Chua1,2*, Irwani Ibrahim1,2 10 11 1 Emergency Medicine Department, National University Hospital, National University Health System, Singapore 12 2 Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 13 3 Accident & Emergency Medicine, Changi General Hospital, Singapore 14 4 Acute and Emergency Care Centre, Khoo Teck Puat Hospital, Singapore 15 5 Singapore Clinical Research Institute, Singapore 16 6 Division of Palliative Care,For National peer University Cancerreview Institute, Singapore only 17 7 Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 18 19 ABSTRACT 20 21 Background 22 23 Patients at their end-of-life (EOL) phase frequently visit the emergency department (ED) due to their symptoms, yet 24 the environment and physicians in ED are not traditionally equipped or trained to provide palliative care. This 25 multicentre study aims to measure the current quality of EOL care in ED to identify gaps, formulate improvements 26 and implement the improved EOL care protocol. We shall also evaluate healthcare resource utilisation and its 27 associated costs. 28 29 Methods and analysis 30 31 This study employs a quasi-experimental interrupted time series design using both qualitative and quantitative methods, involving the EDs of 3 tertiary hospitals in Singapore, over a period of 3 years. There are 5 phases in this

40 improvement and protocol changes. Patient, family and staff relevant outcomes from these changes will be measured on September 29, 2021 by guest. Protected copyright. 41 using time series regression. 42 43 Ethics and dissemination 44 45 The study protocol has been reviewed and ethics approval obtained from the National Healthcare Group Domain 46 Specific Review Board, Singapore. The results from this study will be actively disseminated through manuscript 47 publications and conference presentations. 48 49 (290 words, excluding subheadings) 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 18 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 ARTICLE SUMMARY 4 5 Strengths and limitations: 6  This novel multicentre study evaluating EOL care provided in the ED will be able to evaluate the quality of 7 clinical management rendered to EOL patients via a multi-prong approach, from clinical charts and 8 perceptions of family members who were at the bedside 9  By involving our community partners, we shall also be able to identify difficulties faced in caring for these 10 patients in a non-institutional setting 11  Results of this study have the potential to identify current gaps and barriers, and inform emergency 12 physicians globally on best practices in providing EOL care in the ED 13  Some limitations to the study design include response rates to questionnaires, selection bias due to non- 14 probability sampling and information bias from retrospective chart reviews 15 16 For peer review only 17 TRIAL REGISTRATION 18 19 Ethics approval for this study was obtained from the National Healthcare Group Domain Specific Review Board 20 (DSRB reference no: 2018/00838) for waiver of informed consent for the retrospective cohort in pre-Phase 1. 21 Written informed consent will be obtained from all study participants in all other phases. The study protocol has 22 been registered in ClinicalTrials.gov (NCT03906747) on April 8, 2019. 23 Enrolment for the study is currently ongoing. 24 25 KEY WORDS: 26 27 Palliative care, emergency medical services, cost effectiveness, terminally ill 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 29, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 4 BACKGROUND 5 6 In a report by the Institute of Medicine in 1997, a decent or good death, is one that should be free of unnecessary 7 distress and suffering for both patients and their caregivers, and should be in general accord with their wishes, 8 within reasonable consistency with clinical, cultural and ethical standards.1 The World Health Organization (2016) 9 defines palliative care as a multi-dimensional approach to improve the quality of life for dying patients and their 10 families by caring for the physical, psychosocial and spiritual needs. Emergency physicians, being at the forefront of 11 medicine and healthcare delivery, frequently encounter death as a daily occurrence. Globally, it is estimated that at 12 least 35 million (60%) out of 58 million people who are dying annually suffer from advanced illnesses and would 13 benefit from palliative care services for better quality of life.2 In addition, up to 80% of terminally ill patients prefer 3,4 14 to die at home and yet only a small percentage are able to achieve this. 15 16 As the first point of contactFor in the healthcarepeer system, review the emergency department only (ED) is particularly important for 17 provision of appropriate end-of-life (EOL) care, especially when a majority of EOL patients utilise emergency services in the last month of their lives.5 Many patients present to the ED due to symptoms of distress, poorly- 18 controlled symptoms at home, or families who are unable to provide care as a result of physical and mental distress.6 19 Traditionally, the ED has not been regarded as an optimal place for EOL care, due to its inherent nature of 20 overcrowding and chaos, resulting in lack of tranquil infrastructure for families to spend meaningful time with their 21 loved ones during their last moments.4,7 ED physicians are also perceived to lack knowledge and confidence in 22 managing such patients as palliative care contradicts their conventional training on aggressive resuscitation 23 measures. The initiation of palliative care from ED by emergency physicians was found to be at a dismal rate of 24 18% in one study.8 This highlights the need for more specialised training for emergency physicians to adequately 25 manage the increasing number of patients facing progressive illnesses and disabilities towards the end of their lives. 26 27 Palliative care services in Singapore have grown and developed over the past 30 years and is acknowledged as an 28 essential specialty in the healthcare system.9 However, palliative care development in ED is limited10 and caring for 29 patients’ psychosocial and spiritual needs is not the usual goal in the ED. On the other hand, studies have shown that 30 implementation of EOL pathways not only increases the knowledge and confidence of nursing personnel in 31 managing dying patients, but also leads to consistent and improved patient care.11–13 Moreover, palliative care in 14

40 on September 29, 2021 by guest. Protected copyright. 41 We propose this multicentre interrupted time series study to improve the quality of EOL care for actively dying 42 patients in the ED. The specific objectives in this study are to: 43  Systematically measure the current quality of EOL care in three Singapore EDs to identify gaps in 44 management 45  Formulate improvements needed to address these gaps and implement the improved EOL care protocol 46  Measure (a) the quality of care post-implementation of the improved EOL care, and (b) the impact on 47 health services utilisation and associated costs 48 49 METHODS/DESIGN 50 51 Study design 52 53 This is a quasi-experimental interrupted time series study conducted in the EDs of 3 public hospitals (National 54 University Hospital [NUH], Changi General Hospital [CGH] and Khoo Teck Puat Hospital [KTPH]) in Singapore to 55 evaluate the quality of EOL care rendered with measures conducted post-implementation of interventions to assess 56 improvement in EOL care. The project is overseen by biostatisticians and epidemiologists from the Singapore 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 18 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 Clinical Research Institute (SCRI) and site investigators from participating hospitals. The study comprises 5 phases: 4 pre-Phase 1, pilot phase, and Phases 1 to 3. Apart from pre-phase 1 which is retrospective, all other phases are 5 conducted prospectively. Details of each phases with variables and data to be collected are detailed below. 6 7 Study setting and sites 8 9 The public hospitals included in this study, namely NUH, CGH and KTPH, belong to the 3 main healthcare clusters 10 in Singapore – the National University Health System, Singapore Health Services and National Healthcare Group, 11 which serve the country’s western, eastern and northern populations, respectively.16 Each of the 3 hospitals are 12 tertiary centres with annual ED census of more than 100,000 attendances. 13 14 Currently in our local healthcare setting, patients who are at their EOL phase may be managed at the following 15 locations: in their own homes with a dedicated caregiver and home hospice services, in inpatient hospice facilities or 16 in acute care hospitals. TheFor healthcare peer cost and resources review incurred will thus only depend on the setting where EOL care is 17 provided. 18 Pre-Phase 1 (retrospective) 19 20 In pre-Phase 1, we aim to determine the baseline data prior to any implementation of EOL protocols in the 21 respective hospitals to facilitate the assessment of the impact on health services utilisation. Using the hospital 22 database, patients who died within 5 days of ED attendance will be identified. This timeframe is chosen based on 23 unpublished data from a previous study done in NUH, which showed that EOL patients who were admitted have a 24 median length of inpatient stay of 1.4 days with an interquartile range of 0.4 to 5.2 days.17 Chart reviews will be 25 performed to determine if the patient fulfilled the criteria for EOL management. Information on premorbid 26 conditions, any prior referral to palliative services, clinical management, health services utilisation and associated 27 costs will be collected. The period of the chart reviews are as follows: 28  NUH: 01 January 2013 to 31 December 2013 29  CGH: 01 July 2016 to 30 June 2017 30  KTPH: 01 July 2016 to 30 June 2017 31

32 These periods of chart reviews correspond to the 12 months that preceded initiation of any EOL initiative in the EDs http://bmjopen.bmj.com/ 33 of the respective hospitals. Being in the forefront of the public healthcare sector, any patient can present to the EDs 34 in Singapore, either by self-referral/ walk-in, ambulance or referral by another healthcare facility. We often 35 encounter patients who are actively dying from their underlying comorbidities or from a sudden catastrophic event. 36 Before the initiation of any EOL care pathway in each of the EDs, management of symptoms in any ED patients 37 who are actively dying were all based on attending physicians’ discretion. For instance, an EOL patient with 38 increased secretions may not be routinely given hyoscine butylbromide for relief. Recognising this gap, each of the 3 39 EDs have independently established their own workflows for a protocolised care bundle to manage such patients, so that EOL symptoms can be more effectively managed based on an established pathway. Despite this, an 40 on September 29, 2021 by guest. Protected copyright. 17 41 observational study in NUH showed that there is room for EOL care to be improved and optimised. 42 43 Pilot phase (January to April 2019) 44 The pilot phase tests the face and construct validity, and reliability of a newly developed questionnaire for 45 measuring the quality of EOL care in EDs in the Asian context. The questionnaire development takes reference from 46 the Care of the Dying Evaluation (CODETM).18 Participants, who are family members of EOL patients, will be 47 requested to complete the newly developed questionnaire (re-named Care of the Dying Evaluation - Emergency 48 Medicine, CODE-EM), and then interviewed about their experience of completing the questionnaire and to 49 participate in a retest of the questionnaire one month later either by phone or mail. The post-questionnaire interview 50 aims to examine the validity of the questionnaire in terms of language, length, timing and relevance. 51 52 Phase 1 (April 2019 to April 2020 projected) 53 54 At the start of this study, all 3 hospitals have EOL workflows in place that was developed within each department. A 55 brief summary and comparison of the 3 workflows are illustrated in Figure 1. In Phase 1 of the study, quality 56 indicators will be measured prospectively to evaluate the aspects of EOL care rendered to patients under existing 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 workflows. Data of recruited patients from the start of study enrolment until demise or terminal discharge from the 4 hospital will be collected. Their electronic medical records and ED consult notes will be reviewed from time of ED 5 attendance until terminal discharge or death to assess their inpatient progress and discharge outcomes. Apart from 6 chart reviews of the EOL patients, the validated CODE-EM questionnaire will also be conducted with the nominated 7 caregivers of these patients. The quality indicators will be divided into patient-, family- and staff-centred outcomes. 8 9 10 Figure 1. EOL care workflow in the Emergency Departments of National University Hospital (NUH), Changi General Hospital 11 (CGH) and Khoo Teck Puat Hospital (KTPH) 12 13 The patient-centred outcomes to be evaluated include proportion of patients who fits EOL criteria and died within 5 14 days, proportion of patients on EOL pathway who have documentation that patient and/or family are given 15 opportunities to discuss an individualised care plan, and proportion of patients with symptoms who are prescribed 16 with medicines with individualisedFor peerindications for use,review dosage and route ofonly administration. 17 18 In addition, the following data will be collected for all recruited EOL patients: demographic information, clinical 19 presentation, comorbidities, healthcare utilisation, interventions post-EOL, length of hospitalisation, 20 communications with the family, comfort care measures, cost of hospitalisation, prior do-not-resuscitate decisions, 21 disposition (death, terminal discharge home or hospice, admission to inpatient unit) and cause of death. 22 23 Family-centred indicators are evaluated using the validated CODE-EM questionnaire (see Supplementary Material 24 1) to assess the quality of care and the level of support provided to patients and their families in the ED. The key 25 indicators include proportion of participants who perceived that their family member was treated with dignity and 26 respect by doctors and nurses, and the proportion of participants who perceived that they were adequately supported. 27 28 For staff-centred indicators, we aim to determine the level of healthcare providers’ knowledge and perception of the 29 EOL care using a locally-designed anonymous survey (see Supplementary Material 2) conducted monthly in the first 30 3-months, followed by 6 and 12 months later. This survey will be self-administered. Information regarding the work 31 experience, existing level of palliative care training and designation will also be collected. The main components to be evaluated are level of confidence in communications and providing emotional support, knowledge, clarity of 32 http://bmjopen.bmj.com/ 33 nurses’ and doctors’ roles, conduciveness of physical environment and satisfaction of overall care quality rendered 34 in managing EOL patients in the ED. 35 Phase 2 (March to April 2020 projected) 36 37 Phase 2 will commence at the end of Phase 1 based on the results obtained and will encompass focus group 38 discussions (FGDs), revision of EOL care components and implementation of improved EOL care in the 3 39 institutions. There will be separate FGDs for healthcare professionals, family members and primary care providers 40 (such as general practitioners and home hospice healthcare workers) conducted to identify the gaps for improvement on September 29, 2021 by guest. Protected copyright. 41 and areas that were done well in the existing EOL care paths. Recommendations to improve the quality of the 42 existing EOL care will also be sought. 43 44 By combining the qualitative and quantitative results of Phases 1 and 2, the study investigators from the 3 hospitals 45 will assimilate all data and revise the current protocol accordingly. Combined and in-house staff training on 46 palliative care and improved EOL management workflow will ensue thereafter. 47 48 Phase 3 (May 2020 to May 2021 projected) 49 50 In Phase 3, new and improved EOL workflows will be rolled out in all 3 hospitals. The quality indicators and 51 outcomes to be evaluated will be the same as in Phase 1 (as detailed above). The results obtained in Phase 3 will 52 then be compared to that in Phase 1. Staff-centred indicators will be mapped against the combined and in-house 53 training sessions to evaluate for any changes. 54 55 Project materials 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 18 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 Permission was obtained from the original developers of the CODE-EM questionnaire to adapt it for local and ED 4 contexts (see Supplementary Material 1), which was validated during the pilot phase. To evaluate staff-centred 5 indicators, a locally designed staff survey was used (see Supplementary Material 2). The staff survey was developed 6 by ED attendings and nurses trained in palliative care for the purpose of evaluating quality of care rendered during 7 the initiation of the EOL workflow in its early stages in NUH in 2013. The staff survey was piloted and improved 8 upon during those initial stages, not within the study period. 9 10 Patient selection 11 12 All the following inclusion criteria must be fulfilled: 13 • Actively dying patient or high likelihood of mortality within this admission (based on attending physician’s 14 judgement using available clinical data such as vital signs, mental status and biochemical or radiological 15 investigations, if available) 16 • Family acceptsFor that the goals peer of care are provisionreview of comfort, symptom only relief and respect of dignity 17 • Patient is not a candidate for cardiopulmonary resuscitation, endotracheal intubation or transfer to the intensive care unit 18 • Any of the life-limiting conditions: chronic frailty with poor functional state and limited reversibility 19 (Karnofsky Performance Scale [KPS] <40%)19; chronic severe illness with poor prognosis (terminal cancer, 20 end-stage renal failure [refusal or withdrawal of dialysis], end-stage respiratory, heart or liver disease, 21 advanced neurological disease including dementia, Parkinson’s disease and severely disabling stroke); or, 22 acute severe catastrophic conditions and at risk of dying with complications that are not reversible, as 23 subject to the treating clinician’s judgment 24 25 We excluded the following subjects: vulnerable population (e.g. prisoners, pregnant women); refusal to participate; 26 patients who have been recruited, or had declined participation, during the previous ED attendance(s); patients in 27 peri-arrest state; and/or family members who are not available or present at the patient’s bedside. 28 29 Patient and public involvement 30 31 There was no patient or public involvement in the design and conduct of this study.

32 http://bmjopen.bmj.com/ 33 Sample size calculation 34 35 Preliminary data from 2014 to 2017 in NUH for calculation indicated an average of 7 patients who fulfilled the EOL 36 criteria in NUH every month. We expect to receive approximately 15 eligible EOL patients per month from the 3 37 centres and about 195 eligible patients in Phase 1 (13 months) and Phase 3 (13 months), respectively. 38 39 Based on NUH’s existing data, the current correct identification rate is 24.4%. We performed a power analysis to estimate the expected power with various combinations of recruited subjects and observed improvement. We could

40 on September 29, 2021 by guest. Protected copyright. achieve an 80% power with an alpha of 0.05 for a two-sided test, if we were to: 41 1) Recruit 200 subjects for each phase and expected to observe an absolute difference ≥ 13%; 42 2) Recruit 180 subjects for each phase and expected to observe an absolute difference ≥ 14%; 43 3) Recruit 160 subjects for each phase and expected to observe an absolute difference ≥ 15%; 44 4) Recruit 140 subjects for each phase and expected to observe an absolute difference ≥ 16%. 45 46 Outcomes of interest and statistical analyses 47 48 (i) Pilot phase analysis planning 49 50 The post-questionnaire interviews about the experience of completing the questionnaire will be transcribed verbatim 51 and analysed using a content analysis framework. Data will be refined into specific categories, with words and 52 phrases of shared meaning. A randomised selection of interview transcripts will be independently reviewed by a 53 second investigator, not directly involved in the data collection, to check for coding, and any discrepancies will be 54 discussed with a third investigator. The stability of the developed questionnaire over time will be assessed using the 55 following measures: percentage agreement; κ statistic (Cohen’s for nominal response options and weighted for 56 ordinal response options); and Spearman’s correlation coefficient (for ordinal data). The criteria for good stability 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 over time is defined as percentage agreement >70%; κ>0.6; and r>0.7 and moderate stability over time as percentage 4 agreement >30%; κ>0.40; and r>0.3. Confirmatory factor analysis will be used to assess construct validity. The 5 suitability of questions will be examined by inspection of the correlation matrix and the Goodness of Fit Index. 6 7 (ii) Statistical analysis for other phases 8 9 Descriptive statistics will be obtained about the baseline characteristics of participants. Continuous variables will be 10 presented as mean ± standard deviation while categorical variables will be reported as absolute numbers and 11 percentages. Differences regarding outcomes will be analysed using independent-samples t-test or Mann-Whitney U 12 test, or paired t-test for pre- and post-EOL measurements. Symptom burden (e.g. differences in frequency of 13 symptoms) and CODE-EM questionnaire will be calculated. 14 15 Time series regression will be used to analyse the results from staff survey, using 7 data points (i.e. collected at 16 every 2 months) from eachFor phase. Thepeer best fit pre- reviewand post-programme revisiononly lines will be estimated by using 17 linear regression, and auto-correction will be adjusted for by using maximum likelihood methods where appropriate. The change in the slopes of the regression lines will be estimated. Data will be analysed using Stata version 15 18 (Stata Corp, College Station, Texas, USA). 19 20 (iii) Health economic impact analysis 21 22 This study will also evaluate the financial viability of EOL care by conducting a health economic evaluation, from a 23 hospital system perspective. The costing for our health economic evaluation has two components. First, for costs 24 directly related to services to patients, itemised data and information on resource utilisation associated with patient 25 management will be recorded for each participant, including initial management in ED, services associated with 26 protocol implementation and follow-up. Second, costs related to protocol implementation including extra 27 administration and medical staff will be evaluated as well, according to the required staff levels and time needed. 28 The potential impact of protocol implementation on medical and administrative management resources will be 29 studied by comparing with historical data from chart reviews. After identifying the extra care components associated 30 with study protocol implementation (e.g. accessing the on-call palliative care specialist), for base case analysis, 31 private rates of each service will be applied to calculate the average cost and savings. In Singapore, the billing in our

32 public hospitals can be broadly categorised into “private rates” and “subsidised rates” (after receiving government http://bmjopen.bmj.com/ 33 subsidies). We will use “private rate” of each service item to best estimate the costs. 34 35 The health economic analysis will consist of 2 components: (1) a cost-benefit analysis to assess the cost-benefit 36 ratio, which will be presented as “for every 1 extra dollar spent on the protocol implementation, how much savings 37 can be realised by the potential reduction in hospital admissions and length of hospital stay”; and (2) a cost- 38 effectiveness analysis to estimate the incremental cost-effectiveness ratio. To address possible variations in patient 39 medical conditions and the outcomes, such as variations in compliance, and length of hospital stay, a series of one- way sensitivity analyses will be conducted to address the impact of each parameter's uncertainty and assess the

40 on September 29, 2021 by guest. Protected copyright. robustness of study estimations for generalisability. 41 42 ETHICS AND DISSEMINATION 43 44 Ethics approval for this study was obtained from the National Healthcare Group Domain Specific Review Board 45 (DSRB reference no: 2018/00838) for waiver of informed consent for the retrospective cohort in pre-Phase 1. 46 Written informed consent will be obtained from all study participants in all other phases. For EOL patients who have 47 the mental capacity for informed consent, their consent for participation will be sought and will not be overridden by 48 their family members. For patients who have impaired cognition, informed consent will be sought from their legally 49 acceptable representative in accordance to DSRB’s guidelines. The study protocol has also been registered in 50 ClinicalTrials.gov (NCT03906747). The findings from this study will be disseminated locally and internationally 51 through manuscript publications in peer-reviewed journals and conference presentations at national and international 52 platforms. 53 54 DISCUSSION 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 18 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 To the best of our knowledge, this is the first multicentre study using both qualitative and quantitative methods and 4 focusing on evaluation of EOL care provided in the ED. Additionally, apart from focusing on patients with chronic 5 irreversible illnesses such as advanced cancer, our study cohort includes other death trajectories such as sudden 6 death due to an acute catastrophic event. This is an area where information on palliative care is lacking. 7 8 Our study design has distinct advantages over the traditional epidemiological cohort (before-and-after) and case- 9 control designs that are limited by confounding and lack of usefulness in intervention studies. The interrupted time 10 series study design would be able to detect changes that are delayed or intermittent and can assess if the change is 11 permanent or temporary. The design is simpler without the need for randomisation, which is not possible in EOL 12 research. There is also the ability to control for confounding variables and regression to the mean. 13 14 Furthermore, our study design allows us to obtain both qualitative and quantitative knowledge with regard to current 15 quality of care and identify areas for improvement. The information obtained from intimate FGDs with the family 16 members and our communityFor healthcare peer partners would review be consequential onlyin providing in-depth information on the 17 current quality of care and difficulties faced in order to guide changes in protocols and policies. The unique setting of a multicultural Singapore society may also provide some insights into previously under-investigated cultural 18 beliefs and values that can be a focus of future qualitative studies in the region. 19 20 Despite the strengths of our study, the design presents some inherent limitations. First, the chart reviews of the 21 retrospective cohort in Pre-Phase 1 may have information bias. We aim to mitigate this by looking through all 22 available paper and electronic medical records to reduce missing information. Second, the response rates from 23 questionnaires and staff surveys may also pose a challenge. In order to encourage participation, small tokens of 24 appreciation and incentives will be given out to study participants. Third, the use of non-probability sampling and 25 enrolment when research assistants are around during office hours may also introduce a degree of selection bias. 26 This is however necessary due to economic constraints that preclude round-the-clock staffing of research personnel. 27 However, we have mitigated this by screening through ED attendances beyond office hours and over the weekend, 28 and contacting eligible family members in the wards whenever appropriate for invitation to participate. 29 30 In conclusion, the results of our study would have the potential to identify current barriers and inform emergency 31 physicians globally on the best practices in providing EOL care in the ED for patients who are at the last stages of

32 their lives from a myriad of illnesses. http://bmjopen.bmj.com/ 33 34 (3733 words) 35 36 DECLARATIONS 37 38 Ethics approval and consent to participate 39 Ethics approval for this study was obtained from the National Healthcare Group Domain Specific Review Board (DSRB reference no: 2018/00838) for waiver of informed consent for the retrospective cohort in pre-Phase 1.

40 on September 29, 2021 by guest. Protected copyright. Written informed consent will be obtained from all study participants in all other phases. The study protocol has also 41 been registered in ClinicalTrials.gov (NCT03906747). 42 43 Consent for publication 44 Not applicable. 45 46 Availability of data and materials 47 Not applicable. 48 49 Competing interests 50 The authors declare that they have no competing interests. This study is funded by a major government funding 51 body (National Medical Research Council, Ministry of Health, Singapore) and there is no funding or assistance from 52 any commercial organisation. 53 54 Funding 55 This work is supported by a grant from the National Medical Research Council Health Services Research Grant 56 (grant number: MOH-HSRG-EoL17Ju-0001). The funding body provides funds for employment of research 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 assistants, engagement of statistical support from Singapore Clinical Research Institute and pays for all study-related 4 expenses such as reimbursements of study participants and manuscripts processing fees. It is not involved in the 5 design of the study, and collection, analysis, and interpretation of data, and in writing of the manuscript. 6 7 Author contributions 8 RYP conceived the study, designed the final study protocol, provided the domain knowledge expertise and is the 9 principal investigator of the grant; WSK helped in the design of the final study protocol, contributed to the technical 10 design and revised the initial manuscript draft; LT and RK helped in the design of the final study protocol and are 11 coordinating site investigators; YKYW, LS and CQZ contributed to the technical design and provided biostatistical 12 and epidemiological support; JL, SL, UCS, WCY and NGCC provided clinical input and perspectives to the 13 qualitative aspects of the study; MTC helped in the design of the final study protocol, drafted the initial manuscript 14 and coordinated ethics approval; II helped in the design of the final study protocol, had a part in the technical design 15 and coordinated collaboration among investigators from all institutions. All authors read and approved the final 16 manuscript. For peer review only 17 Acknowledgements 18 Not applicable. 19 20 Author details 21 1 Emergency Medicine Department, National University Hospital, National University Health System, Singapore 22 2 Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 23 3 Accident & Emergency Medicine, Changi General Hospital, Singapore 24 4 Acute and Emergency Care Centre, Khoo Teck Puat Hospital, Singapore 25 5 Singapore Clinical Research Institute, Singapore 26 6 Division of Palliative Care, National University Cancer Institute, Singapore 27 7 Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 28 29 30 REFERENCES 31

32 1. Field MJ, Cassel CK. Approaching death: improving care at the end of life. N Engl J Med 1998; 339: 274– http://bmjopen.bmj.com/ 33 275. 34 2. Stjernswärd J, Foley KM, Ferris FD. The Public Health Strategy for Palliative Care. J Pain Symptom 35 Manage 2007; 33: 486–493. 36 3. Tang ST. When death is imminent: where terminally ill patients with cancer prefer to die and why. Cancer 37 Nurs 2003; 26: 245–251. 38 4. Gloss K. End of life care in emergency departments: a review of the literature. Emerg Nurse 2017; 25: 29– 39 38. 5. Smith A, McCarthy E, Weber E, et al. Half of older Americans seen in emergency department in last month

40 on September 29, 2021 by guest. Protected copyright. of life; most admitted to Hospital, and many die there. Health Aff 2012; 31: 1277–1285. 41 6. Krikorian A, Limonero JT, Maté J. Suffering and distress at the end-of-life. Psychooncology 2012; 21: 799– 42 808. 43 7. Beckstrand RL, Wood RD, Callister LC, et al. Emergency nurses’ suggestions for improving end-of-life care 44 obstacles. J Emerg Nurs 2012; 38: e7. 45 8. Kistler EA, Sean Morrison R, Richardson LD, et al. Emergency department-triggered palliative care in 46 advanced cancer: proof of concept. Acad Emerg Med 2015; 22: 237–239. 47 9. Lien Centre for Palliative Care, Duke-NUS Graduate Medical School. Report on the National Strategy for 48 Palliative Care., https://www.singaporehospice.org.sg/en/wp- 49 content/uploads/2018/08/Report_on_National_Strategy_for_Palliative_Care-5Jan2012.pdf (2011, accessed 50 13 June 2017). 51 10. Chan GK. End-of-life models and emergency department care. Acad Emerg Med 2004; 11: 79–86. 52 11. Selman L, Robinson V, Klass L, et al. Improving confidence and competence of healthcare professionals in 53 end-of-life care: An evaluation of the ‘Transforming End of Life Care’ course at an acute hospital trust. BMJ 54 Support Palliat Care 2016; 6: 231–236. 55 12. Luhrs CA, Penrod JD. End-of-life care pathways. Curr Opin Support Palliat Care 2007; 1: 198–201. 56 13. Weng TC, Yang YC, Chen PJ, et al. Implementing a novel model for hospice and palliative care in the 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 18 BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 1 2 3 emergency department. Med (United States) 2017; 96: e6943. 4 14. Aldridge MD, Kelley AS. The myth regarding the high cost of end-of-life care. Am J Public Health 2015; 5 105: 2411–2415. 6 15. Threapleton DE, Chung RY, Wong SY, et al. Care towards the end of life in older populations and its 7 implementation facilitators and barriers: a scoping review. J Am Med Dir Assoc 2017; 18: 1000-1009.e4. 8 16. Poon CH. Public healthcare sector to be reorganised into 3 integrated clusters, new polyclinic group to be 9 formed. The Straits Times, https://www.straitstimes.com/singapore/health/public-healthcare-sector-to-be- 10 reorganised-into-3-integrated-clusters-new (2017, accessed 16 October 2019). 11 17. Chor WPD, Wong SYP, Ikbal MF bin M, et al. Initiating End-of-Life Care at the Emergency Department: 12 An Observational Study. Am J Hosp Palliat Med 2019; 36: 941–946. 13 18. Mayland CR, Lees C, Germain A, et al. Caring for those who die at home: The use and validation of ‘Care 14 Of the Dying Evaluation’ (CODE) with bereaved relatives. BMJ Support Palliat Care 2014; 4: 167–174. 15 19. Karnofsky D, Burchenal J. The clinical evaluation of chemotherapeutic agents in cancer. Eval Chemother 16 agents 1949; 191–205.For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

32 http://bmjopen.bmj.com/ 243x175mm (300 x 300 DPI) 33 34 35 36 37 38 39

29 30 In your opinion, during the patient’s stay in the ED, did the Yes, all of Yes, some No, s/he did not appear to 31 7. 32 patient appear to be in pain? the time of the time be in pain 33 In your view, did the doctors and nurses in the ED do enough Yes, all of Yes, some No, not Not applicable, 34 8. s/he was not in to help relieve the pain? the time of the time at all 35 pain 36 http://bmjopen.bmj.com/ 37 38 In your opinion, during the stay in the ED, did the patient Yes, all of Yes, some No, s/he did not appear 9. 39 appear to be restless? the time of the time to be restless 40 41 In your view, did the doctors and nurses in the emergency Yes, all of Yes, some No, not Not applicable, 10. s/he was not 42 department do enough to help relieve the restlessness? the time of the time at all restless 43

44 on September 29, 2021 by guest. Protected copyright. 45 No, s/he did not have a In your opinion, during the stay in the ED, did the patient Yes, all of Yes, some 46 11. noisy rattle to the the time of the time 47 appear to have a “noisy rattle” when breathing? breathing 48 No, Not applicable, 49 In your view, did the doctors and nurses in the ED do enough Yes, all of Yes, some s/he did not have 12. not at the time of the time a noisy rattle to 50 to help relieve the “noisy rattle” when breathing? all 51 the breathing 52 53 In your opinion, during the stay in the emergency 54 Yes, all of Yes, some No, s/he did not have 13. department, did the patient appear to have difficulty 55 the time of the time difficulty breathing breathing? 56 In your view, did the doctors and nurses in the emergency Not applicable, 57 Yes, all of Yes, some No, not s/he did not 14. department do enough to help relieve the breathing 58 the time of the time at all have difficulty 59 difficulty? breathing 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 18 BMJ Open

1 2 During the time in the ED, the patient's care and treatment Strongly Strongly 3 15. Agree Neutral Disagree agree disagree 4 was discussed with you and/or the family. BMJ Open: first published as 10.1136/bmjopen-2019-036598 on 28 April 2020. Downloaded from 5 The healthcare team at the ED explained the patient's Very Very 6 16. condition and treatment in a way you found easy or difficult Easy Neutral Difficult easy Difficult 7 to understand. 8 9 How would you assess the overall level of emotional support Very 17. Poor Fair Good Excellent 10 given to you by the ED healthcare team? poor 11 12 The ED healthcare team discussed the patient's religious or Strongly Strongly 18. Agree Neutral Disagree 13 spiritual needs. agree disagree 14 15 16 17 19. Were you told that the patient would be likely to die soon? Yes No 18 For peer review only 19 19a) If yes, who told you s/he was likely to pass away soon? 20 21 22 23 Did a member of the ED healthcare team talk to you about 24 20. what to expect during the dying process (e.g. what Yes No 25 symptoms may arise)? 26 If yes → Was the discussion about what to expect during the 27 20a) Yes No 28 dying process helpful? 29 If no → Would a discussion about what to expect during the 30 20b) Yes No 31 dying process have been helpful? 32 33 34 In your view, the patient was treated with respect and Strongly Strongly 35 21. Agree Neutral Disagree agree disagree 36 dignity by the ED doctors. http://bmjopen.bmj.com/ 37 In your view, the patient treated with respect and dignity by Strongly Strongly 22. Agree Neutral Disagree 38 the ED nurses. agree disagree 39 40 41 42 Overall, in your opinion, were you adequately supported 23. Yes No 43 during the patient's stay in ED? 44 on September 29, 2021 by guest. Protected copyright. 45 46 47 48 ------THE END ------

49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 18

18 For peer review only 19 Section A: Demographics 20 Please put a tick the box below to indicate your choice. 21 22 1. Please tell us your gender 23  Male 24  Female 25 26

27 2. Please tell us your profession 28  Medical (if medical, please answer questions 3 and 4) 29  Nursing (if nursing, skip questions 3 and 4) 30 31 3. If you are a medical doctor, please select your designation 32  Associate consultant and above 33  Senior resident 34 35  Resident physician with supervisory role 36  MOPEX MO http://bmjopen.bmj.com/ 37  Clinical associate/ resident physician with MO role 38  EM resident 39  Non-EM resident (e.g. Internal Medicine, Radiology) 40  Locum doctor 41 42 43 4. If you are doing a rotation in the Emergency Department, please indicate the duration of your 44 posting on September 29, 2021 by guest. Protected copyright. 45  1 month 46  2 months 47  3 months 48  6 months 49 50  For other duration, please state in months: ______51 52 5. Your years of clinical experience from graduation 53  < 2 54  2-5 55  5-10 56

57  10-20 58  > 20 59 60 6. Your palliative care training (Please select the most appropriate option)

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 18 BMJ Open

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 18

6 7 Section B: Care provided by the emergency department 8 Strongly Strongly 9 Disagree Neutral Agree Disagree Agree 10 11 8. I am confident of managing a dying patient in the emergency 12 1 2 3 4 5 13 department. 14 15 9. I am satisfied by the overall 16 standard of palliative care 17 provided in the emergency 1 2 3 4 5 18 department. For peer review only 19 20 10. I am clear on the roles and 21 responsibilities of the 22 emergency department nurses 23 1 2 3 4 5 24 and doctors in the care of a 25 dying patient. 26 27 11. The physical environment in the 28 emergency department is 29 conducive for the provision of 1 2 3 4 5 30 care in a dying patient. 31 32 12. The overall experience of a 33 dying patient and their family in 34 the emergency department is 1 2 3 4 5 35 satisfactory. 36 http://bmjopen.bmj.com/ 37

38 39 Section C: Communication with patients and relatives 40 Strongly Strongly 41 Disagree Neutral Agree Disagree Agree 42 43 13. I am able to give emotional 44 support to a dying patient and 1 2 3 4 5 on September 29, 2021 by guest. Protected copyright. 45 their family. 46 47 14. I routinely discuss a patient’s 48 religious or spiritual needs with 1 2 3 4 5 49 the family. 50 51 52 53 54 55 56 57 58 59 60

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 18 BMJ Open

1 2 3

52 53 54 55 56 57 58 59 Thank you for your participation 60

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml