Supporting dying in place in Hong Kong – Retrospective analysis of a cohort of terminally ill patients in public sector with a preference for dying at home

Jeffrey SC Ng 吳常青, Pui-chi Chiu 趙佩芝, Po-tin Lam 林寶鈿, Ka-yee Lee 李家儀, Kai-ming Li 李啟明

HK Pract 2021;43:80-88

Summary

Introduction: The prevalence of dying at home in Hong Kong is extremely low. A cross-specialty multidisciplinary program supporting dying at home was established since 2011 in a cluster of regional public hospitals.
Objective: This study aimed to review the profile of patients in the program, to explore factors favouring home death, interventions delivered at home, reasons for admission for hospital death and events after home death.
Methods: patients who joined the programme before 2019 were included. Information was searched through the electronic medical records of Hospital Authority
Results: 53 patients were recruited, of median age 81 (range 29 - 102), 66% female and 50.7% with cancer diagnoses. Most (96.2%) patients had family caregiver(s) and 49.1% also had foreign domestic helper (FDH). Thirty-two patients (60.4%) achieved home death. Factors favouring home death included having FDH (62.5% and 28.6%, p = 0.016), nursing visits ≥ 5 times per week (71.9% and 42.9%, p = 0.035) and having less emergency admission and admission to palliative care (PC) wards in the programme (3.1% and 28.6%, p = 0.012; 3.1% and 23.8%, p = 0.031). Interventions at home covered symptom control, psycho-spiritual support and caregiver education. Having refractory symptoms was the main reason (85.7%) for hospital death. For patients who died at home, 87.5% (28/32) received cardio-pulmonary resuscitation (CPR) by emergency rescue personnel (ERP) upon transfer to hospital.

摘要

導言: 香港的居家離世率極低。從2011年起，一組地區公立醫院設立了一項支持居家離世的跨專業多學科專案。
目的: 本研究旨在對專案患者的基本情況進行梳理，探討居家離世的有利因素、可在家開展的介入措施、在醫院離世的原因以及居家離世後的事件。
方法: 納入了2019年之前加入本專案的患者。資訊通過檢索醫院管理局的電子病歷獲得。
結果: 共納入了53名患者，年齡中位數81歲(29歲–102歲)，女性佔66%，癌症患者佔50.7%。大多數患者(96.2%)有家人照護，49.1%的患者還有外籍家庭傭工照護。三分之二的患者(60.4%)實現了居家離世。有利於居家離世的因素包括：聘用外籍家庭傭工(62.5%和28.6%，p = 0.016)，護士家訪≥每週5次(71.9%和42.9%，p = 0.035)，以及在專案中較少入住急症和紓緩治療病房(3.1%和28.6%，p = 0.012；3.1%和23.8%，p = 0.031)。在家中開展的介入措施包括症狀管理、心理精神支持以及照護者教育培訓。在醫院離世的主要原因為有難治性症狀(85.7%)。在居家離世的患者中，87.5% (28/32)在轉運至醫院的途中接受過緊急救援人員的心肺復甦術。
結論: 居家離世在香港是可以實現的。在家中提供專業支援是重要的決定因素之一。因法律障礙而導致緊急救援人員違背患者意願進行無效心肺復甦術的情況尚待克服。

Introduction

An important element of End-Of-Life Care (EOLC) is to let people with terminal illness to receive care and to die in their preferred place. The percentage of people being able to die in place has been proposed to be an element of good death¹ and a quality indicator in palliative care/ EOLC in some countries.² Family satisfaction with EOLC is also strongly associated with their relatives dying in their preferred place.³ Among different locations of death, home is commonly acknowledged as the preferred place.⁴ Dying peacefully in one’s bed at home has also been recognised as a blessing in traditional Chinese belief.⁵

About 90% of Hong Kong (HK) people died in government hospitals⁶ , but a recent HK survey suggested that 31.2% of general public chose home to be the preferred place of death.⁷ The preference for home death may decrease with progression of illness⁸ , consistent with local findings that only 13-19% of patients receiving palliative care preferred to die at home.^9,10 Nonetheless, an earlier study in HK reported extremely low prevalence – only 6 in 1300 cancer patients died at home as preferred.¹¹ Barriers for dying at home in HK are multifold, including the inadequate care available at home, the lack of community professional support as well as the legal and logistic concern after home death.¹²

As an attempt to support care-in-place and dying-in-place, a cross-specialty multidisciplinary programme has been established in regional public hospitals in Kowloon East Cluster (KEC) of Hospital Authority (HA) since July 2011. This dying-at-home programme consists two parts: 1. care-in-place until death supported by a community professional team with expertise in palliative care and 2. dying-in-place facilitated by verification of death in the Accident and Emergency Department (AED).

The professional support at home is provided by the KEC Virtual Ward service, which aims to deliver “hospital-at-home” to patients with high risk of hospital admissions¹³ and was found to be effective in reducing unplanned emergency medical readmissions and in improving the quality of life in frail older patients after discharge.¹⁴ Terminally ill patients are prone to repeated hospital admissions and are one of the targeted service recipients of this service. With collaboration with community nursing service, community support up to daily home visit can be arranged as needed. The service provides on-site medical and nursing support at home and enquiry hotline from 9am to 5pm on Monday to Sunday. Clinical admission to PC ward could be arranged so as to avoid AED attendance. Team members consist of community nurses, a specialist palliative care (PC) nurse, a PC physician and a geriatrician.

The verification of death after a patient passed away at home was smoothed out by a workflow established between AED and PC team. This workflow, which is named “Palliative Care Last Journey”, ensures that the Certificate of the Cause of Death (Form 18) would be issued to the patient who died in AED, such that the deceased body would not be sent to public mortuary as a reportable death.¹⁵ The AED is informed beforehand of the list of patients recruited in the dying-at-home program. Recruited patients would be given a letter with attention to AED stating patient’s terminal illness and the preference for dying at home and refusal of cardio-pulmonary resuscitation (CPR). The standardised HA Advance Directives (AD) forms¹⁶ (if the patient is mentally competent) and Do-Not-Attempt Cardio-Pulmonary Resuscitation (DNACPR) forms for non-hospitalised patients (implemented since October 2014)¹⁷ would be signed in the programme. CPR would be avoided in AED when patients were transferred from home after death. The patient would be certified dead by the AED doctors and the last office would be performed by the AED nurses. The body is retained in the hospital mortuary and the Certificate of the Cause of Death (Form 18) would be completed by the PC physicians.

People with advanced diseases receiving palliative care service were referred for the dying-at-home program. The criteria for enrolment are: (i) a clearly stated preference for dying at home, (ii) such preference accepted by the family and/or caregivers, (iii) having at least one full-time caregiver and (iv) living in the catchment area of United Christian Hospital (UCH) (later extended to Tseung Kwan O Hospital in mid-2018).Patients with difficult symptom control or inadequate support from caregivers would be excluded upon initial assessment. Since the establishment of this programme, a significant portion of recruited patients have achieved dying at home.

Objectives

This retrospective study would analyse the profiles of the cohort of patients recruited into the dying-at-home programme. By comparing the characteristics of the subjects who died at home with those in hospital, this study attempts to identify factors associated with the place of death. This study will also explore the interventions delivered during terminal care at home, the reasons for final admissions in those who died in hospital as well as the events after home death.

Methods

It is a retrospective cohort study. Patients who are terminally ill and have been recruited in the above-mentioned dying-at-home programme before 2019 and living in UCH catchment area will be included in this study. There is no exclusion criterion.

Information on patient’s place of death and data on patient’s profile are extracted by review of the electronic medical records of HA by the investigators. Data collected are based on conceptual models on determinants of place of death in systematic reviews^18,19, which include individual factors (including demographics, patient’s preference), social factors (including household members, characteristics of caregiver, housing and social assistance), disease factors (including diagnosis, functional status and symptom profiles) and health care factors (including care interventions at home and healthcare utilisation). Functional status is assessed by using Palliative Performance Scales (PPS) in this study. PPS uses five observer-rated parameters (ambulation, activity and evidence of disease, self-care, intake, and level of consciousness). It is a reliable and valid tool and PPS score has been found to be a strong predictor of survival in patients receiving palliative care.^20,21 Relatives' Stress Scale (RSS) is used to measures caregiving stress of family caregivers. It is a validated self-rated 15-item scale with total score ranging 0 to 60 and score </= 23 suggesting low risk, > 30 suggesting high risk.^22,23

Information on intervention in terminal care and communication with patient and caregivers at home are also based on documentation in the electronic medical records. For the reasons of the last admissions in those who died in hospital and the events after patients dying at home, including any CPR by emergency rescue personnel (ERP) or in AED, body being sent to public mortuary and referral to coroner, they are retrieved from AED records with supplemented information from community nurses in dying-at-home programme.

The Statistical Package for the Social Sciences (SPSS) version 21 is used for statistical analysis. Descriptive statistics will be used to summarise the patient's health conditions, socio-demographic characteristics and treatments/ interventions received after recruitment into the dying-at-home programme until death. Normality in continuous variables is assessed by Kolmogorov-Smirnov Test. Comparison between the two cohorts (those dying at home and those in wards) will be done by Pearson Chi-square (or Fisher’s exact test (FET) if the sample value is less than 5) and Mann–Whitney U test. A significance level with a p value < 0.05 will be used for all analyses.

This study received the approval from Kowloon East Cluster/ Kowloon Central Cluster Research Ethics Committee, HA (Reference number: KC/KE-18-0180/ ER-1).

Result

Totally 53 patients were enrolled in the dying-at-home program in the period from Jul 2011 to Nov 2018 and were all included in this study. Table 1 summarises the information on patient demographics, social background and disease characteristics. Most patients (96.2%) had full-time family caregivers and nearly half (49.1%) had foreign domestic helpers (FDH) as caregivers living with patients. The overall caregiver stress by RSS was low with median score of 12 (Interquartile range (IQR) 6 – 23), as score ≤ 23 suggesting low risk.^22,23 Half of the patients had cancer diagnosis, including colorectal 8 (29.6%), lung 6 (22.2%), liver 3 (11.1%), breast 3 (11.1%) and other types 7 (25.9%) of malignancy. Functional state was poor with PPS of median 40, i.e. mainly in bed, unable to do most activities and mainly requiring assistance in self-care. The median number of symptoms were 6 (IQR 5 – 7), with generalised weakness and poor oral intake being the most prevalent (in >90% of patients).

Table 2 shows the healthcare utilisation in patients recruited in dying-at-home program. Most (94.3%) patients had hospitalisation in the 6 months before recruitment. Median duration of joining the dying-at-home program was 12 days (IQR 8 – 27.5 days). More than half of the patients received doctor home visit once or more per week and nursing home visit 5 times or more per week within the programme. Around 10% of patients still had hospitalisation despite intensive support during the dying-at-home programme.

Thirty-two patients (60.4%) in the dying-at-home programme achieved home death and the rest (39.6%) died in hospital. Comparing those who died at home with those in hospital, factors favouring home death included having FDH as caregiver living with patient (62.5% and 28.6% respectively, p = 0.016, FET), received nursing visits ≥ 5 times per week (71.9% and 42.9%, X2 = 4.463, df = 1, p = 0.035) and having less emergency admission and admission to PC wards during the programme (3.1% vs. 28.6%, p = 0.012, FET, and 3.1% and 23.8%, p = 0.031, FET, respectively). Other demographic, social, disease and health care variables are not found to have a statistically significant association with the place of death.

Table 3 shows the interventions and communication in the dying-at-home programme. All except one patient who died upon the first doctor visit had received anticipatory prescription for symptom management. Around 30% of patients used systemic opioids, including tramadol (oral/ subcutaneous injection/infusion), morphine (oral) and fentanyl (as transdermal patch). One third of patients required continuous subcutaneous infusion (CSCI) of medication, which included tramadol, haloperidol, metoclopramide, hyoscine butylbromide and furosemide. No strong opioids were administered via CSCI route. Caregiver education was delivered in all cases, which included symptom management, use of medications, skills in bodily care as well as explanation on signs of dying, potential catastrophic events and practical procedures after home death. All caregivers/ families received psycho-spiritual support and bereavement care from community nurses in dying-at-home programme and from the multidisciplinary PC team. Advance care planning was reviewed in all patients and HA standardised DNACPR forms for non-hospitalized patients were signed in all after its rolling out in October 2014. For the recorded reasons for choosing home death, more than half mentioned that their needs were better attended and there was more time with family and more freedom at home. There is no statistically significant difference identified in the items listed in Table 3 between patients who died at home and those in hospital.

Among the patients who died in hospital, majority (85.7%) was admitted for refractory physical symptoms in which shortness of breath being the most common (38.1%) (Table 4). In 3 (14.3%) patients, families changed mind and preferred dying in hospital when patients' condition further declined.

Table 5 shows the events in the cohort of patients who died at home. Despite the presence of documents stating the preference against CPR, families of 28 (87.5%) patients reported that CPR was performed by ERP before arrival to AED. CPR was not initiated or continued in AED. There was no referral to Coroner or body being sent to public mortuary.

Discussion

Despite the extremely low prevalence of home death in HK, more than half of patients in this study died at home. This suggested that home can be an achievable option in Hong Kong as the preferred place of death.

Factors associated with place of death

The criteria for recruitment in the program are known to be factors associated with an increased likelihood of home death, including the preference of patient^18,19, caregiver^18,24 and family¹⁹ for dying at home, as well as the presence of relatives¹⁸ or informal caregiver¹⁹ living with the patient. The presence of home palliative care provided by doctors and nurses is another important factor associated with an increased likelihood of home death^18,19,24,25

When comparing patients dying at home and those in hospital in this study, having FDH as caregiver living with patients is found to be associated with home death. FDHs comprised 5% of HK population and covered 11% of local households.^26,27 As most of the patients who died at home (81.3%) and in hospital (95.2%) already have full-time family caregivers living with them, the specific caregiving role shared by FDH would worth further exploration. The employment of FDH may also indirectly reflect a better social condition and higher economic status which favour home death.^18,28,29 The current study does not find other sociodemographic or illness factors associated with place of death. Lower psychological stress in caregivers has been reported to be associated with home death, but our study does not find significant difference in the stress by RSS between family caregivers of patients dying at home and in hospital (10.5 and 13, Mann-Whitney U = 79.0, n1 = 22, n2 = 17, p = 0.624). A reason may be that families under obvious high psychological stress could have been excluded from the programme by the clinical team upon recruitment assessment.

In this study, more patients who died at home than those in hospital received high intensity of nursing home visit (5 times or more per week). It is consistent with previous review that intensity of home care increased odds of home death¹⁸, reflecting the crucial and heightened role of community nursing support in the last days of life for patients dying at home. Patients who died at home had fewer emergency admissions through AED, consistent with previous report that prior crisis related re-hospitalization would predict hospital death.²⁴ Patients who died in hospital had more clinical admissions to PC ward and appeared to have received longer duration (median 109 versus 54 days, not statistically significant) of PC service. The experience of support from in-hospital PC team or hospice unit may increase the acceptance of admission for symptom relief and has been identified as a determinant of hospital death.¹⁹

Around 40% of patients admitted to hospital when they were imminently dying, despite initial preference on home death and presence of both lay and professional support at home. Evolving physical symptoms, which might not be well anticipated by patients and caregivers, were the major reason of admission. As an example, shortness of breath was known to have its prevalence increased significantly before death and was a common reason for PC patients to visit AED.³⁰ Taking care of a dying family member at home could be physically and psychologically very demanding. We found that some patients’ family might finally choose hospital death, especially when the imminently dying patient became mentally dull and was apparently not aware of his/her location.

Professional support in community for patients and families in the last days of life

Whether a patient with advanced illness can die at home depends heavily on the healthcare services available in the community. This study captured the holistic approach in the dying-at-home programme which includes symptom management to maintain comfort and dignity as well as communication and shared decision with the dying patient and people important to them.³¹ For patients in community, evidence suggested that palliative home care services increased odds of dying at home and were effective in reducing symptom burden.³² The importance of caregiver education and empowerment could not be overemphasised since the sustainability of keeping terminally ill patients at home depends on how able they are to care for their loved ones at home.¹⁸ The fact that caregivers in this programme were trained to monitor fluid and medication infusion and were informed of the dying process may both have contributed to home death.²⁴

Use of strong opioids via CSCI is not feasible in the dying-at-home programme. It is due to the present lack of corporate guidance which could back up the administration of dangerous drug (DD) via syringe driver outside hospital setting. CSCI is a common technique in contemporary palliative care and the use of the compact syringe driver for CSCI allowed symptom control at home for patients who no longer tolerate oral medication.³³ Use of strong opioids via CSCI is an important intervention for patient dying at home with pain and/ or dyspnoea. Guidelines on use of syringe driver with DD in community should be established by HA to overcome this barrier in the public sector.

Legal and logistic issues of home death in Hong Kong

Without the availability of ad hoc doctor visit for certification of death in the dying-at-home programme, all patients would be sent to AED just after death. Unfortunately, majority (87.5%) of these patients had received CPR delivered by ERP, despite the presentation of the valid and applicable advance directives made by the patients or the DNACPR forms written by the doctors. Currently in HK, ERP of Fire Services Department (FSD) are bound by Fire Services Ordinance (FSO) to perform CPR and other related resuscitation to any person who appears to need prompt or immediate medical attention.³⁴ This issue was addressed by Food and Health Bureau in a public consultation in 2019.³⁵ The recently released Consultation Report have proposed measures including the use of a statutory DNACPR form and the corresponding amendment in FSO provisions to accept this form.³⁶ Such legislative changes are urgently needed to resolve the conflict between the wish of the patient and the obligation of ERP.

Before the legal resolution, an alternative way might come from partnership with family physicians, in which a registered doctor who has attended the patient during his/her last illness can certify the death at home and issue the Certificate of the Cause of Death (Form 18).³⁷ With Form 18, the family can proceed to register the death and thence legally move the body³⁷ from home through funeral service

Limitations

This study has a small sample size but provided the very low preference of dying at home in Hong Kong, this study offered initial findings in this area of local research. Patients recruited in the dying-at-home program were highly selected: clear preference for dying at home, presence of full-time caregiver and receiving a high-intensity professional support in community. Therefore, the findings may not be generalisable to other people with advanced life-limiting illness in Hong Kong. However, the recruited patients still cover a wide range of cancer and non-cancer diagnoses, age groups and social background which may still provide meaningful account of the profile of this specific group of patients.

Conclusion

The findings of this study suggest that home as a preferred place of death is achievable in HK. Presence of community palliative care service supporting patient and caregiver is an important determinant. To avoid unnecessary and futile CPR, there is a pressing need in legislative changes to resolve the conflict between the duty of ERP in saving life and the wish of patients to refuse CPR.