With the increased focus on patient-centered care, there is a greater emphasis on understanding patients’ preferences and striving to help them achieve their goals. This approach is crucial throughout the patient’s journey, including its final stages [1,2]. Numerous studies across various settings and cultures have consistently identified the home as the preferred place for passing away [1,3-5]. In Singapore, the Lien Foundation’s “Death Attitude Survey” conducted in 2014 revealed that 77% of Singaporeans preferred to die at home [3], yet only 30% actually did. By 2022, the percentage that passed on at home increased to 39% [6].

Compassionate discharges, also known as “rapid discharges” in the literature, are implemented when a patient is determined to be in the dying phase, with a prognosis ranging from a few hours to 1 week [2,4,7-9]. The primary goal of these discharges is to replace burdensome hospital care with comfort care at home, enabling the patient to spend their final hours in their preferred setting [9,10]. These complex healthcare interventions involve acute care providers, community hospice, and nursing services, as well as the patient’s family [2,7]. They require careful coordination to ensure a smooth transition from hospital to home [2].

In Singapore, the compassionate discharge process can be initiated for any patient in the dying phase, provided there is an expressed desire to die at home and the home environment can offer appropriate care. According to referral statistics from HCA Hospice, the largest home-based hospice organisation in Singapore, the demographics of patients receiving compassionate discharge reflect those receiving home-based palliative care more broadly. Approximately 80% of these patients have a cancer diagnosis, while the remaining 20% have non-cancer conditions.

There are several apparent enablers for successful compassionate discharges in Singapore, such as a national framework designed to standardise discharge practices throughout the country, a dependable transportation infrastructure that facilitates timely patient transfers, and the presence of community hospice providers who offer specialised palliative care in the home setting [8].

Despite these facilitating factors, there are still instances of “unsuccessful” compassionate discharges. These occur when a patient, despite being discharged home on compassionate grounds, returns to an institution for care [7]. As part of a service evaluation, a retrospective audit was conducted on all compassionate discharges from Singapore General Hospital, managed by the Division of Palliative and Supportive Care, from October 2022 to October 2023. The purpose of this audit was to characterise these unsuccessful discharges and to identify potential reasons for them.

Out of 106 compassionate discharges, six were unsuccessful. Two cases, detailed in Table 1, were specifically chosen for further analysis because these patients were ultimately transferred to an inpatient hospice for end-of-life care. The medical records of these two cases, from both the acute hospital and home hospice, were reviewed and analysed to identify common themes and uncover the gaps and challenges that hindered the desired death at home. The other four cases were excluded from analysis because the patient or family opted to return to the Emergency Department for another attempt at active treatment, implying that the goals of care had changed.

Table 1 . Demographics of the Selected Cases..

Case study	1	2
Age range/Sex	80’s/Female	70’s/Male
Life-limiting condition	Ischemic heart disease	Pneumonia
Past medical history	・Ischemic heart disease with an ejection fraction of 35% ・Stage 5 chronic kidney disease from diabetes mellitus, not on renal replacement therapy ・Hypertension ・Hyperlipidaemia ・Chronic normochromic normocytic anaemia ・Osteoporosis	・Ischemic heart disease with moderate aortic regurgitation and ejection fraction of 50% ・End stage renal failure secondary to hypertensive nephropathy and use of non-steroidal anti-inflammatories on dialysis on long term peritoneal dialysis ・Obstructive sleep apnoea
Social history	・Widow with 5 children ・Lived with one son and one daughter	・Married with 2 children ・Lived with wife
Pre-morbid function	・Independent in basic activities of daily living ・Ambulatory at home with walking stick. Wheelchair in the community	・Independent in basic activities of daily living ・Ambulatory without aid ・Worked as a taxi driver
Clinical condition at time of discharge from hospital
Blood pressure (mmHg)	109/58	114/61
Heart rate (beats/min)	80	65
Oxygen (%)	92 (on 2 L via nasal cannula)	100 (on 2 L via nasal cannula)
Mental state	Lethargic	Alert
Nursing needs	・Oxygen concentrator (nasal cannula) ・Hospital bed	・Nasogastric tube ・Oxygen concentrator (nasal cannula)
Primary symptom	Agitation	Dyspnoea
Baseline medications	Subcutaneous haloperidol 0.5 mg twice a day	Transdermal fentanyl patch 6 mcg/hour every 72 hours
Caregiver	Two daughters	Wife
Number of days between discharge and readmission	10	2
Number of days between readmission and demise	7	2
Place of death	Inpatient hospice	Inpatient hospice
Contact with the home hospice team
Number of house visits	6	3
Number of phone calls	4	9

Per local institutional guidelines, Institutional Review Board (IRB) approval was not required for this case presentation since it was conducted as part of a service evaluation and involved fewer than three patients. All patient data were anonymised to ensure confidentiality and privacy [11].

Case 1 was initially admitted to the geriatric service for congestive cardiac failure. Her prolonged hospital stay was further complicated by recurrent episodes of hospital-acquired pneumonia, acute kidney failure, and deep vein thromboses. Due to her clinical deterioration, she was assessed to be in the dying phase. In accordance with her wishes, arrangements were made for a compassionate discharge. To facilitate her care at home, a hospital bed and an oxygen concentrator were procured, and two of her daughters underwent caregiver training provided by the hospital staff in basic nursing care. Once at home, the home hospice team observed that the caregivers were struggling with their duties and feeling overburdened. In response, the team conducted additional home-based caregiver training and applied for government-subsidised nursing aides to assist at home. However, these aides were not immediately available, and the family could not afford a private nurse. Consequently, they reached out to family members abroad for assistance. As these relatives could only stay for 1 week, the family decided to transfer her to an inpatient hospice for end-of-life care once they had to return.

Case 2 was initially admitted to the cardiothoracic surgery department for a sternal wound infection following a cardiac bypass. His medical history included ischemic heart disease and end-stage renal failure, for which he was undergoing long-term peritoneal dialysis. His stay was extended due to candida peritonitis, which necessitated a switch from peritoneal dialysis to haemodialysis. He was later discharged to a community hospital for rehabilitation. One month into his rehabilitation, he was transferred back to an acute hospital due to sepsis. This admission for severe pneumonia required multiple stays in the intensive care unit for multiorgan support. When it was determined that this aggressive treatment approach was futile, a family conference was convened to discuss treatment goals and the extent of care. At this juncture, the patient expressed a desire to discontinue renal replacement therapy and requested a compassionate discharge. His family recognised that they could not manage his high care needs at home, as he was bedbound, required nasogastric tube feeding, and needed subcutaneous injections for symptom relief. Nevertheless, they chose to respect the patient’s wishes and decided to hire private nurses to assist them at home following the compassionate discharge. Two days after returning home, the family decided to transfer the patient to an inpatient hospice, as they were unable to continue with private nursing support and felt uncomfortable managing his care on their own.

The scope of this paper does not extend to identifying all possible factors that may result in an unsuccessful compassionate discharge. However, our findings highlight some patient, caregiver, and healthcare professional factors that may influence the outcome of a compassionate discharge.

The patient factor identified was that of a drastic increase in the care needs. Not unexpectedly, both patients experienced a steep decline in physical and cognitive function after discharge. Prior to admission, they were self-caring. However, being discharged in the terminal phase meant that they were bedbound and required full nursing care. As they previously did not require full-time caregivers, their baseline home setups were neither physically nor socially equipped to meet these heightened needs. Consequently, these families struggled to address these demands emergently [12]. In both cases, sources of help external to the family nucleus were tapped on to allow the patients to be cared for at home, albeit only temporarily.

Beyond addressing the physical needs of patients, caregivers also face a significant psychological burden when caring for the dying at home. For instance, one caregiver for case 1 reported feeling nauseated while cleaning the patient. This suggests that caregivers need to come out of their comfort zones to assume these new challenging responsibilities. In case 2, the private nurses hired showed distress. During the two nights at home, the private nurses called the hospice helpline multiple times for clarification. Patients in the dying phase are often drowsy and caregivers must use non-verbal cues to assess them. Furthermore, they are susceptible to sudden changes in clinical condition, and new distressing symptoms might arise post-discharge. Caregivers, even professional ones, without palliative care training struggle with this. There are fears that giving injectable medications can cause unwanted side effects to the patient and/or hasten death, along with a simultaneous worry about omitting medications that can relieve suffering [13]. These judgment calls, which are difficult in specialised inpatient units, are even more difficult at home where caregivers do not always have trained clinicians with them to make assessments together. The use of video calls with the home hospice team to troubleshoot distressing symptoms has been instrumental in overcoming this issue.

Finally, in both scenarios, family members were willing to care for the dying person only under specific conditions. The discrepancy between the patients’ prognoses and the initial plans prevented the patients from dying at home. Previous studies have indicated that physicians generally prognosticate poorly, although accuracy can improve as death nears [14]. To better advise families, clinicians might benefit from examining the typical or historical time intervals between compassionate discharge and death in their settings. This information could help determine the minimum number of days for which comprehensive care plans should account post-discharge. Prognostication inherently involves uncertainty, which should be clearly communicated to families from the outset. During this emotionally charged period, the clinician’s role is to support families in managing this uncertainty. By leveraging their knowledge and expertise, clinicians should guide and anticipate alongside the family, being sure not to overwhelm them. Both parties must remain flexible and responsive, adjusting the care plan as circumstances evolve.

In conclusion, our findings highlight the complexities involved in managing end-of-life care at home and the factors that may necessitate admission to an inpatient hospice. However, there is still a need for more comprehensive research, particularly on the group of patients who wish to return home to die but encounter barriers that prevent even an attempted compassionate discharge. A deeper exploration of these factors could enhance support systems and better prepare families, enabling more patients to fulfil their wish to die at home surrounded by loved ones.

No potential conflict of interest relevant to this article was reported.

Conception or design of the work: AT. Data collection: AT. Data analysis and interpretation: all authors. Drafting the article: all authors. Final approval of the version to be published:all authors.