Actual and Preferred Location of Death for Children Known to a Palliative Care Service

Abstract

This study aimed to establish the frequency and congruence between preferred and actual locations of expected pediatric deaths and to identify associated factors. A retrospective review of death review forms and medical records was conducted in two pediatric palliative care services in Sydney, Australia, involving 486 patients who died between January 2017 and May 2024). Hospital was the most common LOD (49.8%), then home (28.4%) and hospice (21.0%). At the last documented LOD conversation, 36.2% of families preferred home, 28.6% hospice and 26.4% hospital. Preferred LOD outcomes were achieved for 87.7% of families, although 21% who preferred a home death did not achieve this. Preferred and actual LODs of home were significantly associated with having a community palliative care referral and palliative care home visit. The findings highlight the importance of enabling families to plan for their child’s LOD, and ensuring preferences are documented and communicated across health services.

Keywords

pediatric location death palliative end-of-life

Pediatric palliative care (PPC) teams provide medical, emotional, and spiritual care to children with life-limiting conditions (LLCs) and their family (Australian Institute of Health and Welfare, 2024; Fields et al., 2023). PPC teams may provide this support from diagnosis of a LLC, through to the end of life and into the bereavement period. Given the nature and length of their involvement, PPC teams are uniquely positioned to guide families in meaningful discussions about their preferred location of death (LOD) and to implement the necessary supports to facilitate their preference.

The available options for the LOD may vary based on where families live and the services accessible to them. For a home death, the home environment is reviewed, specialized equipment may be installed, community healthcare services involved, and parents supported to manage the child’s medical needs (Hansson et al., 2023; Winger et al., 2020). A hospital death can occur at a local or tertiary pediatric hospital, and may occur in a hospital ward, intensive care unit, or sometimes the emergency department (Trowbridge et al., 2018). Pediatric hospices are purpose built to support the needs of children with LLCs including those at the end of life, providing a home-like environment, but with the medical equipment and appropriate multidisciplinary support (Bear Cottage, n.d.; Paediatric Palliative Care Australia & New Zealand, 2023). However, pediatric hospices are limited in number in developed countries and often non-existent in less developed countries (Arias-Casais et al., 2020).

In one of the largest population-based reviews and retrospective analyses of pediatric deaths conducted in 2008, hospital emerged as the most common LOD in all countries (Håkanson et al., 2017). More recent studies have also mostly found hospital to be the most common location of expected pediatric deaths in well-developed countries (Australian Institute of Health and Welfare, 2024; Gibson-Smith et al., 2021; Peláez-Cantero et al., 2023), although some studies have found very similar proportions across locations (Siden et al., 2008; Sneha et al., 2023; Wiener et al., 2020). Notably, parents have reported a better PPC experience when they were able to plan the LOD, regardless of the location chosen (Bluebond-Langner et al., 2013; Rainsford et al., 2018). Therefore, empowering families to plan for the LOD and providing them the flexibility to change their choice is considered an important aspect of high-quality PPC (Dussel et al., 2009; Paediatric Palliative Care Australia & New Zealand, 2023).

Families will differ in their preferences of LOD, depending on the accessibility of services and the needs and priorities of the family (Papadatou et al., 2021). Families may also change their preferences as new factors emerge, or as they prioritize different considerations at different stages of the child’s disease progression (Bluebond-Langner et al., 2013; Vickers et al., 2007). To our knowledge, only one study (Vickers et al., 2007) has recorded LOD family preferences at multiple time points. In that study, 68% of families initially preferred a home death, but by the last month of their child’s life, this preference increased to 80% (Vickers et al., 2007). There is a need for a better understanding of how parental preferences for LOD change over time.

Although healthcare professionals endeavor to support the family’s goals of care for LOD, there are various reasons why achieving these goals is not always attainable. It may be difficult to instate certain supports if there is a lack of timely planning, thus limiting the options for LOD (Montel et al., 2009). Across varied international contexts - including Australia (n = 301) (Australian Institute of Health and Welfare, 2024), Spain (n = 164) (Peláez-Cantero et al., 2023), and the US (two studies, n = 140 (Dussel et al., 2009) and n = 150 (Lee et al., 2023))- between 87% and 99% of children receiving palliative care have been found to die at their, or their family’s, preferred LOD (Australian Institute of Health and Welfare, 2024; Dussel et al., 2009; Lee et al., 2023; Peláez-Cantero et al., 2023). While achieving a family’s preferred LOD is not always realistic or achievable due to a variety of factors related to service availability (Dussel et al., 2009), a better understanding of the circumstances in which a family may not be able to achieve their desired LOD may help inform parental decisions and enable clinicians to set more realistic expectations for families and address the barriers to achieving their preferred location.

The first aim of this study was to establish the proportion of expected pediatric deaths that occurred at home, hospital or pediatric hospice facility, in a sample of patients/families known to two Sydney-based pediatric palliative care services. These palliative care services, servicing families across New South Wales (Australia), routinely facilitated end-of-life care arrangements at all three location types. During the full period of data collection, one pediatric hospice facility was available to families. The second aim was to identify LOD preferences expressed by families and the reasons for these preferences, giving consideration to the frequency with which documented preferences changed over the last two weeks of the patient’s life. The third aim was to assess the level of congruence between the preferred and actual LOD, and determine what factors were associated with achieving a preferred LOD. Given the potential challenges documented in the literature in achieving home deaths (Montel et al., 2009; Sagha Zadeh et al., 2018; Zelcer et al., 2010), a secondary aim of the study was to analyze factors associated with actual and preferred home deaths relative to deaths in other locations.

Methods

Study Design

This multi-centred, retrospective study utilized information from death review (Clinical Excellence Commission, n.d.) forms and electronic medical records, including both quantitative and qualitative data.

Ethics approval was granted by the Sydney Children’s Hospital Network (SCHN) Human Research Ethics Committee (2023/ETH02510).

Cohort

The cohort was drawn from pediatric patients cared for by one of the two pediatric palliative care services in the SCHN (Sydney Children’s Hospital Randwick or The Children’s Hospital at Westmead). Within the catchment areas, families had access to two tertiary pediatric hospitals, a specialized pediatric hospice facility and community teams to support home care. The inclusion criteria were: (i) patients died between January 2017 and May 2024, and (ii) died before 19 years of age. There were no exclusion criteria.

Data Extraction and Coding

Data was extracted from death review forms and electronic medical records. Death review forms are completed by a member of the palliative care team based on feedback received during a death review meeting, typically held within one month of the patient’s death, with HCPs present from the various teams involved in the care of the patient. Data from death reviews was extracted for: demographic information (e.g., primary language spoken at home, cultural background, religion/spirituality), referral information (e.g., reasons for palliative care referral, if referral was timely, diagnosis) and LOD information (all documentation regarding conversations between HCPs and family members regarding possible locations of an impending death, as well as the location of the actual death). Although all LOD conversations were coded, these are most commonly documented by palliative care teams, as part of their routine discussions and planning for end-of-life.

Once death review forms were accessed, data for each patient was also extracted from electronic medical records. Key aspects of data retrieved from the electronic medical records included demographical information, whether a community palliative care team referral was carried out, whether a PPC home visit was made, and the Index of Relative Socio-economic Disadvantage (IRSD) percentile (Australian Bureau of Statistics, 2023). The IRSD is an Australian index used to provide information about socioeconomic disadvantage relative to other residential areas based on postal code, with a lower quintile denoting greater disadvantage (Australian Bureau of Statistics, 2023). Other key information collected from the medical records included the timing and preferences for each documented LOD conversations (namely any conversation with the family about considerations or decisions regarding the location at which death may occur), as well as the congruence between preferred and actual LOD. All medical records within three months of the patient’s death were reviewed and coded by one investigator (VS).

Figure 1 shows the information retrieved from the cohort that satisfied the inclusion criteria.

Figure 1.

Information retrieved from cohort

Statistical Analyses

Data elicited were primarily used to carry out descriptive analyses such as reporting the frequency of PPC patient deaths at different locations, the frequencies of different LODs preferred at different points in time, the timing of LOD discussions, and the frequency that congruence is achieved between preferred and actual LODs.

Chi-square tests were used to determine whether the actual LOD (home vs hospital or hospice) was associated with a range of categorical variables; whether the preferred LOD (home vs hospital or hospice) was associated with a range of categorical variables; and whether patients who had congruence between their preferred and actual LOD differed from those who did not, in terms of patient, family, or health service categorical factors extracted.

Continuous variables were screened for normality (skewness <1 and kurtosis <3) and if non-parametric, Mann-Whitney U tests were used to test for any significant differences between home and non-home deaths and preferences.

The total sample size (n = 486) afforded sufficient statistical power (at p < .05) to detect medium effect sizes, where they existed, when comparing home vs non-home LOD and preferences (Cohen, 1992) for patient, family and health service factors. Bonferroni corrections were applied due to the multiple tests conducted (Armstrong, 2014).

Thematic analysis was carried out for the documented reasons for a LOD preference. One researcher (VS) coded all data using codebook thematic analysis (Braun & Clarke, 2023; Roberts et al., 2019). This involved the development of initial codes based on a review of the literature and data familiarization, and editing and movement of themes/subthemes upon further reads and trial coding until agreement was reached with a second researcher (TJ) that the codebook could be assumed to represent the data.

Results

Patient and Family Demographics

Patient and family demographic variables for the 486 patients who died over the 7-year study period are presented in Table 1.

Table 1.

Demographic Characteristics of PPC Patients Who Died Between 2017 and 2024 (n = 486)^a

Child demographics	n	%
Sex (n = 485)
Female	225	46.4
Male	260	53.6
Age (n = 485)
Infant (0 to <1 year)	114	23.5
Child (1 to <13 years)	256	52.7
Young person (13 to <19 years)	115	23.7
Primary language at home (n = 483)
English only	381	78.9
Non-English	81	16.8
Both English and non-English	21	4.3
Primary diagnosis (n = 485)
Oncology	213	43.9
Neurology	81	16.7
Cardiology	37	7.6
Metabolic	29	6.0
Respiratory	9	1.9
Genetic	89	18.4
Gastroenterology	18	3.7
Other	9	1.9
IRSD percentile (n = 484)
0–20% (lowest quintile)	127	26.2
21–40%	49	10.1
41–60%	99	20.5
61–80%	90	18.6
81–100%	119	24.6

^aWhere there was insufficient documented information in order to code for specific variables, a smaller sample size is noted.

Actual LOD

Frequency of Deaths at Different Locations

During the study period, 49.8% of pediatric palliative care patients (n = 242) died at hospital, 28.4% (n = 138) died at home, 21.0% at a hospice (n = 102) and 0.8% of patients (n = 4) died at another location. See Supplemental materials for information regarding temporal trends in deaths occurring at different locations between 2017 and 2024, revealing a relative spike in home deaths in 2021 and a spike in hospice deaths in 2022 (these being years in which COVID-19 had the greatest impact in New South Wales, Australia).

Factors Associated With Actual LOD

No patient or disease factor tested, such as primary diagnosis and age, nor family factors such as primary language spoken at home and IRSD percentile, were found to be significantly associated with actual LOD (Table 2). Having a community palliative care referral and a PPC home visit were significantly associated with patients having a home death (χ2 (3, n = 448) = 63.98, p = .001). The number of months between PPC referral and death was significantly greater for patients who died at home (Mdn = 3.98 months, IQR = 13.50) than for those who died at a non-home location (Mdn = 2.33 months, IQR = 7.64, U = 12,567.500, z = −2.996, p = .003). Patients who died at hospital were more likely to have no recorded LOD plan at all, compared to those who died at home, hospice, or another location (p < .001).

Table 2.

Associations (Chi-Square Tests) Between Actual LOD (Home Versus Non-Home) and Patient, Family, and Health Service Factors

	Actual location of death		p value
	Home n (%)	Non-home location n (%)	p value
Patient & disease factors
Primary diagnosis (n = 485)
Oncological	59 (27.7)	154 (72.3)	.745
Non-oncological	79 (29.4)	193 (70.6)
Age classification at death (n = 485)
Infant (0 to <1 year)	33 (28.1)	81 (71.9)	.955
Child (1 to <13 years)	72 (28.1)	184 (71.9)
Young person (13 to <19 years)	34 (29.6)	81 (70.4
Difficult to manage symptom (n = 449)
Yes	26 (32.1)	55 (67.9)	.345
No	99 (26.9)	269 (73.1)
Identified to be at terminal phase (n = 467)
Yes	112 (28.1)	287 (71.9)	.983
No	19 (27.9)	49 (72.1)
Family factors
Primary language at home (n = 424)
English	105 (30.3)	242 (69.7)	.096
Non-English	16 (20.8)	61 (79.2)
IRSD percentile (n = 484)
Lowest quintile	34 (26.8)	93 (73.2)	.613
Above the lowest quintile	104 (29.1)	253 (65.8)
Number of parents/caregivers at child’s primary household (n = 463)
One parent/caregiver only	27 (31.4)	59 (68.6)	.578
Two parents/caregivers	107 (28.4)	270 (71.6)
Siblings in the household (n = 471)
Yes	106 (28.9)	261 (71.1)	.842
No	29 (27.9)	75 (72.1)
Health service factors
Timeliness of PPC referral (n = 415)
Timely	99 (27.2)	265 (72.8)	.740
Untimely or unsure	15 (29.4)	36 (70.6)
Community palliative care referral (n = 381)
Referral made	109 (52.9)	97 (47.1)	<.001*
Offered and declined or not offered	24 (13.7)	151 (86.3)
Home visit (n = 448)
Carried out	111 (51.4)	105 (48.6)	<.001*
Not carried out	18 (7.8)	214 (92.2)

Note. As there were 11 statistical tests, the Bonferroni-adjusted p value needed for significance is .005.

*p < .01.

Post hoc analyses revealed that patients who had previously stayed at, or visited, a pediatric hospice were more likely to die at a hospice relative to at home or hospital (χ2 (3, n = 467) = 13.55, p < .001). Patients who died at a hospice did not differ from those who died at other locations in terms of existence of siblings versus no siblings (χ2 (3, n = 471) = .020, p = .888) or whether caregivers were single/separated versus a couple (χ2 (3, n = 463) = .005, p = .945).

Preferred LOD

Frequencies of Preferences and Timing of Discussions

Of the 476 patients whose medical records could be accessed, 292 (61.3%) had a recorded family preference for LOD, as recorded at the final LOD conversation documented in the medical records. Home was the most preferred LOD at all time points (see Table 3). The median number of days between the last documented conversation regarding LOD and the patient’s death was 6 days (IQR = 2–16).

Table 3.

Preferred LOD at Different Conversations About LOD Prior to Death

LOD preference	Last LOD conversation (N = 318); n (%)	Second-last LOD conversation (N = 165); n (%)	Third-last LOD conversation (N = 66); n (%)
Hospital (total)	84 (26.4)	17 (10.3)	6 (9.1)
Children’s (general)	57 (17.9	10 (6.1)	4 (6.1)
Children’s (ICU)	22 (6.9)	1 (0.6)	0 (0)
Local	5 (1.6)	6 (3.6)	2 (3.0)
Home	115 (36.2)	65 (39.4)	25 (37.9)
Children’s hospice	91 (28.6)	52 (31.5)	22 (33.3)
Other	2 (0.6)	1 (0.6)	0 (0)
No preference recorded	26 (8.2)	30 (18.2)	13 (19.7)

Reasons for LOD Preferences

As seen in Table 4, psychosocial factors were the most common reason for LOD preferences at all time points. However, documented conversations about clinical care factors influencing the family’s LOD preference, increased in frequency closer to the time of death.

Table 4.

Primary Reasons for LOD Preferences as Documented in Medical Records

Primary reason for LOD preference	Last LOD conversation (N = 131); n (%)	Second-last LOD conversation (N = 59); n (%)	Third-last LOD conversation (N = 27); n (%)
Psychosocial factors of family
Total	70 (53.4)	31 (52.5)	15 (55.6)
Family demographics	43 (32.8)	18 (30.5)	10 (37.0)
Caregiver burden	15 (11.5)	4 (6.8)	2 (7.4)
Wishes of patient	12 (9.2)	9 (15.3)	3 (11.1)
Clinical care
Total	39 (29.8)	10 (16.9)	4 (14.8)
Symptom management	18 (13.7)	3 (5.1)	1 (3.7)
Availability of appropriate clinical services	11 (8.4)	5 (8.5)	3 (11.1)
Family preferences or familiarity with personnel	10 (7.6)	2 (3.4)	0 (0)
Not wanting to change location	7 (5.3)	4 (6.8)	0 (0)
Multiple reasons of equal priority	9 (6.9)	6 (10.2)	3 (11.1)
Other	6 (4.6)	8 (13.6)	5 (18.5)

Over the last two weeks of life, 32.1% (n = 153) of families had at least one change of LOD preference (as documented in their medical records), and 3.8% had more than one change documented.

Factors Associated With Preferred LOD

Having a community palliative care referral or any PPC home visit carried out were both associated with the last recorded LOD preference being for home (Table 5). The patient and family factors tested were not significantly associated with the preferred LOD. Time between PPC referral and death was greater for patients who preferred home (Mdn = 3.58 months, IQR = 7.10) than for patients with a non-home LOD preference (Mdn = 1.68 months, IQR = 5.45, U = 5,656.000, z = −3.306 p = 0.001). Families who preferred a hospice death were significantly more likely to have either stayed at, or visited, a pediatric hospice facility as opposed to having never visited (χ2 (1, n = 288) = 6.34, p = .012).

Table 5.

Associations (Chi Square Tests) Between Last Recorded LOD Preference (Home Versus Non-Home) and Child, Family, and Health Service Factors

	Preferred LOD		p value
	Home; n (%)	Non-home location; n (%)	p value
Child factors
Primary diagnosis (n = 291)
Oncological	55 (36.9)	94 (63.1)	.352
Non-oncological	60 (42.3)	82 (57.7)
Age classification at death (n = 291)
Infant (0 to <1 year)	28 (42.4)	38 (57.6)	.224
Child (1 to <13 years)	56 (35.2)	103 (64.8)
Young person (13 to <19 years)	31 (47.0)	35 (53.0)
Difficult to manage symptom (n = 272)
Yes	21 (42.9)	28 (57.1)	.426
No	82 (36.8)	141 (63.2)
Family factors
Primary language at home (n = 258)
English	88 (40.2)	131 (59.8)	.614
Non-English	14 (35.9)	25 (64.1)
IRSD percentile (n = 292)
Lowest quintile	29 (42.0)	40 (58.0)	.607
Above the lowest quintile	86 (38.6)	137 (61.4)
Number of parents/caregivers at child’s primary household (n = 284)
One parent/caregiver only	21 (41.2)	30 (58.8)	.823
Two parents/caregivers	92 (39.5)	141 (60.5)
Siblings in the household (n = 288)
Yes	86 (39.4)	132 (60.6)	.896
No	27 (38.6)	43 (61.4)
Health service factors
Timeliness of PPC referral (n = 252)
Timely	84 (38.5)	134 (61.5)	.974
Untimely or unsure	13 (38.2)	21 (61.8)
Community palliative care team referral (n = 217)
Carried out	91 (74.0)	32 (26.0)	<.001*
Offered and declined or not offered	19 (20.2)	75 (79.8)
Home visit (n = 277)
Carried out	91 (66.9)	45 (33.1)	<.001*
Offered and declined or not offered	19 (13.5)	122 (86.5)

* As there were 10 statistical tests, the Bonferroni-adjusted p value needed for significance is .005.

Congruence Between Preferred and Actual LOD

Frequency of Congruence

Of the 292 cases with a documented preference regarding LOD at the last recorded discussion about LOD prior to death, 87.7% had congruence between the preferred and actual LOD. In response to the question on the death review form “was the end-of-life location goal met”, 83.5% (n = 345) of responses indicated that the location goal was met, 4.8% that the goal was not met, and 11.6% that it was unknown.

Factors Associated With Preferred and Actual LOD Congruence

The majority of families (97.6%) who preferred a hospital death for their child (as documented in their last LOD discussion prior to death) achieved their goal. In comparison, 92.3% of families who preferred a hospice death and only 78.9% of families who preferred a home death achieved this outcome (Table 6).

Table 6.

Congruence Between Last Preferred LOD and Actual LOD

Last preferred LOD	Actual LOD				p value
Last preferred LOD	Hospital n (%)	Home n (%)	Hospice n (%)	Total N	p value
Hospital	82 (97.6)	2 (2.4)	0 (0)	84	<.001*
Home	19 (16.7)	90 (78.9)	5 (4.4)	114
Hospice	6 (6.6)	1 (1.1)	84 (92.3)	91

* Significant if p<.05.

There were no significant differences in sociodemographic disadvantage (IRSD), age at death, and days between PPC referral and death, between patients who died at the family’s preferred location and those who did not (Table 7).

Table 7.

Congruence Between Last Preferred LOD and Actual LOD by Patient, Family, and Health Service Factors

	Congruence		p value
	Yes; n (%)	No; n (%)	p value
Patient & disease factors
Primary diagnosis (n = 296)
Oncological	139 (91.4)	13 (8.6)	.145
Non-oncological	124 (86.1)	20 (13.9)
Age classification at death (n = 296)
Infant (0 to <1 year)	58 (85.3)	10 (14.7)	.517
Child (1 to <13 years)	142 (89.3)	17 (10.7)
Young person (13 to <19 years)	63 (91.3)	6 (8.7)
Difficult to manage symptom (n = 275)
Yes	39 (79.6)	10 (20.4)	.035
No	204 (90.3)	22 (9.7)
Identified terminal phase at death (n = 294)
Yes	243 (90.0)	27 (10.0)	.102
No	19 (79.2)	5 (20.8)
Family factors
Primary language at home (n = 261)
English only or with other languages	196 (88.3)	26 (11.7)	.844
Non-English	34 (87.2)	5 (12.8)
IRSD percentile (n = 297)
Lowest quintile	61 (86.1)	9 (13.9)	.595
Above the lowest quintile	203 (89.4)	24 (10.6)
Number of parents/caregivers at child’s primary household (n = 289)
One parent/caregiver only	43 (82.7)	9 (17.3)	.114
Two parents/caregivers	214 (90.3)	23 (9.7)
Health service factors
Timeliness of PPC referral (n = 254)
Timely	192 (87.3)	28 (12.7)	.095
Untimely or unsure	33 (97.1)	1 (2.9)
Community palliative care referral (n = 221)
Referral made	107 (84.9)	19 (15.1)	.321
Offered and declined or not offered	85 (89.5)	10 (10.5)
Home visit (n = 282)
Carried out	119 (86.2)	19 (13.8)	.210
Not carried out	131 (91.0)	13 (9.0)

Note. As there are 10 statistical tests, the Bonferroni-adjusted p value needed for significance is .005.

Discussion

This study provided data regarding the actual and preferred LOD of children known to two tertiary pediatric palliative care services in Sydney, Australia. In line with the first aim of the study, 50% of the pediatric palliative care patients were found to die at hospital, 28% at home, and 21% at a pediatric hospice facility. The proportion of patients who died at home was lower than most figures reported across other pediatric studies and geographic contexts: England: 16% (Gibson-Smith et al., 2021), Canada/Australia/New Zealand: 35% (Siden et al., 2008), Australia: 38% (Australian Institute of Health and Welfare, 2024), Spain: 41% (Peláez-Cantero et al., 2023), India: 48% (Sneha et al., 2023), various countries: 43% (Wiener et al., 2020) to 58% (Noyes et al., 2022), Brazil 59% (Kurashima et al., 2005). The relatively low proportion of home deaths in the current study may in part be due to the relative ease of availability of a pediatric hospice facility to the current cohort.

For patients in the current dataset who died at home, referral to PPC generally occurred earlier than for those who died in non-home locations. Earlier PPC referrals likely allow for more opportunity for the necessary preparations for a home death to occur: a referral to the community-based palliative care team (Malcolm et al., 2020), the set-up of specialized equipment in a suitable home space (Sagha Zadeh et al., 2018), and parental education (Winger et al., 2020). Not surprisingly patients who received a referral to a community palliative care team or had any PPC home visit, were more likely to die at home as opposed to hospital or hospice, as community teams can provide the symptom-management and support that families need in order to feel comfortable at home (Hannan & Gibson, 2005; Noyes et al., 2022; Papadatou et al., 2021). Patients who died at home versus out-of-home were not found to differ on a range of patient, family or sociodemographic variables tested. This finding may suggest that equitable access of end-of-life options were available to families. For example, families who were non-English speaking or with greatest socioeconomic disadvantage did not differ in terms of their child’s actual LOD (home versus non-home location). However, caution should be exercised in drawing conclusions from null results (Harms & Lakens, 2018).

Patients who had previously visited or stayed at the hospice were significantly more likely to die at that facility, perhaps due to familiarity and previous positive experiences of receiving respite care at the facility (Dunbar et al., 2020). Encouraging families to visit a pediatric hospice before the end-of-life may help them realize the unique benefits (Gola et al., 2016; Hain, 2019) of the location, and to consider hospice as a potential LOD.

When addressing the second aim of the study, data was collected regarding family preferences for the location of an expected child death, and how these preferences changed over time. The percentage of families who preferred a hospital LOD more than doubled from the second-last (10.3%) to last (26.4%) documented LOD conversation. In contrast, the percentage of families that preferred a home LOD decreased across the second last (39.4%) to last (36.2%) documented conversation. These results highlight how LOD preferences may change over the illness trajectory, as families prioritize differently a range of psychosocial factors and clinical care. As the patient is nearing death, concerns about managing clinical symptoms, family psychosocial dynamics, or perceived limitations in support for a home death can lead families to feel less confident in their ability to successfully manage a home death (Hannan & Gibson, 2005; Noyes et al., 2022; Papadatou et al., 2021; Sneha et al., 2023).

It should be noted that less than two-thirds of pediatric palliative care patients had any medical record documentation regarding the family’s preferences regarding LOD. This implies that for one-third of pediatric palliative care patients, conversations regarding LOD preferences either did not occur with the family prior to the patient’s death, or that these conversations were not appropriately documented. There are several reasons why LOD conversations do not occur in a timely way. Some families are reluctant to engage in difficult discussions about their child’s impending death (Ulrich et al., 2018). Healthcare professionals may hold off on engaging in such discussions until they feel there is a more appropriate time to have this discussion, or until they have developed more rapport with the patient (Cox et al., 2011). Families and healthcare professionals may not recognize that death is imminent (Chu et al., 2019), with prognostication often difficult in the context of early phase clinical trials and medical interventions (Kennedy et al., 2021). Consequently, conversations about end-of-life planning and death may be postponed too long.

Regarding the third aim of the study, it was found that about 89% of patients died at the preferred LOD. This figure is consistent with findings from other studies, across a range of locations, with congruence figures ranging between 70% and 99% (Dussel et al., 2009; Kassam et al., 2014; Lee et al., 2023; Peláez-Cantero et al., 2023; Stilwell et al., 2022). While nearly all families in the current study preferring a hospital or hospice death achieved their desired outcome (98% and 92% respectively), only about 79% of families who preferred a home death achieved this location, with 17% of these patients dying at hospital. Other research has similarly reported that a significant proportion of families desiring a home death for their child, may not achieve their goals (Kassam et al., 2014; Stilwell et al., 2022). Families who prefer a home death may have their child die at hospital if the child rapidly deteriorates at hospital and is not stable enough to be transported (Stilwell et al., 2022) or if their medical needs cannot be adequately managed at home (Zelcer et al., 2010). Congruence between preferred and actual LOD was more likely for patients whose medical records documented that they were at the terminal stage, as this recognition may have provided time for transport to their preferred LOD.

Limitations

It is likely that not all conversations regarding LOD were documented in the patient’s medical records. Staff may have been more likely to record a preference for dying at home or in a hospice, as these require more preparation and coordination than a hospital death. This is consistent with the finding in the current study that patients without a recorded preference for LOD were most likely to die in hospital.

Although the current study assessed congruence between preferred and actual LOD, it was not possible to determine from the available documentation whether families were satisfied with the actual LOD. Moreover, some documented LOD preferences may have been a compromise, given the lack of feasibility of other options. Consequently, the level of congruence between preferred and actual LODs may be somewhat inflated.

When examining the frequencies of actual and preferred LOD death in the period 2017 to 2024, the impact of COVID-19 on this data is uncertain. During the period when the NSW health sector was most impacted by COVID-19 (2020 to 2022) (Australian Bureau of Health Information, 2023), strict restrictions were placed on the number of visitors allowed in hospitals, and to a lesser extent in hospices. There were also restrictions on the access of community palliative care teams as well as informal supports for patients and families in the home environment (Rattner, 2021), with greater reliance on telehealth (Weaver et al., 2021).

A strength of this study was its inclusion of children with a wide range of diagnoses requiring palliative care, given that previous studies examining LOD have drawn more extensively on pediatric oncological cohorts (Dussel et al., 2009; Kassam et al., 2014; Stilwell et al., 2022; Vickers et al., 2007).

Clinical and Research Directions

Given that only 62% of families known to pediatric palliative care teams had a documented LOD preference (at the time of the last LOD recorded discussion prior to death), there is a need to ensure that all families discuss and plan for the LOD, and that preferences are appropriately documented. Previous studies have shown that healthcare workers may experience difficulties discussing, recording and tracking goals of care (Cox et al., 2011); therefore more work is required to establish how this information may be more routinely recorded and communicated between health services. As LOD preferences are influenced by changing clinical, psychosocial and logistical factors, there is value in having LOD conversations with the family at multiple time points. Earlier PPC referral may provide more opportunities for these discussions, and if the preference is for a home death, more time for home supports to be instated and for adequate parental education (Malcolm et al., 2020; Sagha Zadeh et al., 2018).

As 21% of families were unable to achieve their preference for a home death, clinicians should consider if those preferring a home death have adequate supports in place. If such supports cannot be instated, it may be necessary to support families in identifying and achieving a more realistic LOD goal. Parents and caregivers who aim for a home death, benefit from clear guidance in how they can respond to clinical situations which may arise. Community palliative care team involvement is key in managing challenges to a home death, such as difficult to manage symptoms and rapid deterioration (Montel et al., 2009; Odejide et al., 2022; Stilwell et al., 2022). As these teams often primarily service adult palliative care patients (Contro et al., 2002), they may benefit from specialist pediatric outreach services (Mherekumombe, 2018; Mherekumombe et al., 2016).

Prospective, longitudinal studies are needed to more accurately identify changes in the LOD preferences of families. Furthermore, it would be valuable to assess familial satisfaction with actual LOD, irrespective of whether it was their preferred location. Parental satisfaction with actual LOD is likely to have a lasting impact, with less decisional regret about the LOD (Dussel et al., 2009) and better bereavement outcomes (Jaaniste et al., 2017).

In light of the significant proportion of families unable to achieve their preference for a home death, there is a need to examine local barriers further. Such barriers may include a desire for more healthcare support in the home (Barlund et al., 2021; Wahid et al., 2018), level of trust in community healthcare providers (Brouwer et al., 2020), and access to medical supports at home to decrease the need to travel for inpatient treatment (Barlund et al., 2021). Strategies used to overcome these barriers, such as better healthcare service coordination or upskilling staff (Wahid et al., 2018), may be evaluated by comparing future preferred and actual LOD congruence figures with the benchmark figures reported in the current study.

Further research into how families prefer to have difficult discussions, such as those around LOD, could inform the creation of educational or clinical tools to help healthcare workers engage in sensitive and timely conversations about end-of-life and LOD planning.

Conclusion

This study found that the most common LOD for expected pediatric deaths was hospital, followed by home and then hospice. As end-of-life approached, the documented preference for death at home decreased, while the preference for a hospital death increased. Families who chose hospital as the desired place of death were more likely to achieve congruence between their preferred and actual location of death. However, nearly one-fifth of families who preferred a home death did not achieve that preferred location. The current study underscores the need to improve how LOD preferences are recorded and to explore and address local barriers that prevent achieving a preferred home death.

Supplemental Material

Supplemental Material - Actual and Preferred Location of Death for Children Known to a Palliative Care Service

Supplemental Actual and Preferred Location of Death for Children Known to a Palliative Care Service by Victoria Sun, Sandra Coombs, Nicole Armitage, Meaghan Dowling, and Tiina Jaaniste in OMEGA - Journal of Death and Dying

Footnotes

ORCID iDs

Sandra Coombs

Nicole Armitage

Tiina Jaaniste

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Victoria Sun is a medical student at the University of New South Wales, Australia, who spent her honours year carrying out research in the field of children’s palliative care at the Sydney Children’s Hospital, Randwick, Australia, under the supervision of A. Prof Jaaniste.

Sandra Coombs has a Masters degree in Clinical Nursing and Post-Graduate degree in Pediatric Palliative Care. She has many years of experience as a pediatric palliative care nurse consultant and currently works in a project-based quality role.

Nicole Armitage is a pediatric palliative care specialist. She has a particular interest in hospice and community based care for young people with life-limiting illness.

Meaghan Dowling is a senior occupational therapist with many years’ experience working in the field of pediatric palliative care.

Tiina Jaaniste, PhD is Conjoint Associate Professor in the School of Clinical Medicine, University of New South Wales, Australia and Head of Research in the Departments of Palliative Care and Pain, Sydney Children’s Hospital, Randwick, Australia. She has a background in clinical psychology and has published widely in the fields of children’s pain and palliative care.

References

Arias-Casais

Garralda

Pons

J. J.

Marston

Chambers

Downing

Ling

Rhee

J. Y.

de Lima

Centeno

(2020). Mapping pediatric palliative care development in the WHO-European region: Children living in low-to-middle-income countries are less likely to access it. Journal of Pain and Symptom Management, 60(4), 746–753. https://doi.org/10.1016/j.jpainsymman.2020.04.028

Armstrong

R. A.

(2014). When to use the Bonferroni correction. Ophthalmic and Physiological Optics: The Journal of the British College of Ophthalmic Opticians (Optometrists), 34(5), 502–508. https://doi.org/10.1111/opo.12131

Australian Bureau of Health Information . (2023). New South Wales and the COVID-19 pandemic from 2020 to 2022. https://www.bhi.nsw.gov.au/BHI_reports/healthcare_in_focus/new_south_wales_and_the_covid-19_pandemic_from_2020_to_2022

Australian Bureau of Statistics . (2023). The index of relative socio-economic disadvantage (IRSD). https://www.abs.gov.au/statistics/detailed-methodology-information/concepts-sources-methods/socio-economic-indexes-areas-seifa-technical-paper/2021

Australian Institute of Health and Welfare . (2024). Specialist paediatric palliative care delivered to children who died in 2021. Australian Institute of Health and Welfare. https://www.aihw.gov.au/reports/palliative-care-services/paediatric-palliative-care-for-children-who-died/summary

Barlund

A. S.

André

Sand

Brenne

A. T.

(2021). A qualitative study of bereaved family caregivers: Feeling of security, facilitators and barriers for rural home care and death for persons with advanced cancer. BMC Palliative Care, 20(1), 7. https://doi.org/10.1186/s12904-020-00705-y

Bear Cottage . (nd). Family resources. Sydney Children’s Hospital Network. https://bearcottage.schn.health.nsw.gov.au/family-resources.php

Bluebond-Langner

Beecham

Candy

Langner

Jones

(2013). Preferred place of death for children and young people with life-limiting and life-threatening conditions: A systematic review of the literature and recommendations for future inquiry and policy. Palliative Medicine, 27(8), 705–713. https://doi.org/10.1177/0269216313483186

Braun

Clarke

(2023). Toward good practice in thematic analysis: Avoiding common problems and be(com)ing a knowing researcher. International Journal of Transgender Health, 24(1), 1–6. https://doi.org/10.1080/26895269.2022.2129597

10.

Brouwer

Maeckelberghe

E. L. M.

van der Heide

Hein

Verhagen

(2020). Barriers in care for children with life-threatening conditions: A qualitative interview study in the Netherlands. BMJ Open, 10(6), Article e035863. https://doi.org/10.1136/bmjopen-2019-035863

11.

Chu

White

Stone

(2019). Prognostication in palliative care. Clinical Medicine, 19(4), 306–310. https://doi.org/10.7861/clinmedicine.19-4-306

12.

Clinical Excellence Commission . (nd). Death screening and database project. NSW Government. https://www.cec.health.nsw.gov.au/improve-quality/teamwork-culture-pcc/person-centred-care/end-of-life/death-screening-and-database-project

13.

Cohen

(1992). A power primer. Psychological Bulletin, 112(1), 155–159. https://doi.org/10.1037//0033-2909.112.1.155

14.

Contro

Larson

Scofield

Sourkes

Cohen

(2002). Family perspectives on the quality of pediatric palliative care. Archives of Pediatrics and Adolescent Medicine, 156(1), 14–19. https://doi.org/10.1001/archpedi.156.1.14

15.

Cox

Moghaddam

Almack

Pollock

Seymour

(2011). Is it recorded in the notes? Documentation of end-of-life care and preferred place to die discussions in the final weeks of life. BMC Palliative Care, 10, 18. https://doi.org/10.1186/1472-684x-10-18

16.

Dunbar

Carter

Brown

(2020). ‘Place bonding’ in children's hospice care: A qualitative study. BMJ Supportive & Palliative Care, 10(3), Article e25. https://doi.org/10.1136/bmjspcare-2018-001543

17.

Dussel

Kreicbergs

Hilden

J. M.

Watterson

Moore

Turner

B. G.

Weeks

J. C.

Wolfe

(2009). Looking beyond where children die: Determinants and effects of planning a child’s location of death. Journal of Pain and Symptom Management, 37(1), 33–43. https://doi.org/10.1016/j.jpainsymman.2007.12.017

18.

Fields

Fraser

L. K.

Taylor

Hackett

(2023). What does ‘good’ palliative care look like for children and young people? A qualitative study of parents’ experiences and perspectives. Palliative Medicine, 37(3), 355–371. https://doi.org/10.1177/02692163231154300

19.

Gibson-Smith

Jarvis

S. W.

Fraser

L. K.

(2021). Place of death of children and young adults with a life-limiting condition in England: A retrospective cohort study. Archives of Disease in Childhood, 106(8), 780–785. https://doi.org/10.1136/archdischild-2020-319700

20.

Gola

Francalanza

P. C.

Galloni

Pagella

Capolongo

(2016). Architectures for paediatric palliative care: How to improve quality of life and environmental well-being. Annali dell'Istituto superiore di sanita, 52(1), 48–55. https://doi.org/10.4415/ann_16_01_10

21.

Hain

R. D. W.

(2019). Hospices and palliative care for children: Converging stories. British Medical Bulletin, 130(1), 81–88. https://doi.org/10.1093/bmb/ldz012

22.

Håkanson

Öhlén

Kreicbergs

Cardenas-Turanzas

Wilson

D. M.

Loucka

Frache

Giovannetti

Naylor

Rhee

Ramos

M. R.

Teno

Beernaert

Deliens

Houttekier

Cohen

(2017). Place of death of children with complex chronic conditions: Cross-national study of 11 countries. European Journal of Pediatrics, 176(3), 327–335. https://doi.org/10.1007/s00431-016-2837-0

23.

Hannan

Gibson

(2005). Advanced cancer in children: How parents decide on final place of care for their dying child. International Journal of Palliative Nursing, 11(6), 284–291. https://doi.org/10.12968/ijpn.2005.11.6.18296

24.

Hansson

Björk

Santacroce

S. J.

Raunkiaer

(2023). End-of-life palliative home care for children with cancer: A qualitative study on parents' experiences. Scandinavian Journal of Caring Sciences, 37(4), 917–926. https://doi.org/10.1111/scs.13066

25.

Harms

Lakens

(2018). Making ‘null effects’ informative: Statistical techniques and inferential frameworks. Journal of Clinical and Translational Research, 3(2), 382–393. https://doi.org/10.18053/jctres.03.2017S2.007

26.

Jaaniste

Coombs

Donnelly

T. J.

Kelk

Beston

(2017). Risk and resilience factors related to parental bereavement following the death of a child with a life-limiting condition. Children, 4(11), 96. https://doi.org/10.3390/children4110096

27.

Kassam

Skiadaresis

Alexander

Wolfe

(2014). Parent and clinician preferences for location of end-of-life care: Home, hospital or freestanding hospice? Pediatric Blood and Cancer, 61(5), 859–864. https://doi.org/10.1002/pbc.24872

28.

Kennedy

Shearsmith

Ayres

Lindner

O. C.

Marston

Pass

Danson

Velikova

(2021). Online monitoring of patient self-reported adverse events in early phase clinical trials: Views from patients, clinicians, and trial staff. Clinical Trials, 18(2), 168–179. https://doi.org/10.1177/1740774520972125

29.

Kurashima

A. Y.

Latorre Mdo

Teixeira

S. A.

De Camargo

(2005). Factors associated with location of death of children with cancer in palliative care. Palliative & Supportive Care, 3(2), 115–159. https://doi.org/10.1017/s1478951505050194

30.

Lee

DeGroote

N. P.

Brock

K. E.

(2023). Early versus late outpatient pediatric palliative care consultation and its association with end-of-life outcomes in children with cancer. Journal of Palliative Medicine, 26(11), 1466–1473. https://doi.org/10.1089/jpm.2023.0063

31.

Malcolm

Knighting

Taylor

(2020). Home-based end of life care for children and their families - A systematic scoping review and narrative synthesis. Journal of Pediatric Nursing, 55, 126–133. https://doi.org/10.1016/j.pedn.2020.07.018

32.

Mherekumombe

M. F.

(2018). From inpatient to clinic to home to hospice and back: Using the “pop up” pediatric palliative model of care. Children, 5(5), 55. https://doi.org/10.3390/children5050055

33.

Mherekumombe

M. F.

Frost

Hanson

Shepherd

Collins

(2016). Pop up: A new model of paediatric palliative care. Journal of Paediatrics and Child Health, 52(11), 979–982. https://doi.org/10.1111/jpc.13276

34.

Montel

Laurence

Copel

Pacquement

Flahault

(2009). Place of death of adolescents and young adults with cancer: First study in a French population. Palliative & Supportive Care, 7(1), 27–35. https://doi.org/10.1017/S1478951509000054

35.

Noyes

Herbert

Moloney

Irving

Bradford

(2022). Location of end-of-life care of children with cancer: A systematic review of parent experiences. Pediatric Blood and Cancer, 69(6), Article e29621. https://doi.org/10.1002/pbc.29621

36.

Odejide

O. O.

Fisher

Kushi

L. H.

Chao

C. R.

Vega

Rodrigues

Josephs

I. A.

Brock

K. E.

Buchanan

Casperson

Cooper

R. M.

Fasciano

K. M.

Kolevska

Lakin

J. R.

Lefebvre

Schwartz

C. M.

Shalman

D. M.

Wall

C. B.

Wiener

Mack

J. W.

(2022). Patient, family, and clinician perspectives on location of death for adolescents and young adults with cancer. JCO Oncology Practice, 18(10), e1621–e1629. https://doi.org/10.1200/OP.22.00143

37.

Paediatric Palliative Care Australia & New Zealand . (2023). A practical guide to palliative care in paediatrics (The green book). Paediatric Palliative Care Australia & New Zealand. https://paediatricpalliativecare.org.au/resource/a-practical-guide-to-palliative-care-in-paediatrics/

38.

Papadatou

Kalliani

Karakosta

Liakopoulou

Bluebond-Langner

(2021). Home or hospital as the place of end-of-life care and death: A grounded theory study of parents' decision-making. Palliative Medicine, 35(1), 219–230. https://doi.org/10.1177/0269216320967547

39.

Peláez-Cantero

M. J.

Morales-Asencio

J. M.

Navarro-Mingorance

Á.

Madrid-Rodriguez

Tavera-Tolmo

Á.

Escobosa-Sánchez

Martino-Alba

(2023). End of life in patients attended by pediatric palliative care teams: What factors influence the place of death and compliance with family preferences? European Journal of Pediatrics, 182(5), 2369–2377. https://doi.org/10.1007/s00431-023-04870-z

40.

Rainsford

Phillips

C. B.

Glasgow

N. J.

MacLeod

R. D.

Wiles

R. B.

(2018). The ‘safe death’: An ethnographic study exploring the perspectives of rural palliative care patients and family caregivers. Palliative Medicine, 32(10), 1575–1583. https://doi.org/10.1177/0269216318800613

41.

Rattner

(2021). COVID-19: Encountering never-before-known suffering. Journal of Social Work in End-of-Life and Palliative Care, 17(2–3), 104–107. https://doi.org/10.1080/15524256.2021.1881692

42.

Roberts

Dowell

Nie

J. B.

(2019). Attempting rigour and replicability in thematic analysis of qualitative research data: A case study of codebook development. BMC Medical Research Methodology, 19(1), 66. https://doi.org/10.1186/s12874-019-0707-y

43.

Sagha Zadeh

Eshelman

Setla

Sadatsafavi

(2018). Strategies to improve quality of life at the end of life: Interdisciplinary team perspectives. The American Journal of Hospice & Palliative Care, 35(3), 411–416. https://doi.org/10.1177/1049909117711997

44.

Siden

Miller

Straatman

Omesi

Tucker

Collins

J. J.

(2008). A report on location of death in paediatric palliative care between home, hospice and hospital. Palliative Medicine, 22(7), 831–834. https://doi.org/10.1177/0269216308096527

45.

Sneha

Narasimhan

Scott

J. X.

Shuba

Rajakumar

P. S.

Rani

(2023). Choice of place of the death of children with cancer during end-of-life care: Parent's perspectives in a developing country. Indian Journal of Palliative Care, 29(4), 407–411. https://doi.org/10.25259/ijpc_28_2022

46.

Stilwell

Bhatt

Mehta

Carter

Bisset

Soanes

Shankar

(2022). Preferred place of death in paediatric, teenage and young adult haemato-oncology patients: A retrospective review. BMJ Supportive & Palliative Care, 12(e5), e650–e653. https://doi.org/10.1136/bmjspcare-2019-002005

47.

Trowbridge

Walter

J. K.

McConathey

Morrison

Feudtner

(2018). Modes of death within a children's hospital. Pediatrics, 142(4), Article e20174182. https://doi.org/10.1542/peds.2017-4182

48.

Ulrich

C. M.

Mooney-Doyle

Grady

(2018). Communicating with pediatric families at end-of-life is not a fantasy. The American Journal of Bioethics: AJOB, 18(1), 14–16. https://doi.org/10.1080/15265161.2017.1401175

49.

Vickers

Thompson

Collins

G. S.

Childs

Hain

Paediatric Oncology Nurses' Forum/United Kingdom Children's Cancer Study Group Palliative Care Working Group . (2007). Place and provision of palliative care for children with progressive cancer: A study by the Paediatric Oncology Nurses' Forum/United Kingdom Children's Cancer Study Group palliative care working group. Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology, 25(28), 4472–4476. https://doi.org/10.1200/jco.2007.12.0493

50.

Wahid

A. S.

Sayma

Jamshaid

Kerwat

Oyewole

Saleh

Ahmed

Cox

Perry

Payne

(2018). Barriers and facilitators influencing death at home: Q meta-ethnography. Palliative Medicine, 32(2), 314–328. https://doi.org/10.1177/0269216317713427

51.

Weaver

M. S.

Rosenberg

A. R.

Fry

Shostrom

Wiener

(2021). Impact of the coronavirus pandemic on pediatric palliative care team structures, services, and care delivery. Journal of Palliative Medicine, 24(8), 1213–1220. https://doi.org/10.1089/jpm.2020.0589

52.

Wiener

Tager

Mack

Battles

Bedoya

S. Z.

Gerhardt

C. A.

(2020). Helping parents prepare for their child's end of life: A retrospective survey of cancer-bereaved parents. Pediatric Blood and Cancer, 67(2), Article e27993. https://doi.org/10.1002/pbc.27993

53.

Winger

Kvarme

L. G.

Løyland

Kristiansen

Helseth

Ravn

I. H.

(2020). Family experiences with palliative care for children at home: A systematic literature review. BMC Palliative Care, 19(1), 165–184. https://doi.org/10.1186/s12904-020-00672-4

54.

Zelcer

Cataudella

Cairney

A. E. L.

Bannister

S. L.

(2010). Palliative care of children with brain tumors: A parental perspective. Archives of Pediatrics & Adolescent Medicine, 164(3), 225–230. https://doi.org/10.1001/archpediatrics.2009.284

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB