key: cord-284145-7le8303x
authors: Taylor, Johanna; Booth, Alison; Beresford, Bryony; Phillips, Bob; Wright, Kath; Fraser, Lorna
title: Specialist paediatric palliative care for children and young people with cancer: A mixed-methods systematic review
date: 2020-05-02
journal: Palliat Med
DOI: 10.1177/0269216320908490
sha: 
doc_id: 284145
cord_uid: 7le8303x

BACKGROUND: Specialist paediatric palliative care services are promoted as an important component of palliative care provision, but there is uncertainty about their role for children with cancer. AIM: To examine the impact of specialist paediatric palliative care for children and young people with cancer and explore factors affecting access. DESIGN: A mixed-methods systematic review and narrative synthesis (PROSPERO Registration No. CRD42017064874). DATA SOURCES: Database (CINAHL, Cochrane Database of Systematic Reviews, Embase, MEDLINE, PsycINFO) searches (2000–2019) identified primary studies of any design exploring the impact of and/or factors affecting access to specialist paediatric palliative care. Study quality was assessed using The Mixed Methods Appraisal Tool. RESULTS: An evidence base of mainly low- and moderate-quality studies (n = 42) shows that accessing specialist paediatric palliative care is associated with less intensive care at the end of life, more advance care planning and fewer in-hospital deaths. Current evidence cannot tell us whether these services improve children’s symptom burden or quality of life. Nine studies reporting provider or family views identified uncertainties about what specialist paediatric palliative care offers, concerns about involving a new team, association of palliative care with end of life and indecision about when to introduce palliative care as important barriers to access. There was evidence that children with haematological malignancies are less likely to access these services. CONCLUSION: Current evidence suggests that children and young people with cancer receiving specialist palliative care are cared for differently. However, little is understood about children’s views, and research is needed to determine whether specialist input improves quality of life.

• • Specialist paediatric palliative care is promoted as an important component of children and young adult cancer services, but there is uncertainty about the factors that affect access and the benefits for children who receive this specialist input. • • Three reviews, which have aggregated evidence for children with all life-limiting conditions, suggest that the benefits of specialist palliative care include less time in hospital and improvements in quality of life and symptom management. • • The growing number of studies investigating the role of specialist palliative care for children with cancer report mixed results and varying provision, and there is a need to aggregate this evidence to inform future policy and practice.

• • Accessing specialist paediatric palliative care is associated with less intensive care at the end of life, more advance care planning and fewer in-hospital deaths for children and young people with cancer, but there is no robust evidence to tell us whether these services lead to improvements in quality of life or symptom management. • • Children with haematological malignancies are less likely to receive specialist palliative care than children with other cancers. • • Uncertainty about when to introduce palliative care services to families, what it comprises and the added value of specialist input was identified as a key barrier to access, as were perceptions that paediatric oncology teams already meet the palliative care needs of their patients.

• • Evidence is still needed to determine whether specialist paediatric palliative care improves the quality of life and symptom management for children and young people with cancer. • • Exploration of why children with certain cancers are less likely to receive specialist palliative care at the end of life may help address this inequality in access. • • A core outcome set study including the views of children and families would help improve future aggregation of evidence in this area.

Approximately 20% of children and young people diagnosed with cancer do not survive despite significant medical advances in recent decades. 1, 2 The majority of deaths are due to the malignancy, with some attributable to anticancer treatments. 3 Distress from symptoms and suffering during the end-of-life phase can be significant, [4] [5] [6] impacting on the child and their family's quality of life. 7, 8 In addition, many children and young people who die from cancer continue to have high-intensity treatments towards the end of life, 9, 10 with nearly half dying in the acute care setting, 11 despite preferences from the majority of children and their parents for being at home during the end-of-life phase. 12, 13 Palliative care, defined by the World Health Organization (WHO) as 'the active total care of the child's body, mind and spirit . . . [that] begins when illness is diagnosed, and continues regardless of whether or not a child receives treatment directed at the disease', 14 is recognised as an important component of children and young people's cancer services. 15, 16 In addition, and over the last 30 years, specialist paediatric palliative care services for children and young people have been developing in many countries including the United States, Canada, United Kingdom and across Europe. 17 The English National Health Service (NHS) defines specialist paediatric palliative care as 'a consultant-led multi-professional specialist palliative care team . . . led by a medical consultant working at Paediatric Palliative Care Competency Level 4'. 16 In practice, however, the models of providing specialist palliative care vary within and between countries, including, for example, hospital-and community-based teams which support all children with life-limiting conditions, 18 teams embedded within paediatric oncology departments, joint working with, or hospice-led provision, and specialist nurse-led teams as well as services led by a paediatric palliative care consultant. 17, 19, 20 Even within the developed world, the availability, referral and uptake of specialist palliative care among children and young people with cancer remains low and variable between and within countries and settings, and it is not clear to what extent these services are addressing all aspects of palliative care as defined by WHO. 17, [21] [22] [23] [24] [25] For children with cancer, referral to palliative care also often occurs late in the trajectory of illness, sometimes only days before death. 26, 27 Recent systematic reviews suggest that access to specialist palliative care services is associated with improvements in quality of life, symptom control, perceived support, reduced time in hospital, less invasive treatment and greater advance care planning. 19, 28, 29 However, these reviews have aggregated the results for children and young people across conditions, and the evidence for those with cancer remains unclear because of conflicting results between individual studies 28 and the lack of work exploring condition-specific factors that may influence access to and benefit from specialist palliative care services. 30, 31 A rigorous review of the evidence on the impacts of specialist paediatric palliative care for children and young people with cancer and their families is both crucial to informing debates within paediatric oncology regarding the positioning and role of these specialist services and for future service development. This mixed-methods systematic review synthesises the existing evidence on the benefits, drawbacks, facilitators and barriers associated with referral to and uptake of specialist paediatric palliative care for children and young people with cancer and their families.

The review questions are as follows:

1. What are the reported benefits and drawbacks of referral to specialist paediatric palliative care for children and young people with cancer and their families? 2. What are the factors (e.g. barriers, facilitators) affecting referral to and uptake of specialist paediatric palliative care for children and young people with cancer?

The protocol registration number is CRD42017064874. 32 

Primary studies of any design (e.g. experimental, observational, surveys, consensus and qualitative studies) examining either the impact of or factors affecting specialist paediatric palliative care access for children and young people (age 0-24 years) with cancer were included. Studies of a mixed population were included if (1) the majority of the participants were children and young people with cancer and/or (2) data were reported separately. We defined specialist paediatric palliative care as care provided by multidisciplinary teams or palliative care services which included clinicians (e.g. oncologists, paediatricians, nurses) with paediatric palliative care training, or services who self-identified as providing specialist paediatric palliative care. We included services delivered in different settings (e.g. inpatient, community or home settings) and both liaison services (e.g. supporting the child's usual care team) and services directly supporting children and their families. Although broad, this reflects the varying provision of specialist palliative care for children within and between different countries and so enabled us to synthesise the evidence about these specialist services.

To understand the different perspectives on referral to specialist palliative care services, studies that included the following participant groups were eligible: children and young people; parents (including bereaved); other family members and health and social care staff. No comparator was required.

We excluded case studies, review articles, descriptive, theoretical or clinical opinion articles, conference abstracts and articles not published in the English language. We also restricted the eligibility to studies conducted in highincome countries (defined as OECD (Organisation for Economic Co-operation and Development) 33 member countries) because of the very different healthcare infrastructure and status of specialist paediatric palliative care in developing countries. 17 

Electronic databases (CINAHL, Cochrane Database of Systematic Reviews (CDSR), Embase, MEDLINE, PsycINFO) were searched on 27 June 2017 (from 2000, in line with increasing availability of specialist paediatric palliative care services internationally 19 ). The search strategy consisted of terms and synonyms for [ 

Titles and abstracts were screened in Covidence, 34 and relevant full-text articles were retrieved and independently assessed for eligibility by two reviewers with disagreements resolved via a third reviewer. 35 Reasons for exclusion at full text were recorded.

Data on study characteristics, methods, study focus (e.g. impact and/or factors affecting access) and quantitative outcome data were extracted into Microsoft Excel using a pre-piloted data extraction template. Qualitative data, including author-reported results, direct quotations and results tables, were imported into NVivo version 11 36 for analysis.

Data extraction was carried out by one reviewer and checked by a second.

The Mixed Methods Appraisal Tool (MMAT) 37,38 was used to appraise the methodological quality of all included studies. MMAT comprises two generic screening questions and an additional four criteria for use with specific study designs. Criterion assessments (e.g. Is the sample representative of the population under study?) are categorised 'Yes', 'No' or 'Can't tell'. An overall quality score (of 0%, 25%, 50% or 100%) based on how many study design specific criterion were met (those categorised as 'Yes') was calculated for each study.

The quality assessment was undertaken independently by two reviewers and informed the synthesis methods and reporting of the review results, along with identifying needs for future research. We did not exclude studies based on quality assessment.

An integrative narrative synthesis was planned drawing on interpretative review methodology 39 with thematic analysis as the principal method of synthesis. 40 The synthesis plan outlined in the protocol 32 was modified following an assessment of the potential for aggregation, configuration and integration of study findings. The final method involved separate syntheses of data reporting the impact of specialist paediatric palliative care and factors affecting access (both included quantitative and qualitative data). Each comprised the following steps:

Data reduction. This involved reviewing and summarising extracted data and identifying recurring categories from across the studies, and distinct service types. Each synthesis included quantitative and qualitative data and all numerical and statistical findings were converted to descriptive summaries. For qualitative data, 'meaning units' (comprehensible segments of text which contain one idea or piece of information) were identified and named (or 'coded') drawing both on author-reported results and participant quotes to capture the full range of concepts or themes across the studies. 41 Data display. Using the method of constant comparison, 39 descriptive summaries of quantitative data and coded qualitative data were compared to ensure that similar data were grouped together to develop a thematic coding matrix consisting of descriptive themes and overarching categories which grouped similar themes together. This was performed by one reviewer with regular input from the wider review team for sense checking and validation. Data from each study were then synthesised into the coding matrix, retaining reference to the service type and critical appraisal score to facilitate greater interpretation. Data display techniques were used to illustrate the spread of themes across studies and specialist paediatric palliative care models, and narrative weaving describes the results of each synthesis. 42 

A total of 8549 unique records were screened by title and abstract, 626 full-text articles were retrieved and reviewed, and 49 articles describing 42 studies 23, were included in the review (see Figure 1 ).

Of the 42 included studies, 11 examined the impact of specialist paediatric palliative care, 44, 48, 57, 58, 62, 64, 66, 70, 73, 74, 88 14 explored factors affecting access 23, 43, 45, 56, 65, 68, 71, 75, 76, [78] [79] [80] [81] [82] [83] 89 and 17 studies investigated both 46, 47, [49] [50] [51] [52] [53] [54] [55] [59] [60] [61] 63, 67, 69, 72, 77, [84] [85] [86] [87] 90 (see Table 1 for study characteristics). Using the MMAT, 25 studies were categorised as quantitative non-randomised, 12 as quantitative descriptive and 5 as qualitative studies. The majority of studies were conducted in the United States (n = 28). Others were in the United Kingdom (n = 3), 23, 47, 76 Canada (n = 3), 50, 51, 63, 72 Germany (n = 3), 57, 73, 90 France (n = 1), 87 Switzerland (n = 1), 79 Israel (n = 1) 49 and two in multiple countries. 65, 71 Study populations Thirty-one studies (all quantitative) examined the impact of and/or characteristics of children receiving specialist palliative care; 24 included a comparator group of children not receiving this, 23,43-51,54-64,78,86-88, 90 1 compared children receiving late and early specialist input, 52, 53, 84, 85 and 6 used a single-group study design. 66, 67, 69, 70, 73, 74 Of the remaining 11 studies (6 quantitative and 5 qualitative), 10 explored the views of healthcare staff 65, 71, 72, [75] [76] [77] [79] [80] [81] [82] [83] 89 and 1 the views of parents and young people. 68 Of the 31 studies examining outcomes and/or characteristics of children, the majority (n = 21) drew their sample of children and young people from a single centre. Several studies used the same or potentially overlapping samples as other included studies (see Figure 2 ).

In total, data for 7933 children and young people, 4289 of whom had received specialist palliative care compared to 3644 who did not, were included. While the majority (n = 23) included children and young people with any cancer, eight studies concerned children and young people with particular diagnoses or treatments (see Table 1 for details). 44, 56, 61, 62, 66, 67, 69, 78 Overall, these 31 studies address diagnoses from infancy to young adulthood, with only three studies focused on young adults, which included any malignancy. 54, 58, 86 Five studies included some children and young people with conditions other than cancer. 61, 67, 70, 73, 88 Of the 11 studies exploring stakeholder views, 3 recruited paediatric oncology staff from single hospitals 75, 77, 82, 83 and 1 from multiple hospitals, 79,80 1 recruited staff involved in providing palliative care to children with cancer from primary, tertiary and community settings, 76 4 recruited paediatric oncologists via professional organisations 65,71,72,81 and 1 recruited parent and young person dyads from three hospitals. 68 In total, these studies represented the views of 1133 physicians, 986 other healthcare professionals (mainly nurses, social workers and other staff working in paediatric oncology, but also in palliative care and other settings), 129 parents and 129 young people.

We identified five broad service types from the included studies: hospital-based palliative care teams with referral triggering an initial consultation 46, [50] [51] [52] [53] 58, [61] [62] [63] [64] 67, 69, [73] [74] [75] 77, [82] [83] [84] [85] 87, 88 (n = 17), hospice services 23, 45, 47, 55, 56, 59, 60, 65, 86 (n = 8), home-based services 48, 57, 70, 90 (n = 4), integrated oncology services 44,66 (n = 2) and an integrated oncology inpatient unit 49 (n = 1). Three studies referred to palliative care consultation only, 43, 54, 78 and three a palliative care team, 68, 79, 81 although it is likely that these were all referring to hospital-based teams. Of the remaining four studies, three included a range of models 71, 72, 89 and one referred to specialist paediatric palliative care explicitly but did not define this. 76 Very few studies provided specific details about the team or service providing palliative care in terms of skills mix, role and the extent of provision. There were also some anomalies between studies; for example, five studies used hospice discussions or enrolment as an outcome of specialist palliative care, 46, 61, 62, 64, 90 whereas in other studies hospices were defined as the source of specialist intervention. Not all studies investigating hospice settings specified the characteristics (e.g. adult vs children's hospice), and it was difficult to determine what 'home-based services' might comprise. Among the studies exploring stakeholder views, three offered only 'hypothetical models' for participants to consider 68, 75, 76 and three focused primarily on views about early integration of specialist palliative care, which was defined as close to the time of diagnosis. 68, 75, 81 Only two studies accounted for the timing of the initial palliative care consult/enrolment when deciding which children to classify as having received specialist intervention; one only included children who had received specialist palliative care for more than 30 days 63 and the second for more than 1 day. 54 A third study, which compared early and late interventions, included all children receiving specialist palliative care but defined early provision as that received for more than 30 days. 52, 53 Only two studies explored how the duration of specialist paediatric palliative care exposure affected outcomes. 61, 70 Study quality

The quality and reporting of studies varied greatly among the 37 quantitative studies, with scores ranging from 0% to 100% (see Table 2 ).

There were concerns about the representativeness of samples in 14 of the quantitative studies, due to the single-site design of many studies, long study period and inappropriate participant selection or recruitment methods. 44, 46, [48] [49] [50] [51] 54, 57, 61, 62, 64, 70, 72, 86, 89 Assessment of representativeness was not possible in another seven studies because of poor reporting. 66, 68, 73, 78, 81, 89, 90 In eight studies, there were concerns about the comparability of groups (e.g. use of historical cohorts, different participant characteristics), which were not accounted for in the analysis. 44, 50, 51, 54, 55, 57, 58, 62, 90 Although inappropriate measures were identified in only three studies, 50, 51, 56, 66 wider concerns about measurement were evident across studies. These included how receipt of specialist palliative care was determined (e.g. from day of initial palliative care consultation); whether outcomes could be attributed to specialist provision or care from the primary oncology team; potential recall problems in studies using bereaved parent-reported outcomes and failure to take account of the chronology of variables. For example, having a do-not-resuscitate (DNR) order was a predictor of specialist palliative care in some studies 46, 59, 60 and an outcome in others. 54, 58 Determining the extent of these limitations was hindered by poor reporting of study methods.

The five qualitative studies 75-77,79,83 used appropriate methods for sampling, data collection and analysis. However, there were limitations to the transferability of findings due to the single-site design in three studies, [75] [76] [77] 83 and for the third, the time period elapsed since the study was conducted (published in 2001 when specialist paediatric palliative care for children with cancer was at an early stage of development). 76 The impact of specialist paediatric palliative care A total of 17 distinct outcome domains were identified and these were categorised under one of six overarching categories: advance care planning (n = 9 studies), end-oflife care provision (n = 16), location of death (n = 18), child's quality of life (n = 6), family support (n = 3) and service satisfaction (n = 7). The synthesis of results by category and outcome domain is summarised below and in detail in Table 3 .

The outcome domains were of end of life discussions and decisions about attempting resuscitation (DNR orders), and the timing of these. Provision of specialist palliative care was found to be associated with an increased likelihood of end-of-life discussions being documented in all six studies that measured this, 50, 54, 61, 62, 64, 74 with evidence from two studies that these occurred earlier in the child's illness in those receiving specialist palliative care. 61, 64 Six studies found that DNR orders occurred earlier in care trajectories in those receiving specialist palliative care. 46, 54, 55, 58, 62, 64 However, the evidence about whether a DNR order was more likely was mixed. 46, 54, 55, 61, 62, 64 End-of-life care provision There was evidence from the seven studies which measured treatment intensity 54, 58, 61, 63, 64, 87, 88 that children who received specialist paediatric palliative care were less likely to receive high-intensity treatments and to spend less time in an intensive care unit (ICU) during the end-of-life phase, compared to children who did not receive this. There was also evidence from the five studies which assessed hospital admissions (either the duration of stay or the number of admissions) that children who received specialist palliative care spent less time in the hospital than those who did not. 44, 47, 55, 70, 90 Only one study compared the types of admission; here specialist palliative care was only found to decrease the number of planned admissions. 47 Two studies examined the cost of care, also focusing on hospital admissions, but did not find any significant differences between before and after specialist input, 70 or between those who received specialist palliative care and those who did not. 54 There was no evidence to suggest that hospice care, either enrolment or utilisation, changed as a result of specialist involvement from the four studies that explored this. 46, 61, 62, 90 Outpatient care, 90 home-based care 57 and preferred location of care 51 were each assessed by a single study so no conclusions can be drawn about these.

A total of 17 studies examined differences in location of death between children receiving and not receiving specialist palliative care. A consistent finding across studies was that children who received this were less likely to die in ICU. 49, 52, 54, 57, 58, 61, 64, 84 However, studies varied in whether or not they found differences in the proportion of home or hospital deaths. 46, 48, 57, 59, 60, 62, 64, 73, 90 For example, the study comparing early and late involvement of a hospital team found that 'late access' children were nearly five times more likely to die in ICU than at home or in a hospice, but observed no differences in terms of non-ICU hospital versus home/hospice deaths. 52 However, they did find lower odds of hospital deaths when comparing hospice and no hospice involvement. 84 Just four studies investigated the impact of specialist palliative care on whether children died in their/families preferred location of death, 48, 51, 57, 90 three of which found that congruence between preferred and actual place of death increased with specialist involvement. 48, 51, 90 Satisfaction with the information received, medical care and PC, rated on a Likert-type scale (1 = very good, 6 = unsatisfactory) Overall satisfaction across the items was very good, with a mean rating of 1.6 (1 = very good, 6 = unsatisfactory). There were no significant differences in satisfaction ratings between parents whose child died at home (which was preferred by most families) or in hospital Szymczak et al. 77 Paediatric oncology providers were asked about their perceptions of SPPC service and how these may influence timing of referral The SPPC service was highly regarded by oncology providers, who identified the following impacts: the SPPC team being able to spend time with families and provide emotional support and explore wider needs, expertise in pain and symptom management, supporting transition from hospital to home, around-theclock support, support for parents and siblings, and helping families to clarify goals QL 100

All SPPC models Wentlandt et al. 72 Paediatric oncologists asked about the quality of SPPC services in a survey More than 83% of participants agreed or strongly agreed that they were satisfied with the quality of SPPC services 

This theme included three outcome domains: quality of life, comfort and symptoms. Data were primarily from parent-reported measures collected after a child had died. Two studies measured quality of life but did not use validated measures; one found that children receiving specialist paediatric palliative care had more fun and more events adding meaning compared to children not receiving this, but found no differences in how afraid or peaceful they felt. 48 In the second, 70% of parents strongly agreed that these specialist services improved their child's quality of life, but there was no comparator group. 66 The study which measured comfort, also with no comparator, found an increase in comfort levels over time reported by parents but no change reported by children. This was the only study that used a child-reported measure. 67 Although three studies suggested that treatment of symptoms increased with specialist involvement, 57, 74, 90 there was little evidence that the extent of symptom control/suffering from physical and emotional symptoms differed between children receiving and not receiving specialist palliative care. Indeed, high levels of suffering from symptoms were described by the four studies which assessed this. 48, 57, 64, 90 Family support

Only three studies investigated whether specialist paediatric palliative care affected provision of support to family members, or the impact of that support. One study found that use of bereavement support for parents and siblings was more likely if a child had received specialist input. 62 Two studies found that specialist palliative care increased parent preparedness for the end-of-life phase of care. 64,66

All seven studies (six surveys 66, 67, 69, 72, 73, 90 and one qualitative 77 ) assessing this reported high levels of family and professional satisfaction with specialist palliative care, with reported benefits including expertise in pain and symptom management, time to plan end-of-life care with families and meeting psychosocial and family needs. None of these studies included a comparator group, although one reported increasing parental satisfaction in line with increasing provision of specialist services.

The synthesis of studies exploring factors affecting specialist paediatric palliative care access identified four overarching categories each containing several linked themes: sociodemographics (n = 14 studies), disease profile (n = 22 studies), end-of-life care characteristics (n = 18 studies) and acceptability of specialist paediatric palliative care (n = 17 studies). The synthesis of results by category and theme are summarised below and presented in detail in Table 4 .

Multiple quantitative studies which investigated the sociodemographic profile of children and/or their families receiving and not receiving specialist palliative care consistently showed that access was not associated with a child's gender 47 

Out of 12 studies examining type of cancer, 10 found that children with solid tumours were more likely to receive specialist palliative care than children with haematological malignancies. 46, 47, 49, 50, [53] [54] [55] 59, 61, 63, 78, 86 Six studies investigated whether disease status (e.g. prognosis, relapse) was associated with access; no consistency was found across these, with some studies reporting conflicting resu lts. 46, 49, 50, 54, 55, 61 In contrast, all the studies investigating staff-reported practices consistently identified children with a poor prognosis as those most likely to be refer red. 65, 71, 72, 75, 77, 79, 81, 83, 89 Staff in two studies believed that this could result in referrals that were too late for children to benefit from specialist input. 71, 79 Uncertainty about a child's prognosis and about the benefits of introducing specialist palliative care earlier in the disease trajectory (e.g. around diagnosis) were identified as key barriers to a timely referral, 65, 71, 75, 77, 79, 81, 83, 89 although non-physician professionals 75, 81 and families 68 were more receptive to early integration than physicians. Automatic referral triggers were used by the majority of providers in one study. However, in line with reported practice, most encouraged referrals for children during the end-of-life phase of care. 71

The end-of-life care provided to children was found to influence whether or not children received specialist palliative care. Four studies consistently found that advance care planning (e.g. documented advance directives or DNRs, hospice referrals) was associated with an increased likelihood of specialist involvement, or earlier compared to late involvement. 46, 53, 55, 59 Provision of palliative care by the oncology team (we already do palliative care) was also reported by oncology staff to influence referral practices 65, 71, 72, 75, 79 and to impede oncologists' abilities and willingness to identify needs for specialist input in their patients and practice, something that was reported to encourage referrals. 65, 71, 72, [75] [76] [77] 79, 89 Similarly, young people were more likely to accept specialist input if they had specific unmet needs, for example, pain and poor quality of life, 68 although staff in one study were concerned that parents might not discuss their child's palliative care needs during active treatment. 81 There was conflicting evidence about whether treatment intensity (e.g. phase 1 trial enrolment) influenced access to specialist palliative care, 43, 46, 50, 53, 56, 61 and this mirrored varying beliefs among oncology staff and families about whether children receiving active treatments should or could be referred. 65, 68, 71, 72, 75 Acceptability of specialist paediatric palliative care Staff acceptability of specialist palliative care was reported to influence access in 16 studies. Family acceptability was identified in 13, although families' views were sought in only three of these. [67] [68] [69] Eight studies measured intervention uptake to explore acceptability. 23, 46, 61, 63, 67, 69, 89, 90 These reported varying rates of referral and uptake of specialist palliative care, but provided evidence of increasing involvement over time.

Staff uncertainties about the benefits of specialist palliative care, and about how a specialist service differs from care provided within oncology (the added value of specialist paediatric palliative care), were identified as key barriers to referral. 68, [75] [76] [77] 79, 81, 83, 89 Concerns that bringing in a new team could undermine continuity of care and impact on relationships with families were also reported to influence referral practices, 68, 72, [75] [76] [77] 79, 89 as were perceptions about availability and capacity of specialist palliative care services (perceived availability). 65, 71, 72 Staff in eight studies identified the association of palliative care with end of life (what specialist paediatric palliative care symbolises) as a barrier to access. 72, [75] [76] [77] 79, 81, 83, 89 A clinician's emotion about a family and their readiness to accept a child's prognosis and discuss this were identified as additional challenges, 75, 77, 79, 81, 83, 89 and contributed to what was described as the emotional labour associated with introducing a service which staff participants referred to as 'death nurses', 76 'the death team' 77 and 'the death squad'. 75 A perception that oncologists need to control patient care ('clinical ownership') was identified as a further barrier to specialist palliative care referral and could deter others involved in a child's care from recommending this for a family. 75, 81, 83, 89 Oncology staff reported that family readiness 75, 77, 81, 83 and family resistance 65, 71, 72, 79, 89 could affect access, although views on this varied. For example, 60% of paediatric oncologist respondents in one survey believed that their patients had negative perceptions of specialist palliative care, but very few (4%) reported that patients refused a referral. 72 In another study, only 38% of staff participants identified family resistance as a barrier, 89 and the study that explored young people and parents' attitudes found mainly positive views, and very few participants (2% and 8%, respectively) believed that referral was associated with the loss of hope for a cure, 68 which was a fear commonly reported by staff. 72, [75] [76] [77] 79, 83, 89 

This systematic review found evidence that children and young people with cancer who receive specialist paediatric palliative care are more likely to be engaged in advance care planning, receive less intensive care at the end of life and are less likely to die in hospital, compared to those who do not receive this. Some of the included studies also indicate that these differences may be more marked when children receive specialist input for a longer duration before they die. The review did not find that receipt of specialist palliative care is associated with improved quality of life or symptom control; however, no conclusions can be drawn because of the significant methodological limitations of the seven studies investigating this. 48, 57, 58, 64, 66, 67 Importantly, only one study sought young people's views about the impact of these specialist services across the 28 studies which examined this, and this was a feasibility study with no comparator group. 67 The review also found that the type of cancer and whether or not paediatric oncology teams themselves engage in palliative care practices may affect access to specialist palliative care services. Thus, studies reported fewer children with haematological malignancies receiving specialist palliative care, and involvement more likely where the oncology team were proactively addressing palliative care needs (e.g. evidence of advance care planning). There was no indication from the studies included about why children with certain cancers are less likely to receive specialist palliative care, and a recent review of barriers to access did not explore this. 31 Evidence from adult cancer 91 indicates that the remitting and relapsing trajectories of haematological malignancies, a more aggressive approach to treatment and greater difficulties predicting prognosis may contribute to the observed inequity of specialist palliative care involvement. 92 Evidence from our review regarding clinician uncertainty about when to involve specialist palliative care, combined with the commonly reported practice of referring children with a poor prognosis, 27 suggests this might be the case.

Clinicians' views regarding the need for specialist palliative care input for their patients and their acceptability of specialist palliative care services were also consistently reported as affecting referral practices. More specifically, views on how these specialist services differ to palliative care provided by the oncology team, perceived drawbacks associated with involving a new team, readiness to accept a change in prognosis and the negative connotations associated with the term 'palliative care' were identified as barriers to referral. Concerns among clinicians about how families might react to the offer of a referral was a common theme in several studies, and although the study that explored young people and parents' views runs counter to this, wider work on this issue reports mixed opinions among young people 93, 94 and parents 73,74 about palliative care and how this should be introduced. 95 

This review is the first to systematically synthesise the available evidence about specialist paediatric palliative care for children and young people with cancer. Strengths of the review include a published protocol, robust search, independent screening and data extraction by two reviewers, and the use of appropriate mixed-methods techniques to synthesise the results. There are, however, limitations in the conclusions which can be drawn from this review due to the heterogeneity of study populations and interventions. This, and the substantial risk of potential bias identified in some of the studies and inconsistency of measurement across studies, meant it was inappropriate to aggregate the results statistically, or to compare results between the different approaches to providing specialist palliative care. In addition, the descriptions of specialist services and the palliative care provided by oncology teams were typically very poor, making interpretation of the differences between these challenging.

Our finding that end-of-life care is different for those who receive specialist paediatric palliative care compared to those who do not broadly aligns with three recent reviews about children with all life-limiting conditions. 19, 28, 29 Two of these reviews 19, 29 concluded that specialist intervention appears to offer benefit in terms of improved quality of life. By focusing only on children and young people with cancer, our review highlights the lack of robust evidence pertaining to both quality of life and symptom burden in this population. Although the broader literature implies that other differences we observed, such as reduced hospital stays and more advance care planning, are indicative of better care, 63 there is no evidence that these changes in the delivery of care reflect family preferences, 48, 51, 57 or lead to reduced symptom burden, which studies continue to show is significant for children with cancer. 5, 90, 97 The integration of findings from evaluation studies and those which have explored factors affecting access tells us that while end-of-life care may be different for children who receive specialist palliative care, children who receive this are also different to those who do not, particularly in terms of their disease profile and care processes. Synthesis of the qualitative studies go some way to explaining these differences, highlighting in particular the central role of clinicians' emotions, beliefs and attitudes in shaping referral practices, and the ongoing uncertainty about when to initiate palliative care and whether this should be provided by a specialist service or the oncology team. It also reveals that, in practice, decisions and discussions about no longer pursuing curative treatments and introducing specialist palliative care go hand in hand, and that families who are comfortable discussing an uncertain future or families cared for by clinicians who have this confidence may be more likely to receive specialist palliative care. 75, 77, 83 Interventions that support clinicians to initiate palliative care with families and improve clinical acceptability of specialist palliative care services therefore offer the potential to address these key barriers to access. 82, 98 Although there are various initiatives to support the delivery of palliative care within oncology services (e.g. palliative care training, 99,100 communication tools, 101,102 early integration models 26,67,69 ), we know very little about whether these are being implemented in practice and how they might influence referral to specialist palliative care. There is also little available evidence about whether families play an active role in the initiation of palliative care, or whether clinicians' concerns about how families will respond to a referral are warranted. 68 Future research should therefore investigate the effectiveness, delivery and acceptability of the different models of delivering palliative care for children and young people with cancer, particularly outside of North America. Future research also needs to examine factors affecting uptake from families' perspectives and to explore the role of socioeconomic factors. This conclusion is supported by the recent priority setting partnership results for teenagers and young adults with cancer, 103 which includes how best to support young people who have incurable cancer and their families. In order to undertake this research, we must first determine what outcomes are the most important to measure and develop appropriate tools to measure them. [104] [105] [106] [107] Development of a core outcome set would meet this requirement. 108 This too will need to include the views of children and young people and their families, particularly if we are to address the methodological challenges that continue to affect the quality of research in this area, and the lack of evidence about whether specialist paediatric palliative care improves quality of life for children and their families.

New policies to address the global burden of childhood cancers

Validation of a classification system for treatment-related mortality in children with cancer

Symptoms and suffering at the end of life in children with cancer

Symptoms and distress in children with advanced cancer: prospective patientreported outcomes from the PediQUEST study

Symptoms in the palliative phase of children with cancer

Quality of life in children with advanced cancer: a report from the PediQUEST study

Psychological distress in parents of children with advanced cancer

Predictors of and trends in high-intensity end-of-life care among children with cancer: a population-based study using health services data

Procedural burden experienced by children with cancer during their terminal hospital admission

Place of death in children and young people with cancer and implications for end of life care: a population-based study in England

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

Experiences and preferences for end-of-life care for young adults with cancer and their informal carers: a narrative synthesis

World Health Organization. WHO definition of palliative care for children

End of life care for infants, children and young people with lifelimiting conditions: planning and management. NICE guideline

Manual for prescribed specialised services

Pediatric palliative care provision around the world: a systematic review

Paediatric palliative care: development and pilot study of a 'Directory' of lifelimiting conditions

Specialist paediatric palliative care services: what are the benefits?

Pediatric palliative care in childhood cancer nursing: from diagnosis to cure or end of life

Availability and use of palliative care and end-of-life services for pediatric oncology patients

Geographic access to hospice care for children with cancer in Tennessee

Referral to a specialist paediatric palliative care service in oncology patients

Pediatric palliative care in Canada in 2012: a cross-sectional descriptive study

Pediatric palliative care programs in children's hospitals: a cross-sectional national survey

Early integration of palliative care for children with high-risk cancer and their families

Palliative care initiation in pediatric oncology patients: a systematic review

Effect of pediatric palliative care programs on health care resource utilization and costs among children with life-threatening conditions: a systematic review of comparative studies

Impact of specialized pediatric palliative care: a systematic review

Factors associated with the use of U.S. community-based palliative care for children with life-limiting or life-threatening illnesses and their families: an integrative review

Barriers to accessing palliative care for pediatric patients with cancer: a review of the literature

Specialist paediatric palliative care for children and young people with malignancies: a mixed methods systematic review

Organisation for Economic Co-operation Development

Veritas Health Innovation. Covidence systematic review software

EndNote. X8.0.1 (Bld 10444)

QSR International Pty Ltd, 2015. 37. National Collaborating Centre for Methods Tools. Appraising qualitative, quantitative, and mixed methods studies included in mixed studies reviews: the MMAT

Testing the reliability and efficiency of the pilot Mixed Methods Appraisal Tool (MMAT) for systematic mixed studies review

The integrative review: updated methodology

Synthesizing qualitative and quantitative health evidence

How to plan and perform a qualitative study using content analysis

Achieving integration in mixed methods designs -principles and practices

Influence of early phase clinical trial enrollment on patterns of end-oflife care for children with advanced cancer

Development of an in-home standardized end-of-life treatment program for pediatric patients dying of brain tumors

Race does not influence donot-resuscitate status or the number or timing of end-oflife care discussions at a pediatric oncology referral center

Trends in end-of-life care in pediatric hematology, oncology, and stem cell transplant patients

Does referral to specialist paediatric palliative care services reduce hospital admissions in oncology patients at the end of life?

Improved quality of life at end of life related to home-based palliative care in children with cancer

Integration of a palliative and terminal care center into a comprehensive pediatric oncology department

Differences in end-of-life communication for children with advanced cancer who were referred to a palliative care team

Parent and clinician preferences for location of end-of-life care: home, hospital or freestanding hospice?

Illness and end-oflife experiences of children with cancer who receive palliative care

Predictors of late palliative care referral in children with cancer

End-of-life care characteristics for young adults with cancer who die in the hospital

Variables influencing end-of-life care in children and adolescents with cancer

Does phase 1 trial enrollment preclude quality end-of-life care? Phase 1 trial enrollment and end-of-life care characteristics in children with cancer

Did increased availability of pediatric palliative care lead to improved palliative care outcomes in children with cancer?

Palliative care involvement is associated with less intensive end-of-life care in adolescent and young adult oncology patients

Racial and ethnic differences in hospice enrollment among children with cancer

Hospice care for children with cancer: where do these children die?

End-of-life care patterns associated with pediatric palliative care among children who underwent hematopoietic stem cell transplant

Patterns of end-oflife care in children with advanced solid tumor malignancies enrolled on a palliative care service

Predictors of specialized pediatric palliative care involvement and impact on patterns of end-of-life care in children with cancer

Easing of suffering in children with cancer at the end of life: is care changing?

Hospice referral practices for children with cancer: a survey of pediatric oncologists

Effects of a unique pediatric hematology-oncology palliative care program on medical decision-making and communication between healthcare providers and families: results of a supportive care survey

Establishing feasibility of early palliative care consultation in pediatric hematopoietic stem cell transplantation

Patients' and parents' needs, attitudes, and perceptions about early palliative care integration in pediatric oncology

Feasibility of early palliative care consultation for children with highrisk malignancies

Exposure to homebased pediatric palliative and hospice care and its impact on hospital and emergency care charges at a single institution

A summary of pediatric palliative care team structure and services as reported by centers caring for children with cancer

Referral practices of pediatric oncologists to specialized palliative care

Impact of a pediatric palliative care program

The impact of palliative care consultation on symptom assessment, communication needs, and palliative interventions in pediatric patients with cancer

Pediatric oncology providers' perceptions of barriers and facilitators to early integration of pediatric palliative care

Palliative care in the community for children with cancer in South East England

Pediatric oncology providers' perceptions of a palliative care service: the influence of emotional esteem and emotional labor

Utilization of palliative care consultations in pediatric oncology phase I clinical trials

The conceptual understanding of pediatric palliative care: a Swiss healthcare perspective

Barriers to palliative care in pediatric oncology in Switzerland: a focus group study

Perceptions of barriers and facilitators to early integration of pediatric palliative care: a national survey of pediatric oncology providers

The codesign of an interdisciplinary team-based intervention regarding initiating palliative care in pediatric oncology

Seven types of uncertainty when clinicians care for pediatric patients with advanced cancer

Predictors of location of death for children with cancer enrolled on a palliative care service

Impact of race and ethnicity on end-of-life experiences for children with cancer

Location of death and endof-life characteristics of young adults with cancer treated at a pediatric hospital

High-intensity end-of-life care among children, adolescents, and young adults with cancer who die in the hospital: a population-based study from the French national hospital database. Cancer. Epub ahead of print 26

Implications of pediatric palliative consultation for intensive care unit stay

Knowledge, beliefs, and behaviors related to palliative care delivery among pediatric oncology health care providers

Specialized pediatric palliative care services for children dying from cancer: a repeated cohort study on the developments of symptom management and quality of care over a 10-year period

Haematological malignancy: are patients appropriately referred for specialist palliative and hospice care? A systematic review and metaanalysis of published data

Palliative care specialists' perceptions concerning referral of haematology patients to their services: findings from a qualitative study

Communication preferences of pediatric cancer patients: talking about prognosis and their future life

Children with cancer share their views: tell the truth but leave room for hope

Supportive and palliative care needs of families of children who die from cancer: an Australian study

Parental experiences with a paediatric palliative care team: a qualitative study

Self-reported fatigue in children with advanced cancer: results of the PediQUEST study

Who is responsible for delivering palliative care to children with cancer?

Training pediatric fellows in palliative care: a pilot comparison of simulation training and didactic education

National impact of the EPEC-Pediatrics enhanced train-the-trainer model for delivering education on pediatric palliative care

Improving the care of children with advanced cancer by using an electronic patient-reported feedback intervention: results from the PediQUEST randomized controlled trial

Communication in pediatric oncology: state of the field and research agenda

Teenage and young adult cancer top 10

Indicators used to assess the impact of specialized pediatric palliative care: a scoping review

Health-related quality-of-life outcome measures in paediatric palliative care: a systematic review of psychometric properties and feasibility of use

Symptoms and concerns among children and young people with life-limiting and life-threatening conditions: a systematic review highlighting meaningful health outcomes

How do professionals assess the quality of life of children with advanced cancer receiving palliative care, and what are their recommendations for improvement?

Core Outcome Set-STAndards for Development: The COS-STAD recommendations

The authors would like to acknowledge the input of members of the Martin House Research Centre Family Advisory Board, who helped with interpretation of the review findings. 

Data are reported fully in the included tables and figures and are derived from published sources only.

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

This review contains data derived from published sources only.

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This article presents independent research partly funded by the Wellcome Trust 

https://orcid.org/0000-0001-5898-0900Alison Booth https://orcid.org/0000-0001-7138-6295

Supplemental material for this article is available online.