Impaired mobility associated with an increased likelihood of death in children: A systematic review

Abstract

Improved identification of children with an increased likelihood of death can support appropriate provision of integrated palliative care. This systematic review aims to consider immobility and the associated likelihood of death in children with disabilities, living in high-income countries. Two reviewers independently searched MEDLINE, Embase, Cochrane Library, OpenGrey and Science Citation Index (1990–2016) for studies that reported hazard ratios (HRs) and relative risk for the likelihood of death related to impaired mobility. Nine papers were included. Three studies reported functioning using the Gross Motor Function Classification Scale (GMFCS) and the remaining studies reported measures of functioning unique to the study. The strongest single prognostic factor for the likelihood of death was ‘lack of sitting ability at 24 months’, HR 44.4 (confidence interval (CI) 6.1–320.8) followed by GMFCS V HR 16.3 (CI 5.6–47.2) and 11.4 (CI 3.76–35.57) and ‘not able to cruise by 24 months’, HR 14.4 (CI 3.5–59.2). Immobility is associated with an increased risk of dying over study periods, but different referent groups make clinical interpretation challenging. Overall, the quality of evidence is moderate. The findings suggest that immobility can support identification of children who may benefit from integrated palliative care.

Keywords

Child health disability functioning GMFCS immobility palliative care systematic review

It is estimated in the World Health Report on Disability (2011) that in 2004 around 13 million children, under the age of 14 years, were living with severe disability (World Health Organization, 2011). The term ‘disability’ is now viewed as an umbrella term for impairments of body functioning (World Health Organization, 2012). Due to either increased survival, changes in disability definitions, or survey procedures, the prevalence of disability in high-income countries is increasing (Friebert and Williams, 2015; Halfon et al., 2012).

For some, there is a significant burden of illness (Read et al., 2012) and uncertainty, since health may be unpredictable and unstable (Horridge, 2015). Variations in definitions and prognostic criteria can hinder the identification of children who are considered life limited (Friebert and Williams, 2015). In the United Kingdom, children with ‘irreversible but non-progressive conditions causing severe disability, leading to susceptibility to health complications and likelihood of premature death’ may be considered to be life limited (Together for Short Lives, 2013), and further delineation is available in a directory of life-limiting conditions (Hain et al., 2013). However, while the prevalence (or identification) of children with life-limiting conditions is also increasing (Fraser et al., 2012), for children with severe impairments of body functioning (World Health Organization, 2012), the focus is often on enhancing participation and improving the quality of life (Rosenbaum, 2015). Consideration of the likelihood of death is sometimes seen as incongruent with this focus, and may be given little consideration. While severe impairments are associated with death in childhood (Wolfe et al., 2014), prognostic uncertainty remains, and suitability for palliative care is, therefore, uncertain.

Children with recognized life-limiting conditions, and their families, can benefit from palliative care, as an active approach to holistic care from diagnosis, regardless of whether treatment interventions are available. This approach involves giving support to the family as well as the child and can be implemented even when resources are limited, such as by the team that knows the child best, even when specialist palliative care teams are not available (World Health Organization, 2015). However, children with severe impairments are not always recognized as needing palliative care due to the uncertainty of prognosis and the fact that many survive into adulthood. Unfortunately, this puts some at risk of not receiving appropriate and timely palliative care, which, for many chronic health conditions, can be of long-term benefit (Siouta et al., 2016).

Palliative care in its different guises ranges from a philosophy of care to specialist services (Hain and Wallace, 2008) but is recognized under the United Nations human right to health. Palliative care for children with chronic health conditions, and their families, can help meet the increasing health needs of cumulative disabilities and decreasing quality of life (Dallara et al., 2014; Graham and Robinson, 2005; Viallard, 2014). However, this needs to be introduced early for maximizing benefit such as improved symptom management, support for decision–making and reduced interventions at the end of life (Hauer and Wolfe, 2014; Walter et al., 2013). For bereaved parents, lack of support at the end of life places them at increased risk of long-term psychological and physical morbidity (Rosenberg et al., 2012).

While there is variability in local availability of palliative care services, and lack of consensus in the definition of integrated palliative care, it is generally accepted that palliative care is not a stand-alone intervention. An integrated approach, aimed at improving the quality and quantity of life, is known to increase the quality of care and support healthcare professionals (Ewert et al., 2016). Even in the absence of focused palliative care services, advance care planning for this population is an important means of managing uncertainty and driving up standards in children’s palliative and end-of-life care (Brook and Hain, 2008). This dual planning concept involves making multiple plans for care and using the one that is most appropriate for the circumstances at the time (Villanueva et al., 2016).

In trying to identify children with life-limiting conditions, research has usually focused on either medical diagnoses (Hain et al., 2013) or impairments of body functioning (Strauss et al., 2000); however, potentially both features could be used together to improve the estimation of prognosis. For example, while spastic quadriplegic cerebral palsy (CP) (international classification of diseases [ICD]10 G80.0) and mixed CP syndromes (ICD10 G80.0) are identified as life limiting (Hain et al., 2013), not all children with these diagnoses will die in childhood, and consideration of functional status (activity, performance and participation) is important (Bergstraesser et al., 2013).

Studies have reported the increased probability of death to be associated with immobility (Ashwal, 2005; Cohen et al., 2008), lack of feeding skills, genetic aetiology, hearing deficit (Cohen et al., 2008), intercurrent illnesses with three or more coexisting disabilities (Decouflé and Autry, 2002), functional disability (Hutton et al., 1994), inability to recognize voices or speak intelligible words and incontinence (Katz, 2003). More recently, the presence of ventilatory airway support, pain/distress associated with feeding and difficulty maintaining sitting position have been proposed as ‘vulnerability factors’ indicative of warranting palliative care (Harrop and Brombley, 2012) but the methods have not been reported in any detail. The focus of this review is mobility, a measurable aspect of body functioning allowing comparison. The International Classification of Functioning, Disability and Health (World Health Organization, 2012) considers mobility to include a range of movements such as changing or maintaining a body position or location, carrying, moving or handling objects, walking or moving and using assorted forms of transportation. There may be specific impairments of functioning, routinely measured, within these diagnostic categories that may be useful in identifying children at an increased likelihood of death. Attention to both diagnosis and impairments to functioning may therefore help healthcare professionals further define the population with an increased probability of death and facilitate earlier interventions.

The impetus behind prognosis research in children with impairments includes planning for future care needs, health insurance (Christakis and Iwashyna, 1998; Katz, 2003; Strauss et al., 2000) and clinical decision-making (Hayden et al., 2013). More recently, attempts have been made to further identify children with an increased risk of death for the provision of palliative care services (Hain et al., 2013; Harrop and Edwards, 2013). Since access to palliative care is limited by an uncertain prognosis, children with severe impairments may not have their needs fully met.

Rationale

Knowledge of specific impairments of body functioning associated with an increased likelihood of death in children can improve the identification of those who may benefit from integrated palliative care.

Objectives

The primary objective of this review was to identify the likelihood of death associated with impaired mobility in children with severe impairments, living in high-income countries. This was done through a systematic review of studies reporting a point estimate of hazard ratio (HR) and relative risk (RR) for mortality related to impaired mobility.

Method

Studies were retrieved using MEDLINE, Embase, Cochrane Library, OpenGrey and Science Citation Index from 1990 to 2016 and reference lists. Relevant studies were located by combining search terms which included child, paediatric, disabilities, mortality, death and survival. For MEDLINE search strategy, see Online Supplementary information. The last search was performed on 01/12/16. The reference lists of included studies, and the grey literature, were explored to locate potentially relevant studies for inclusion. No published protocol of the present review exists.

All retrieved papers were reviewed to include studies using the following criteria: (1) observational cohort study with >50% follow-up after 1980, published after 1990; (2) majority of participants were children (<18 years) for most of the study; (3) children described as having disability/impairments; (4) studies relating to mobility; (5) country of study defined by World Bank as high income; and (6) studies reporting a HR or RR for mortality to allow comparison of data. Studies of cancer, gestation, psychiatry, seizures, trauma, interventions, medicinal products and the neonatal period were excluded as outside the focus of this review, as were studies of prognostic factors unrelated to body functioning such as biomarkers. As the focus of this review is prognostic factors, which are distinct from the actual cause of death, studies dealing with cause of death alone were also excluded.

All data were extracted onto piloted forms (SN) and then checked independently (EP). The point estimates and associated confidence intervals (CIs) of the HR and RR were extracted for synthesis by two authors independently (SN and EP). Two authors (SN and EP) then independently assessed the quality of studies using the Quality in Prognosis Studies (QUIPS) tool (Hayden et al., 2013), which assesses quality in terms of study participation, attrition, prognostic factor and outcome measurement, confounding, statistical analysis and reporting, resulting in a judgement of high, moderate, or low risk of bias. A forest plot was drawn in the R package ‘forest plot’ (Gordon and Lumley, 2017).

Results

Nine studies met the inclusion criteria (Online Supplementary Figure 1). Data were extracted for impairments of body functioning related to mobility; most studies reported multiple factors. Figure 1 presents the findings in a forest plot, with immobility described as per study. This shows an indication of the magnitude of risk associated with each impairment but they are not necessarily directly comparable. The majority of studies reported functioning assessed using non-standardized measures, two studies reported the Gross Motor Function Classification Scale (GMFCS) (Palisano et al., 2008) level V (Touyama et al., 2013; Westbom et al., 2011) and one study reported GMFCS combining level IV and V (Baird et al., 2011).

Figure 1.

Forest plot of HR and RR for mortality, related to specific impairments.

The characteristics of included studies are displayed in Table 1 (also see Supplementary Table S1). The risk of bias of included studies was judged using the QUIPS checklist. Of the nine papers included, four were judged to be at low risk of bias in all sections of QUIPS checklist (see Table 1), and the remaining studies had at least one section rated as moderate risk of bias. The most common aspect causing risk of bias was statistical analysis.

Table 1.

Risk of bias from individual studies, judged using QUIPS.

	Baird et al. (2011)	Cohen et al. (2008)	Evans et al. (1990)	Hemming et al. (2005)	Nielsen et al. (2002)	Strauss et al. (1998a)	Strauss et al. (1998b)	Touyama et al. (2013)	Westbom et al. (2011)
Study participation	Low	Low	Mod^b	Low	Low	Low	Low	Low	Low
Study attrition	Low	Low	Mod^c	Low	Low	Low	Low	Low	Low
Prognostic factor measurement	Low	Low	Low	Low	Low	Low	Mod^d	Low	Low
Outcome measurement	Low	Low	Low	Low	Low	Low	Low	Low	Low
Study confounders	Low	Low	Low	Low	Low	Low	Low	Low	Low
Statistical analysis	Mod^a	Mod^a	Mod^a	Low	Low	Low	Low	Mod^a	Low

QUIPS: Quality in Prognosis Studies; Low: low risk of bias; Mod: moderate risk of bias; High: high risk of bias.

^aEvent rate too low for statistical analysis used.

^bCase definition changed during study period.

^cInsufficient data on those lost to follow-up.

^dReliability not reported, methods to limit bias not reported.

Seven papers were reported as retrospective cohort studies (Baird et al., 2011; Evans et al., 1990; Hemming et al., 2005; Nielsen et al., 2002; Strauss et al., 1998a, 1998b; Touyama et al., 2013), one as case control (Cohen et al., 2008) (with two comparison groups) and one retrospective study of prospectively collected data (initial recruitment for another study, reporting vital status at a later point) (Westbom et al., 2011). This study suggested the data were collected as part of a research project, and the remaining studies used routinely collected clinical data.

Studies used existing data sets, with two using the Client Development Evaluation Report (CDER) (Strauss et al., 1998a, 1998b), two using registers of CP (Nielsen et al., 2002; Westbom et al., 2011), and the remaining five, using clinical records (Baird et al., 2011; Cohen et al., 2008; Evans et al., 1990; Hemming et al., 2005; Touyama et al., 2013). Sample sizes ranged from 277 to 12,719 but were not consistently reported. All but two studies reported children with CP as their population, one reported children as having ‘developmental disabilities’ and one with traumatic brain injury. Two studies originated in the United States, three studies originated in the United Kingdom and singular studies were identified from Japan, Denmark, Sweden and Israel. Eight were published in English and one in Danish (translated by a native speaker). Three studies used the GMFCS (Palisano et al., 2008), although this scale categorizes functioning rather than measuring it. Otherwise, the measurement of functioning was unique to each study (or as in the case of the CDER, a local measure forming part of the data collection). Aspects of impairment, such as potential prognostic factors, are presented as described by the study. In most cases, the referent groups constituted those children who did not have impairment or the degree of impairment differed, leading to a high degree of clinical heterogeneity.

Figure 1 shows that the singular factor associated with the highest ratio was ‘lack of sitting ability at 24 months’ with HR 44.4 (CI 6.1–320.8) (Baird et al., 2011); however, sitting ability is defined as ‘getting to any sitting position from any lying position on the floor and then sitting without propping with either arm for 15 seconds’. This equated to 56 of 200 in the exposed group compared to 1 of 132 children in the referent group. Two studies reported HR for GMFCS level V, with comparable results: 16.3 (CI 5.6–47.2), constituting 29 of 166 children in the exposed group compared to 4 of 413 in the referent group and 11.4 (CI 3.76–35.57) which equated to 25 of 102 children in the exposed group compared to 5 of 605 in the referent group (Touyama et al., 2013; Westbom et al., 2011). One study combined GMFCS level IV and V with a reduced HR of 6.2 (CI 2.8–14) equating to 55 of 176 children in the exposed group compared to 2 of 160 (Baird et al., 2011). In contrast, seemingly less severe impairments such as ‘no functional hand use’ and ‘rolls/sits but cannot walk unaided’ (Strauss et al., 1998a) were associated with lower HRs of less than 3.

All studies reported a range of follow-up times according to date of birth and end point of the study (Table 1; also see Supplementary Table S1), therefore it is not possible to make any meaningful judgements about the effect of the length of follow-up on the hazard.

Discussion

As far as we are aware, this is the first systematic review examining the impact of immobility on the likelihood of death in children with severe impairments. We have extracted HRs, which give an instantaneous risk at any given point over a period of time, and RR, which gives a cumulative risk over a period of time. While only a small number of studies were identified, this review adds to the growing body of literature seeking to identify children most at risk of death, and who might benefit from palliative care. These findings suggest an association with an increased likelihood of death, rather than prediction of death or imminent death, for which more data would be required. However, the findings have implications for practice, policy and research.

The clear majority of studies focused on CP, seen to be a static or non progressive brain injury (National Institute for Health and Care Excellence, 2017). However, children may, in reality, have an unpredictable illness trajectory. The clinical signs and level of disability for CP, for example, may change over time due to the long-term existence of hypertonia, contractures and musculoskeletal deformity and bone density, and life expectancy may be influenced by the presence of scoliosis, seizures, cardiac or respiratory factors (Wimalasundera and Stevenson, 2016). For the static conditions captured within this review, impairments to functioning are an early indicator of risk.

Given the increased likelihood of death in children who are not able to sit or ‘cruise’ at 24 months of age (‘cruising’ refers to walking while holding onto something, defined in the study as ‘along furniture or wall, even if placed in standing, for at least two steps sideways in either direction’) (Baird et al., 2011) and those in the GMFCS level V category, healthcare professionals should consider these as specific high-risk factors for mortality. Nevertheless, all reported impairments were associated with an increased risk.

With regard to the assessment of functioning itself, measurement was diverse and study specific. Gross motor function was assessed using the GMFCS, CDER, or non-standardized measures based on clinical judgement. Notably, GMFCS category V is directly related to immobility, with consideration of head and truck control, but where this was not used, immobility was gauged in ways unique to each study, which renders further statistical analysis problematic. Evans et al. (1990) defined immobility as ‘confined to a wheelchair which they did not propel themselves’, Cohen et al. (2008) defined immobility as ‘no mobility’, and Hemming et al. (2005) defined immobility as ‘unable to walk even with aids, uses a wheelchair or is bed ridden’. Strauss et al. (1998a) use ‘rolls/sits but cannot walk unaided’ and Strauss et al. (1998b) rate immobility by the lowest of five CDER items (rolling and sitting, hand use, arm use, ability to creep and crawl and ambulation). Nielsen et al. (2002) described immobility as ‘ingen gangfunktion’, translated into English as ‘no walking function’. The GMFCS is considered the gold standard for categorizing gross motor function and classifies into five levels. A more detailed measurement of gross motor function, such as the GMFM (CanChild, 2017), a 66-item scale, could potentially give more comprehensive information by measuring changes in functioning but would require more resources. This measure would also define immobility more specifically and includes the ability to roll and head control.

Hand use was categorized as no functional use as opposed to some functional use by Strauss et al. (1998a) and Baird et al. (2011), using a four-point hand manipulation scale. For hand use, validated measures of functioning exist such as the Manual Ability Classification Scale (MACS) (Eliasson et al., 2006). For more detailed assessment of hand and arm use, the quality of upper extremity skills test (QUEST; DeMatteo et al., 1993) and the Melbourne (Randall et al., 2001) are available.

The findings of this review suggest that immobility (however measured or categorized) is a potential prognostic factor. However, it is uncertain if this is specific to the populations studied, which is primarily children with CP. The direct links between immobility and causes of death have not been examined in this review. Causes of death, such as respiratory infections (Rousseau et al., 2015; Sidebotham et al., 2014), may be due to factors such as immune insufficiency, poor swallowing and inability to cough or achieve full lung expansion. Research is needed to determine the links between immobility and causes of death, as immobility is unlikely to be a single prognostic factor but comprises of multiple impairments, making it at best a broad surrogate for these. This review only offers a reminder that children who are immobile are more at risk of death, and that palliative care provision should be considered.

For consideration of the likelihood of risk, accurate functional measurement is imperative, and we would recommend that the minimum level of functional assessment should include GMFCS and MACS with consideration given to the more detailed assessments suggested above. With accuracy of functional measurement, changes in body functioning can also be monitored.

In the context of this review, the overall quality of the body of evidence was defined using the grading of recommendations assessment, development and evaluation (GRADE) definition that is the extent to which we can be confident that these results reflect the association between prognostic factors and death in the underlying population (Huguet et al., 2013). Generalizability of these results to diagnosis other than CP is limited, given all but two studies were conducted in this population. More specifically, according to GRADE (Huguet et al., 2013), the quality of the body of evidence involves a number of aspects. Consideration of within-study risk of bias (methodological quality) was judged to be moderate; the directness of evidence was judged to have no serious limitations as the samples were similar to the population of interest, although not necessarily to the broader population of children requiring palliative care. Although no formal tests of heterogeneity were undertaken, it was clear that there was heterogeneity of patient populations and in some cases between outcomes. Furthermore, many of the CIs are very wide, suggesting imprecise estimates of the population parameter.

However, many of the effect estimates were strong, exceeding the GRADE levels of >2 and >5 for upgrading evidence by one and two levels, respectively (Schünemann et al., 2013). Given the existing limitations, the evidence would have been given an overall judgement of very low or low, had it not been for the high magnitude of effect associated with some results. Therefore, there is moderate-quality evidence for GMFCS level V and immobility for being associated with an increased likelihood of death in children living in high-income countries and low-quality evidence for head control and hand use. The small number of studies is of concern, this was primarily due to the requirement for HR or RR to be reported but we were unable to assess the risk of publication bias, which for prognosis research is a known issue (Huguet et al., 2013; Riley et al., 2013).

It is important to consider these findings alongside other available information, such as diagnosis (Hain et al., 2013). Communication of risk of death can support the difficult news of a potentially life-limiting condition (Bluebond-Langner et al., 2017) which needs to be carefully discussed and considered alongside the individual preferences of the child and family for information (Andrews et al., 2013; Maltoni et al., 2005), with particular acknowledgement of the uncertainty reflected in the wide CIs found in this analysis. However, if these factors are to be used to inform the provision of palliative care, there needs to be some consensus relating to the level of hazard indicative of the need.

Palliative care should be seen not as an end point but as part of a broader health system. For example, palliative care can be combined with rehabilitation approaches to facilitate self-management and self-care (Tiberini and Richardson, 2015), although this model is most commonly found in the adult literature (Kelley et al., 2012; Oldervoll et al., 2011). Palliative care as ‘active total care’ and rehabilitation are not, even for children, mutually exclusive. Integrating palliative care into existing service provision, tailored to the needs of the child and family, acknowledges the inherent uncertainty, while maximizing the potential for health and well-being, alongside the risk that the child may deteriorate and die.

More rigorous research using prospective observational studies in defined populations with validated measures of functioning are needed if we are to advance the identification of children with an increased likelihood of death, improve the identification of life-limiting conditions and thus improve access to palliative care. However, such studies are likely to be costly, difficult to conduct and not able to report their findings for many years. New guidelines for reporting prognosis studies of tumour markers now exist (McShane et al., 2005), but none yet related to functioning. Because of these problems, researchers should collaborate to develop a consensus about how to make better use of existing data, including validated measures of functioning commonly used by healthcare professionals, particularly therapists, in this population of children.

Conclusions

This systematic literature review suggests that GMFCS level V and overall immobility are associated with a significantly increased likelihood of death in children, with existing severe impairments, living in high-income countries. The GMFCS can therefore support clinical judgement about whether integrated palliative care may be appropriate, as it provides a categorization of immobility. In order to confirm this finding, researchers and healthcare professionals collecting routine data should use validated measures of functioning.

Footnotes

Declaration of Conflicting Interests

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

Funding

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 funded by the National Institute for Health Research (NIHR), as part of an MRes. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.

Supplementary material

Supplementary material for this article is available online.

References

Andrews

Schünemann

Oxman

. (2013) GRADE guidelines: 15. Going from evidence to recommendation—determinants of a recommendation’s direction and strength. Journal of Clinical Epidemiology 66: 726–735.

Ashwal

(2005) Recovery of consciousness and life expectancy of children in a vegetative state. Neuropsychological Rehabilitation 15: 190–197.

Baird

Allen

Scrutton

. (2011) Mortality from 1 to 16-18 years in bilateral cerebral palsy. Archives of Disease in Childhood 96: 1077–1081.

Bergstraesser

Hain

Pereira

(2013) The development of an instrument that can identify children with palliative care needs: the Paediatric Palliative Screening Scale (PaPaS Scale): a qualitative study approach. BMC Palliative Care 12: 20.

Bluebond-Langner

Hargrave

Henderson

. (2017) ‘I have to live with the decisions I make’: laying a foundation for decision making for children with life-limiting conditions and life-threatening illnesses. Archives of Disease in Childhood 102: 468–471.

Brook

Hain

(2008) Predicting death in children. Archives of Disease in Childhood 93: 1067–1070.

CanChild (2017) What is the GMFM? https://www.canchild.ca/en/resources/44-gross-motor-function-measure-gmfm (accessed 12 September 2017).

Christakis

Iwashyna

(1998) Attitude and self-reported practice regarding prognostication in a national sample of internists. Archives of Internal Medicine 158: 2389–2395.

Cohen

Asor

Tirosh

(2008) Predictive factors of early mortality in children with developmental disabilities: a case-comparison analysis. Journal of Child Neurology 23: 536–542.

10.

Dallara

Meret

Saroyan

(2014) Mapping the literature: palliative care within adult and child neurology. Journal of Child Neurology 29: 1728–1738.

11.

Decouflé

Autry

(2002) Increased mortality in children and adolescents with developmental disabilities. Paediatric and Perinatal Epidemiology 16: 375–382.

12.

DeMatteo

Law

Russell

. (1993) The reliability and validity of Quality of Upper Extremity Skills Test. Physical and Occupational Therapy in Pediatrics 3: 1–18.

13.

Eliasson

Krumlinde-Sundholm

Rösblad

. (2006) The Manual Ability Classification System (MACS) for children with cerebral palsy: scale development and evidence of validity and reliability. Developmental Medicine and Child Neurology 48: 549–554.

14.

Evans

Alberman

(1990) Cerebral palsy: why we must plan for survival. Archives of Disease in Childhood 65: 1329–1333.

15.

Ewert

Hodiamont

van Wijngaarden

. (2016) Building a taxonomy of integrated palliative care initiatives: results from a focus group. BMJ Supportive & Palliative Care 6: 14–20.

16.

Fraser

Miller

Hain

. (2012) Rising national prevalence of life-limiting conditions in children in England. Pediatrics 129: e923–e929.

17.

Friebert

Williams

(2015) NHPCO’s Facts and Figures: Pediatric Palliative & Hospice Care in America. Alexandria, VA: National Hospice and Palliative Care Organization.

18.

Gordon

Lumley

(2017) Forestplot: Advanced Forest Plot Using ‘grid’ Graphics. R package version 1.7. https://CRAN.R-project.org/package=forestplot.

19.

Graham

Robinson

(2005) Integrating palliative care into chronic care for children with severe neurodevelopmental disabilities. Journal of Developmental and Behavioral Pediatrics 26: 361–365.

20.

Hain

Devins

Hastings

. (2013) Paediatric palliative care: development and pilot study of a ‘Directory’ of life-limiting conditions. BMC Palliative Care 12: 1–5.

21.

Hain

Wallace

(2008) Progress in palliative care for children in the UK. Paediatrics and Child Health 18: 141–146.

22.

Halfon

Houtrow

Larson

. (2012) The changing landscape of disability in childhood. The Future of Children 22: 13–42.

23.

Harrop

Brombley

(2012) Vulnerability factors considered for acceptance of children with cerebral palsy or other static neurological conditions to children’s hospice services. Together for Short Lives. Birmingham.

24.

Harrop

Edwards

(2013) How and when to refer a child for specialist paediatric palliative care. Arch Dis Child 98: 202–208.

25.

Hauer

Wolfe

(2014) Supportive and palliative care of children with metabolic and neurological diseases. Current Opinion in Supportive and Palliative Care 8: 296–302.

26.

Hayden

Van der Windt

Cartwright

. (2013) Assessing bias in studies of prognostic factors. Annals of Internal Medicine 158: 280–286.

27.

Hemming

Hutton

Colver

. (2005) Regional variation in survival of people with cerebral palsy in the United Kingdom. Pediatrics 116: 1383–1390.

28.

Horridge

(2015) Advance Care Planning: practicalities, legalities, complexities and controversies. Archives of Disease in Childhood 100: 380–385.

29.

Huguet

Hayden

Stinson

. (2013) Judging the quality of evidence in reviews of prognostic factor research: adapting the GRADE framework. Systematic Reviews 2: 71.

30.

Hutton

Cooke

Pharoah

(1994) Life expectancy in children with cerebral palsy. BMJ 309(6952): 431–435.

31.

Katz

(2003) Life expectancy for children with cerebral palsy and mental retardation: implications for life care planning. NeuroRehabilitation 18: 261–270.

32.

Kelley

Ettner

Morrison

. (2012) Disability and decline in physical function associated with hospital use at end of life. Journal of General Internal Medicine 27: 794–800.

33.

Maltoni

Caraceni

Brunelli

. (2005) Prognostic factors in advanced cancer patients: evidence-based clinical recommendations—a study by the Steering Committee of the European Association for Palliative Care. Journal of Clinical Oncology 23: 6240–6248.

34.

McShane

Altman

Sauerbrei

. (2005) Reporting recommendations for tumor marker prognostic studies (REMARK). Nature Clinical Practice Urology 2: 416–422.

35.

National Institute for Health and Care Excellence (2017) Cerebral Palsy in Under 25s: Assessment and Management, NICE Guideline (CG 62). London: NICE.

36.

Nielsen

Uldall

Rasmussen

. (2002) Survival of children born with cerebral palsy. Children born 1971-1986. Ugeskrift for Laeger 164: 5640–5643.

37.

Oldervoll

Loge

Lydersen

. (2011) Physical exercise for cancer patients with advanced disease: a randomized controlled trial. Oncologist 16: 1649–1657.

38.

Palisano

Rosenbaum

Bartlett

. (2008) Content validity of the expanded and revised Gross Motor Function Classification System. Developmental Medicine and Child Neurology 50: 744–750.

39.

Randall

Carlin

Chondros

. (2001) Reliability of the Melbourne assessment of unilateral upper limb function. Developmental Medicine and Child Neurology 43: 761–767.

40.

Read

Blackburn

Spencer

(2012) Disabled children and their families: a decade of policy change. Children & Society 26: 223–233.

41.

Riley

Hayden

Steyerberg

. (2013) Prognosis Research Strategy (PROGRESS) 2: prognostic factor research. PLoS Med 10(2): e1001380.

42.

Rosenbaum

(2015) The importance of context: what are our assumptions about childhood disability? Developmental Medicine and Child Neurology 57: 1084–1084.

43.

Rosenberg

Baker

Syrjala

. (2012) Systematic review of psychosocial morbidities among bereaved parents of children with cancer. Pediatric Blood & Cancer 58: 503–512.

44.

Rousseau

Mathieu

Brisse

. (2015) Aetiologies, comorbidities and causes of death in a population of 133 patients with polyhandicaps cared for at specialist rehabilitation centres. Brain Injury 29: 837–842.

45.

Schünemann

Brozek

Guyatt

. (2013) GRADE Handbook; Handbook for grading the quality of evidence and the strength of recommendations using the GRADE approach. Available at: http://gdt.guidelinedevelopment.org/app/handbook/handbook.html (accessed 13 September 17).

46.

Sidebotham

Fraser

Covington

. (2014) Understanding why children die in high-income countries. The Lancet 384: 915–927.

47.

Siouta

Van Beek

van der Eerden

. (2016) Integrated palliative care in Europe: a qualitative systematic literature review of empirically-tested models in cancer and chronic disease. BMC Palliative Care 15: 56.

48.

Strauss

Shavelle

DeVivo

. (2000) An analytic method for longitudinal mortality studies. Journal of Insurance Medicine-New York 32: 217–225.

49.

Strauss

Shavelle

Anderson

(1998a) Life expectancy of children with cerebral palsy. Pediatric Neurology 18: 143–149.

50.

Strauss

Shavelle

Anderson

(1998b) Long-term survival of children and adolescents after traumatic brain injury. Archives of Physical Medicine and Rehabilitation 79: 1095–1100.

51.

Tiberini

Richardson

(2015) Rehabilitative Palliative Care. London: Hospice UK.

52.

Together for Short Lives (2013) Key Information. http://www.togetherforshortlives.org.uk/professionals/childrens_palliative_care_essentials/approach (accessed 13 September 2017).

53.

Touyama

Ochiai

. (2013) Long-term survival of children with cerebral palsy in Okinawa, Japan. Developmental Medicine and Child Neurology 55: 459–463.

54.

Viallard

(2014) Some general considerations of a human-based medicine’s palliative approach to the vulnerability of the multiply disabled child before the end of life. Culture, Medicine and Psychiatry 38: 28–34.

55.

Villanueva

Murphy

Vickers

. (2016) End of life care for infants, children and young people with life limiting conditions: summary of NICE guidance. BMJ 355: i6385.

56.

Walter

DeCamp

Warrier

. (2013) Care of the complex chronically ill child by generalist pediatricians: lessons learned from pediatric palliative care. Hospital Pediatrics 3: 129–138.

57.

Westbom

Bergstrand

Wagner

. (2011) Survival at 19 years of age in a total population of children and young people with cerebral palsy. Developmental Medicine and Child Neurology 53: 808–814.

58.

Wimalasundera

Stevenson

(2016) Cerebral palsy. Practical Neurology 16(3): 184–194.

59.

Wolfe

Macfarlane

Donkin

. (2014) Why Children Die: Death in Infants, Children and Young People in the UK. Royal College of Paediatrics and Child Health and National Children’s Bureau. Royal College of Paediatrics and Child Health: London.

60.

World Health Organization (2011) World Report on Disability. Geneva: World Health Organization.

61.

World Health Organization (2012) International Classification of Functioning, Disability, and Health. World Health Organization.

62.

World Health Organization (2015) WHO Definition of Palliative Care. Geneva: World Health Organization http://www.who.int/cancer/palliative/definition/en/ (accessed 13 September 2017).

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

0.53 MB