You are here

Article Text

Article menu

Download PDFPDF

Original article
How many children and young people with life-limiting conditions are clinically unstable? A national data linkage study
  1. Stuart Jarvis1,
  2. Roger C Parslow2,
  3. Pat Carragher3,
  4. Bryony Beresford4,
  5. Lorna K Fraser1
  1. 1Department of Health Sciences, University of York, York, UK
  2. 2Division of Epidemiology and Biostatistics, University of Leeds, Leeds, UK
  3. 3Children's Hospice Association Scotland, Edinburgh, UK
  4. 4Social Policy Research Unit, University of York, York, UK
  1. Correspondence to Dr Lorna K Fraser, Department of Health Sciences, University of York, Area 2, Seebohm Rowntree Building, Heslington, York YO10 5DD, UK; lorna.fraser{at}york.ac.uk

Abstract

Objective To determine the clinical stage (stable, unstable, deteriorating or dying) for children and young people (CYP) aged 0–25 years in Scotland with life-limiting conditions (LLCs).

Design National cohort of CYP with LLCs using linked routinely collected healthcare data.

Setting Scotland.

Patients 20 436 CYP identified as having LLCs and resident in Scotland between 1 April 2009 and 31 March 2014.

Main outcome Clinical stage based on emergency inpatient and intensive care unit admissions and date of death.

Results Over 2200 CYP with LLCs in Scotland were unstable, deteriorating or dying in each year. Compared with 1-year-olds to 5-year-olds, children under 1 year of age had the highest risk of instability (OR 6.4, 95% CI 5.7 to 7.1); all older age groups had lower risk. Girls were more likely to be unstable than boys (OR 1.15, 95% CI 1.06 to 1.24). CYP of South Asian (OR 1.61, 95% CI 1.28 to 2.01), Black (OR 1.58, 95% CI 1.04 to 2.41) and Other (OR 1.33, 95% CI 1.02 to 1.74) ethnicity were more likely to experience instability than White CYP. Deprivation was not a significant predictor of instability. Compared with congenital abnormalities, CYP with most other primary diagnoses had a higher risk of instability; only CYP with a primary perinatal diagnosis had significantly lower risk (OR 0.23, 95% CI 0.19 to 0.29).

Conclusions The large number of CYP with LLCs who are unstable, deteriorating or dying may benefit from input from specialist paediatric palliative care. The age group under 1 and CYP of South Asian, Black and Other ethnicities should be priority groups.

  • Palliative Care
  • PICANet
  • Routine data
  • End-of-life care
  • Life-limiting conditions

https://doi.org/10.1136/archdischild-2016-310800

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    What is already known on this topic?

    • National prevalence of children and young people (CYP) with life-limiting conditions (LLCs) is rising in England.

    • CYP with LLCs have complex healthcare needs—often with repeated hospital admissions, particularly at end of life.

    What this study adds?

    • In each year, over 2200 CYP with LLCs in Scotland are unstable, deteriorating or dying.

    • Children under 1 year of age are more likely than older children to be unstable, deteriorating or dying.

    • CYP from South Asian, Black or Other ethnic groups are more likely to be unstable, deteriorating or dying than White children.

    Introduction

    The prevalence of children and young people (CYP) with life-limiting conditionsi (LLCs) is increasing, with numbers in England estimated to be over 40 000.1 ,2 Although many of the individual diagnoses are rare, as a group, CYP with LLCs represent a larger patient population than many other long-term conditions in children (eg, diabetes mellitus).3 CYP with LLCs have unpredictable disease trajectories, but typically have complex healthcare needs—often with repeated hospital admissions, particularly at end of life.4 ,5 During childhood, care is usually coordinated by tertiary paediatric specialists or community paediatricians. Although the WHO definition of children's palliative care states that “It begins when illness is diagnosed, and continues regardless of whether or not a child receives treatment directed at the disease”,6 paediatric palliative care specialists are often not involved until later in the disease process. Identifying CYP who would benefit from specialist input earlier may be beneficial to them and health and social care services.

    Until now, there has been no published research estimating the proportion of CYP with LLCs who were not clinically stable and therefore in greater need of input from specialist paediatric palliative care services. This study used routinely collected national healthcare data to identify CYP with LLCs in Scotland, their clinical stage and risk factors for instability.

    Patients and methods

    Definition of LLCs

    Individuals with LLCs were identified using a previously developed ICD-107 coding framework,2 further refined by removing International Classification of Diseases, 10th Edition (ICD-10) codes where, in the current data set, no individuals had died (information not available when the original framework was developed).

    Data sets used

    Extracts from the following routinely collected data sets were used (see online supplementary figure S1 and table S1):

    • Scottish Birth Record (SBR)

    • Scottish Morbidity Record (SMR01)—General Acute Inpatient and Day Case

    • Prescribing Information System data (available from 2009/2010)

    • Scottish Outpatient Dataset (SMR00)

    • Scottish Cancer Registry (SMR06)

    • General Records Office Death Registration data

    • Paediatric Intensive Care Audit Network (PICANet) (available from 2007/2008).

    Data access was approved by the Privacy Advisory Committee (ref: XRB14010). The Health Research Authority approved linkage of PICANet data with NHS National Services Scotland (NSS) data (PIAG 4-07(c)/2002 Amendment 16 February 2015).

    All data were analysed within the NSS Electronic Data Research and Innovation Service (eDRIS) safe haven.8 The presented results underwent disclosure control.9

    Data management

    Data linkage was undertaken by the eDRIS team. A probabilistic algorithm using surname; first initial (forename and second initial if available); sex; year, month and day of birth and postcode was used to match individuals in each data set to the population spine data set (Community Health Index (CHI) index) which contains personal identifiable data for all individuals in Scotland who have used an NHS Scotland service (including general practitioner registration).10 False-positive and false-negative rates in the matching are kept close to 1%.10 The data sets were then linked deterministically using the CHI number.

    There were multiple sources of demographic information. Individuals were assigned the most commonly recorded gender. The various ethnic groups recorded were collapsed to four main groups: White; South Asian (Indian, Pakistani and Bangladeshi); Black; Other (including mixed ethnicity) and then the most commonly recorded ethnic group (excluding ‘not known’) was assigned to each individual. Date of birth was assigned as the most commonly recorded date. Individuals were assigned an age group (under 1; 1–5; 6–10; 11–15; 16–20; 21–25 years) in each financial year based on age at the start of the first record for that year.

    Five population-weighted deprivation categories, using the Scottish Index of Multiple Deprivation (SIMD) 2009, with 20% of the Scottish population in each category, were assigned by eDRIS using Data Zone (the Scottish Government's preferred small geography area)11 of residence. SIMD is a national area-based measure of relative deprivation which consists of 35 indicators in seven domains.12 Individuals were assigned the first deprivation category recorded each year.

    LLC diagnoses were categorised into 11 groups based on the main ICD-10 chapters: neurology, haematology, oncology, metabolic, respiratory, circulatory, gastrointestinal, genitourinary, perinatal, congenital and ‘other’. Individuals may have more than one LLC diagnosis. A primary diagnosis category was defined as the modal diagnosis category in the first diagnostic field in SMR01 and SBR records for that financial year. If there was no mode, the first record for the financial year was used. If there was no SMR01 or SBR record in a given year, the preceding year's primary diagnostic group was used.

    Dates of death were based on the death registration data. Other sources (SMR01, SMR06 and PICANet) replaced only invalid (more than a day before the last inpatient admission) or missing dates.

    Analyses

    All data analyses were performed using Stata V.12 (StataCorp. Stata Statistical Software: Release 12. College Station, TX: StataCorp LP, 2011).

    Cohort identification

    The study cohort was defined as all individuals with LLCs resident in Scotland and aged 0–25 years between 1 April 2009 and 31 March 2014 (in this period, all data sets were available). To minimise left edge effects (where some CYP diagnosed prior to 2009 but who were still alive and resident in Scotland with LLCs may be missed), LLC was assigned to an individual if any of their SBR or SMR01 records from 1 April 2003 to 31 March 2014 contained a framework ICD-10 code. Each year, individuals were deemed alive and resident in Scotland if they had a record in the SBR, SMR01 or community prescription records. LLC prevalence per 10 000 population was calculated with population at risk determined from census-derived mid-year estimates.13 95% CIs were calculated using a normality assumption.14

    Clinical stage

    Four clinical stages (stable, unstable, deteriorating and dying) in palliative/end-of-life patients have been previously defined.15 ,16 The Palliative Care Funding Review (PCFR) in England of adults and children aimed “to develop a classification system which categorising palliative/end-of-life patients in according to level of need, which is capable of categorising palliative and end of life patients into meaningful groups based on comparable intensity of care needs and similarity in resource use”.15 In the PCFR, the phase of illness (clinical stage) was a subjective measure by the clinician in charge of care but for both adults and children this was the most important factor in determining the intensity of care need and resource use. In this current study, we have reversed this process by defining clinical stage using an objective measure of healthcare use (figure 1):

    • Entering unstable stage: an unplanned admission to hospital lasting >48 hours. More than 48 hours was chosen to avoid including children who had emergency admissions for service availability issues (eg, 17:00 on Friday) rather than level of illness.

    • Entering deteriorating stage: an emergency admission to an intensive care unit or paediatric intesive care unit (PICU).

    • Entering dying stage: the last 28 days before death (an arbitrary time prior to death was required and due to very small numbers of CYP dying, this was the only acceptable classification).

    Individuals were stable when not in one of the other stages, for example, only appearing in the prescribing data or only having planned inpatient hospital stays.

    The most severe clinical stage was determined for each cohort member in each financial year. Sensitivity analyses were undertaken, classifying individuals at 1-month, 3-month and 6-month intervals and varying the transition criteria, for example, entering unstable: any unplanned admission; entering deteriorating: any PICU admission requiring ventilation and entering dying: 14 days prior to death. These results could not be released from the safe haven due to disclosure concerns.

    Figure 1
    Figure 1

    Criteria for transition between stages of condition for children and young people with life-limiting conditions in Scotland. The charts at the bottom show two example patient trajectories, with multiple changes in clinical stage, and the most severe clinical stage recorded in each year. ICU, intensive care unit.

    The yearly percentage of cohort members in each clinical stage was calculated overall, by age group, deprivation category and diagnostic group. Analyses by ethnicity could not be disclosed due to small numbers. In each set of analyses, individuals with data missing for the variable under consideration were excluded.

    Statistical modelling

    Due to small numbers, the non-stable categories were combined to create a single binary outcome variable (stable=0; not stable=1) for modelling. Age group, deprivation category, ethnicity, primary diagnostic category and gender were identified as predictors.

    The data were clustered by individual with possible dependence in stability across multiple years, requiring a multilevel model. A random intercept logistic regression model was used, showing individual rather than population effects. Level 1 corresponded to financial years and level 2 to the individual. Further details are given in the online supplement.

    Univariable analyses

    Univariable random intercept logistic regression models were generated for each of the candidate predictors: age group, deprivation category, ethnicity, primary diagnostic category and gender.

    Multivariable analyses

    For the multivariable model, candidate predictors were added in turn and retained if their OR for risk of instability was significantly (p<0.05) different from 1 or their inclusion improved model fit (a decrease of >2 in Schwarz’s Bayesian Information Criterion (BIC)17 ,18). Individuals were excluded from the model in a given year if, in that year, they had missing data for any of the predictors.

    Results

    Refinement of diagnostic codes indicating LLCs

    Codes F80.3 (acquired aphasia with epilepsy (Landau–Kleffner)), Q74.8 (other specified congenital malformations of limbs) and Q44.5 (other congenital malformations of bile ducts) were removed from the coding framework.

    Missing and conflicting data

    Gender, date of birth and deprivation category were missing or conflicting for less than 1% of individuals. Ethnicity was missing for 4620 individuals (22.6%), with 173 conflicts.

    Fewer than 10 individuals had inconsistent dates of death. Sixty individuals had dates of death discarded due to at least one SBR or SMR01 episode beginning after the last recorded date of death.

    Numbers of CYP with LLCs in Scotland

    There were 20 436 CYP with LLCs alive and resident in Scotland between 1 April 2009 and 31 March 2014. Numbers increased from 12 039 in 2009/2010 to 15 404 in 2013/2014 (table 1).

    View this table:
    Table 1

    Cohort demographics: for demographics that are constant across years, the number of individuals over the whole study period is provided; for demographics that do change across years, such numbers are not meaningful and these are marked ‘N/A’

    Prevalence increased from 75.0 (95% CI 74.3 to 75.7) per 10 000 population in 2009/2010 to 95.7 (95% CI 94.9 to 96.5) per 10 000 population in 2013/2014 (see online supplementary table S2). Prevalence increased for all age groups except under 1-year-olds (195.0 per 10 000 population, 95% CI 189.3 to 200.6 in 2009/2010; 192.1 per 10 000 population, 95% CI 186.3 to 197.8 in 2013/2014) (see online supplementary table S2).

    Clinical stage

    Each year, between 2201 and 2310 cohort members (14–19%) had at least one period of instability (see online supplementary table S3). The number of individuals stable for the whole year increased from 9729 in 2009/2010 (80.8%; 95% CI 80.5 to 81.2%) to 13 203 in 2013/2014 (85.7%; 95% CI 85.4 to 86.0%) (figure 2, see online supplementary table S3). Numbers of individuals in all other stages decreased: 1857 individuals were unstable in 2009/2010 compared with 1783 in 2013/2014; 262 were deteriorating in 2009/2010 and 254 in 2013/2014 and 191 were dying in 2009/2010 and 164 in 2013/2014. There were variations by age, deprivation category and diagnostic group (see online supplementary figures S2–S4) with under 1-year-olds and those in the most deprived categories most likely to experience instability, while those with perinatal conditions were most likely to be stable.

    Figure 2
    Figure 2

    Clinical stage among children and young people in Scotland with life-limiting conditions. Recorded status is the most severe clinical stage in the year.

    Univariable analyses

    Individuals under 1 year of age were most likely to experience instability, 5.84 (95% CI 5.26 to 6.48) times more likely to experience instability than the 1–5-year-old reference group. Children aged 6–10 years were less likely to experience instability (OR 0.60; 95% CI 0.54 to 0.67), while those aged 11-years old to 25-years old were more likely to experience instability (table 2).

    View this table:
    Table 2

    Unadjusted ORs for risk of instability from univariable random intercept logistic regression models of instability in each year

    Females were 1.12 (95% CI 1.03 to 1.21) times more likely to experience instability than males. Individuals of South Asian (1.69, 95% CI 1.34 to 2.12 times), Black (2.01, 95% CI 1.31 to 3.07) or Other ethnicity (1.57, 95% CI 1.19 to 2.06 times) were more likely to experience instability than those of White ethnicity.

    There were differences in risk of instability by deprivation category, with the least deprived group less likely (0.81, 95% CI 0.72 to 0.91 times) to experience instability than the most deprived group.

    There were also differences by primary diagnostic category. With congenital conditions as the reference category, all other primary diagnostic groups were associated with a higher risk of instability except perinatal conditions (OR 0.27, 95% CI 0.23 to 0.32) and circulatory conditions (OR 0.64, 95% CI 0.53 to 0.78).

    Multivariable analyses

    Age category, gender, ethnicity, deprivation category and primary diagnostic category were included in the final model (table 3).

    View this table:
    Table 3

    ORs for risk of instability from multivariable random intercept logistic regression models of instability in each year

    Age was a significant predictor of instability. With 1–5 years as the reference category, under 1-year-olds had the highest risk of instability (OR 6.4, 95% CI 5.7 to 7.1). Older age groups had lower risk: 6–10 years OR 0.54 (95% CI 0.49 to 0.60); 11–15 years OR 0.73 (95% CI 0.65 to 0.82); 16–20 years OR 0.80 (95% CI 0.71 to 0.90) and 21–25 years OR 0.66 (95% CI 0.59 to 0.75).

    Instability was significantly more likely for girls than for boys (OR 1.15, 95% CI 1.06 to 1.24). Instability was more likely for CYP of South Asian (OR 1.61, 95% CI 1.28 to 2.01), Black (OR 1.58, 95% CI 1.04 to 2.41) and Other (OR 1.33, 95% CI 1.02 to 1.74) than White CYP.

    Deprivation category was not a significant predictor of instability, but inclusion improved model fit (as indicated by BIC).

    Primary diagnostic category was a significant predictor of instability. With congenital abnormalities as the reference category, most other primary diagnoses indicated a higher risk of instability. Only CYP with a perinatal primary diagnosis had significantly lower risk (OR 0.23, 95% CI 0.19 to 0.29). Primary circulatory diagnoses did not indicate a significantly different risk to the reference group (OR 0.89, 95% CI 0.72 to 1.09).

    As many ethnicity data were missing, a sensitivity analysis was carried out excluding ethnicity from the model (see online supplementary table S4). Only circulatory primary diagnoses had an OR for instability under this model that was significantly (p<0.05) different from the OR under the model including ethnicity. ORs for other predictors did not differ significantly between the two models, although individuals in the most deprived category were significantly more likely than those in the least deprived category to experience instability according to the model excluding ethnicity—the two categories were not significantly different in the model including ethnicity.

    Discussion

    This study identified increasing numbers of CYP with LLCs living in Scotland, with over 2200 experiencing at least one episode of instability each year. These individuals may benefit from input from specialist paediatric palliative care services. Children's Hospice Association Scotland, the largest provider of children's palliative care in Scotland, received approximately 115 new referrals each year and currently looks after 380 children and families (personal communication). Therefore, the potential demand for palliative care in the age group 0–25-years in Scotland outstrips the current provision.

    The prevalence of CYP with LLCs was higher in this study than previous estimates in Scotland and England,1 ,2 due to rising prevalence over time and the ability to include individuals alive in Scotland with LLCs who did not have a hospital admission in a single year. The prevalence increased in all age groups except under 1-year-olds suggesting, in common with earlier studies,1 ,2 ,19 that increased prevalence is due to improved survival times rather than increased incidence.

    This is the first study to assign clinical stage based on routinely collected clinical data. Previous approaches have used subjective clinical assessments of care needs, often linked to resource usage.16 ,20–24 This limits assessments to individuals who come into contact with clinicians and may underestimate the stable fraction. Such estimates can vary widely by location, possibly due to differences between individual clinicians' assessments.20 While the transition criteria used in this study were arbitrary, their application to the data was objective and consistent. All CYP identified with LLCs were included, whether or not they had any known contact with clinicians in a given year.

    The results from the multivariate analyses highlighted groups with higher risk of instability who may therefore benefit from targeted input from paediatric palliative care specialists (eg, the under 1 age group, more than 6 times more likely than any other to experience instability). This input should be a combination of direct provisioning of palliative care and training and education of healthcare professionals.

    CYP in minority ethnic groups are more likely to experience instability than the White population. This may be due to different diagnoses within the same broad diagnostic categories used in the model or differences in health-seeking behaviours. There are known differences between ethnic groups in the prevalence of LLC and critical illness,1 ,2 ,25 diagnoses,26 ,27 general access to medical healthcare and likelihood of hospital admission28–30 and access to specialist palliative care services.31 ,32

    Deprivation category was not a significant predictor of risk of instability. It has been previously found that socioeconomic status did not significantly influence access to healthcare within Britain.29 ,33

    Other than age, the strongest predictors of instability were the diagnostic groups. The low risk of instability associated with perinatal diagnoses may reflect inclusion of individuals who had life-threatening diagnoses/events around the time of birth or in the neonatal period but survived those and are no longer life limited.

    Given that the clinical stage categories are based on clearly defined patterns of healthcare usage (eg, emergency hospital admission or PICU admission), they are useful in identifying clinical areas where integration of children's palliative care would be beneficial.

    Strengths and limitations

    This study used high-quality, linked, national healthcare data. The clinical stage definitions were arbitrary but based on clinical knowledge and their application is reproducible and objective. The definitions share some similarity with those previously defined,16 particularly in emergency treatment defining the transition from stable to unstable phases, but differ in defining the deteriorating phase using PICU admission and the dying phase using a fixed period before death.

    Cohort identification only required a life-limiting or life-threatening diagnosis to be recorded once within the hospital data sets. This may result in including individuals who have had a life-threatening event but would no longer be considered life limited.

    Cohort effects are evident and the lack of availability of the community prescribing data prior to 2009 may have accounted for some of the rise in prevalence and the rise in the number of stable CYP with LLCs (CYP only present in the prescribing data are, by the definitions used, stable).

    Although this is a national study, small numbers in some of the categories restricted the analyses (eg, clinical stage by ethnic group or on a quarterly rather than yearly basis). It is possible that the high number of missing ethnicity data influences the presented association between ethnic group and instability, but the sensitivity analysis suggests there is no significant effect on other predictors. As ethnicity recording is improving over time, future studies should be undertaken to validate the present findings.

    Conclusion

    Each year, over 2200 CYP with LLCs in Scotland are unstable, deteriorating or dying and therefore would benefit from input from specialist paediatric palliative care services.

    The under 1 age group and CYP of South Asian, Black or Other ethnicities are most likely to be unstable, deteriorating or dying and would benefit from targeted input from paediatric palliative care specialists.

    Acknowledgments

    We are grateful for the input of the Children in Scotland with Palliative Care Needs (ChiSP) Project Steering Group. We acknowledge the support from The Farr Institute @ Scotland. The Farr Institute @ Scotland is supported by a 10-funder consortium: Arthritis Research UK, the British Heart Foundation, Cancer Research UK, the Economic and Social Research Council, the Engineering and Physical Sciences Research Council, the Medical Research Council, the National Institute of Health Research, the National Institute for Social Care and Health Research (Welsh Assembly Government), the Chief Scientist Office (Scottish Government Health Directorates) and the Wellcome Trust (MRC Grant No: MR/K007017/1). The PICANet Audit is commissioned by the Healthcare Quality Improvement Partnership (HQIP) as part of the National Clinical Audit Programme (NCA). HQIP is led by a consortium of the Academy of Medical Royal Colleges, the Royal College of Nursing and National Voices. Its aim is to promote quality improvement, and in particular to increase the impact that clinical audit has on healthcare quality in England and Wales. HQIP holds the contract to manage and develop the NCA Programme, comprising more than 30 clinical audits that cover care provided to people with a wide range of medical, surgical and mental health conditions. The PICANet Audit is funded by NHS England, NHS Wales, NHS Lothian/National Service Division NHS Scotland, the Royal Belfast Hospital for Sick Children, The National Office of Clinical Audit (NOCA), Republic of Ireland and HCA Healthcare.

    References

    1. ↵
      1. Fraser LK,
      2. Miller M,
      3. Aldridge J, et al
      . Life-limiting and life-threatening conditions in children and young people in the United Kingdom; national and regional prevalence in relation to socioeconomic status and ethnicity. Final Report for Children's Hospice UK. 2011.
    2. ↵
      1. Fraser LK,
      2. Miller M,
      3. Hain R, et al
      . Rising national prevalence of life-limiting conditions in children in England. Pediatrics 2012;129:E923–9. doi:10.1542/peds.2011-2846
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      National Paediatrics Diabetes Audit. National Paediatrics Diabetes Audit Report 2012–2013. London, 2014.
    4. ↵
      1. Feudtner C,
      2. DiGiuseppe DL,
      3. Neff JM
      . Hospital care for children and young adults in the last year of life: a population-based study. BMC Med 2003;1:3. doi:10.1186/1741-7015-1-3
      OpenUrlCrossRefPubMed
    5. ↵
      1. Cohen E,
      2. Berry JG,
      3. Camacho X, et al
      . Patterns and costs of health care use of children with medical complexity. Pediatrics 2012;130:e1463–70. doi:10.1542/peds.2012-0175
      OpenUrlAbstract/FREE Full Text
    6. ↵
      World Health Organization. WHO definition of palliative care. Secondary WHO definition of palliative care, 2011. http://www.who.int/cancer/palliative/definition/en/
    7. ↵
      World Health Organisation. International statistical classification of diseases and related health problems. 10th edn. Geneva, Switzerland: World Health Organisation, 1992.
    8. ↵
      ISD Scotland. Electronic Data Research and Innovation Service (eDRIS). Secondary Electronic Data Research and Innovation Service (eDRIS). http://www.isdscotland.org/Products-and-Services/EDRIS/
    9. ↵
      ISD Scotland. ISD statistical disclosure control protocol. NHS National Services Scotland, 2009.
    10. ↵
      1. Kendrick S,
      2. Clarke J
      . The Scottish Record Linkage System. Health Bull (Edinb) 1993;51:72–9.
      OpenUrlPubMed
    11. ↵
      ISD Scotland. GPD Support: Geography. http://www.isdscotland.org/Products-and-Services/GPD-Support/Geography/ (accessed 6 July 2016).
    12. ↵
      Scottish Government. Scottish Index of Multiple Deprivation. http://www.gov.scot/Topics/Statistics/SIMD (accessed 6 July 2016).
    13. ↵
      National Records of Scotland. Mid-year population estimates: Scotland and its pre-April 2014 NHS Board areas by single year of age and sex: 1981 to 2013. National Records of Scotland, 2014.
    14. ↵
      1. Bland JM
      . Standard error and confidence interval for a proportion. An introduction to medical statistics. 4th edn. Oxford: Oxford University Press, 2015:105–6.
    15. ↵
      1. Hughes-Hallet T,
      2. Craft A,
      3. Davies C, et al
      . Palliative care funding review: funding the right care and support for everyone. London: Department of Health, 2011.
    16. ↵
      1. Clapham S,
      2. Holloway A
      . Palliative care outcomes collaboration clinical manual. Australian Health Services Research Institute, 2014.
    17. ↵
      1. Schwarz G
      . Estimating the dimension of a model. Ann Stat 1978;6:461–4. doi:10.1214/aos/1176344136
      OpenUrlCrossRefWeb of Science
    18. ↵
      1. Kass RE,
      2. Raftery AE
      . Bayes factors. J Am Stat Assoc 1995;90:773–95. doi:10.1080/01621459.1995.10476572
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Norman P,
      2. Fraser L
      . Prevalence of life-limiting conditions in children and young people in England: time trends by area type. Health Place 2014;26:171–9. doi:10.1016/j.healthplace.2014.01.002
      OpenUrlCrossRefPubMed
    20. ↵
      1. Currow DC,
      2. Eagar K,
      3. Aoun S, et al
      . Is it feasible and desirable to collect voluntarily quality and outcome data nationally in palliative oncology care? J Clin Oncol 2008;26:3853–9. doi:10.1200/JCO.2008.16.5761
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Currow DC,
      2. Allingham S,
      3. Yates P, et al
      . Improving national hospice/palliative care service symptom outcomes systematically through point-of-care data collection, structured feedback and benchmarking. Support Care Cancer 2015;23:307–15. doi:10.1007/s00520-014-2351-8
      OpenUrlCrossRefPubMed
    22. ↵
      1. Eagar K,
      2. Gordon R,
      3. Green J, et al
      . An Australian casemix classification for palliative care: lessons and policy implications of a national study. Palliat Med 2004;18:227–33. doi:10.1191/0269216304pm876oa
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Eagar K,
      2. Green J,
      3. Gordon R
      . An Australian casemix classification for palliative care: technical development and results. Palliat Med 2004;18:217–26. doi:10.1191/0269216304pm875oa
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Dalgaard KM,
      2. Thorsell G,
      3. Delmar C
      . Identifying transitions in terminal illness trajectories: a critical factor in hospital-based palliative care. Int J Palliat Nurs 2010;16:87–92. doi:10.12968/ijpn.2010.16.2.46754
      OpenUrlCrossRefPubMed
    25. ↵
      1. Parslow RC,
      2. Tasker RC,
      3. Draper ES, et al
      . Epidemiology of critically ill children in England and Wales: incidence, mortality, deprivation and ethnicity. Arch Dis Child 2009;94:210–15. doi:10.1136/adc.2007.134403
      OpenUrlAbstract/FREE Full Text
    26. ↵
      1. Corry PC
      . Consanguinity and prevalence patterns of inherited disease in the UK Pakistani community. Hum Hered 2014;77:207–16. doi:10.1159/000362598
      OpenUrlCrossRefPubMed
    27. ↵
      1. Sheridan E,
      2. Wright J,
      3. Small N, et al
      . Risk factors for congenital anomaly in a multiethnic birth cohort: an analysis of the Born in Bradford study. Lancet 2013;382:1350–9. doi:10.1016/S0140-6736(13)61132-0
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Krakauer EL,
      2. Crenner C,
      3. Fox K
      . Barriers to optimum end-of-life care for minority patients. J Am Geriatr Soc 2002;50:182–90. doi:10.1046/j.1532-5415.2002.50027.x
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Saxena S,
      2. Eliahoo J,
      3. Majeed A
      . Socioeconomic and ethnic group differences in self reported health status and use of health services by children and young people in England: cross sectional study. BMJ 2002;325:520. doi:10.1136/bmj.325.7363.520
      OpenUrlAbstract/FREE Full Text
    30. ↵
      1. Netuveli G,
      2. Hurwitz B,
      3. Levy M, et al
      . Ethnic variations in UK asthma frequency, morbidity, and health-service use: a systematic review and meta-analysis. Lancet 2005;365:312–17. doi:10.1016/S0140-6736(05)70195-1
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Koffman J,
      2. Burke G,
      3. Dias A, et al
      . Demographic factors and awareness of palliative care and related services. Palliat Med 2007;21:145–53. doi:10.1177/0269216306074639
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Karim K,
      2. Bailey M,
      3. Tunna K
      . Nonwhite ethnicity and the provision of specialist palliative care services: factors affecting doctors’ referral patterns. Palliat Med 2000;14:471–8. doi:10.1191/026921600701536390
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Cooper H,
      2. Smaje C,
      3. Arber S
      . Use of health services by children and young people according to ethnicity and social class: secondary analysis of a national survey. BMJ 1998;317:1047–51. doi:10.1136/bmj.317.7165.1047
      OpenUrlAbstract/FREE Full Text

    Footnotes

    • Twitter Follow Stuart Jarvis at @swjarvis and Lorna Fraser at @lornafraser10

    • Contributors SJ contributed to the design of the study, carried out the analysis, contributed to the interpretation of data and wrote the first draft of the manuscript. RCP made substantial contributions to the conception/design of the work, to interpretation and to revision of the manuscript. PC, BB made substantial contributions to the conception/design of the work, interpretation of the work and to revision of the manuscript. LKF conceived and designed the work and took part in the analysis and played a major role in interpretation of data. She also contributed to drafting and revision of the manuscript. All authors have approved the manuscript prior to submissions and agree to be accountable for all aspects of the work.

    • Funding This study forms part of the ChiSP project, funded by the Managed Service Network for Children and Young People with Cancer (MSNCYPC) through Children's Hospice Association Scotland.

    • Competing interests PC is employed by the funder, Children's Hospice Association Scotland.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement No data are available—the data are potentially sensitive and were accessed within the NSS eDRIS safe haven with only aggregated results disclosed after undergoing disclosure control.

    • ↵i Life-limiting conditions (LLCs) are those for which there is no reasonable hope of cure and from which children or young people will ultimately die prematurely, for example, Duchenne muscular dystrophy or neurodegenerative disease. Life-threatening conditions (LTCs) are those for which curative treatment may be feasible but can fail, for example, cancer. LLC will be used throughout this paper to include LLCs and LTCs.