End of Life Care for Infants, Children and Young People: a mixed methods evaluation of current practice in England

Unfortunately, around 450 children and teenagers in England with cancer will require end of life care each year. Currently, the provision of this care varies across the country and little is known about how this variation impacts on children and their families. There are growing numbers of specialist palliative (end of life) care services and children’s hospices in the UK, but there is little evidence to tell us how these services should be developed and what their role should be in supporting children and young people at the end of life. There is very little known about the costs of care and how best to use these limited resources to improve care for these children and their families.

There are studies from North America which have shown that many children with cancer receive high intensity treatments (for example intensive care, IV chemotherapy) very close to the end of their life but whilst palliative care specialists are involved in their care they receive less of these treatments and have more choice on place of care. There have been no recent studies of end of life care for children, teenagers and young adults in the UK.

The University of York requires access to NHS England data for the purpose of the following research project: End of Life Care for Infants, Children and Young People: a mixed methods evaluation of current practice in England.

The following is a summary of the aims of the research project provided by or on behalf of the University of York:

In this study the University of York wish to describe the use of high intensity treatments in children who have died from cancer and assess whether this varies according to the model of end of life care that was available in their treating service.

The University of York requests NHS England data for the purposes of a study to use data collected as part of clinical care, from around 4000 children, teenagers and young adults treated in cancer services in England who died between 2012-2020.

Using routinely collected data sources the University of York will assess whether the use of ‘high intensity’ treatments in children, teenagers or young adults who have died from cancer varies depending on the model of End of Life care that their service delivered.

The objectives of this study are:

1. Compile a comprehensive dataset of the hospital use, treatments and death records of all children, teenagers and young adults who have died from cancer in England from 2012-2020.

2. Describe the use of high intensity treatments and assess any change over time in the use of these ‘high intensity’ treatments

3. Assess whether the use of these ‘high intensity’ treatments varied depending on the model of End of Life care used by the treatment centre (using models identified from WS1 of main study).

4. Assess whether the use of these ‘high intensity’ treatments varied depending on any clinical or demographic characteristics of the children or young person used by the treatment centre.

5. Assess the resource use of the different models of End of Life care in children with cancer.

The following NHS England data will be accessed:

• Hospital Episode Statistics

- (NDRS) Admitted Patient Care

- (NDRS) Accident & Emergency

- (NDRS) Outpatients

• (NDRS) Cancer Registration

• (NDRS) Systemic Anti-Cancer Therapy Dataset (SACT)

• (NDRS) National Radiotherapy Dataset (RTDS)

The level of the data will be:

• Identifiable data received to NHS England by The Paediatric Intensive Care Audit Network (PICANet) & The Intensive Care National Audit Research Centre (ICNARC) – necessary solely for data linkage purposes by the National Cancer Registration Service (NCRAS) data production team.

• Pseudonymised data disseminated to the University of York from the NCRAS team at NHS England—for the purposes of data processing to fulfil the study objectives.

The data will be minimised and limited to data for a study cohort identified as per below.

• Patient age is between 0-25 who have died between 01 January 2012 to 31 December 2020 with a diagnosis of cancer in England.

For each individual patient, data will only be provided from the cancer start date [Diagnoses must have occurred from 01/01/1988 to 31/12/2020] and until Date of Death.

The quantum of data requested is the minimum necessary and could not be further reduced without impacting the ability to achieve the stated aims. The justification for the data products requested is that the study team need to acquire data which will provide the necessary information to quantify and assess the granular level of care this population received before death. This includes detail of hospital use, treatments and death records.

The University of York is only requesting data necessary to ensure that the aims of the study are achieved and to allow for comparability with previous scientific evidence. The research team at the University of York will only have access to pseudonymised data. The University of York is requesting data restricted to the minimum number of years required 01 January 2012 to 31 December 2020. Under 25’s receiving end of life care for a diagnosis of cancer is a relatively small sample size, as such a period of 8 years is indicated in order for the study team to capture a large enough cohort and to assess the change in care over time. There are no less intrusive ways to achieving the purpose of this project

The scope of the project and minimisation was peer reviewed by the National Institute of Health Research (NIHR) and subsequently reviewed by The Confidentiality Advisory Group (CAG) and Health Research Authority (HRA).

The University of York is the sponsor, sole data controller and responsible for ensuring that the data will only be processed for the purpose described above. The project is being funded by a grant provided by the NIHR Health Services and Delivery. Although NIHR is the study funder, NIHR will not carry out any data controllership activities.

The lawful basis for processing personal data under the UK GDPR is:

Article 6(1)(e) - processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller;

The University of York is recognised as a public authority per Schedule 1 of the Freedom of Information Act 2000. Power is conferred upon the University of York by the University's Royal Charter "to advance learning and knowledge by teaching and research, and to enable students to obtain the advantages of University education."

The lawful basis for processing special category data under the UK GDPR is:

Article 9(2)(j) - processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject.

In the University's capacity as a public authority, the research undertaken will be in the public interest as the outcomes of this study will quantify the effects of current healthcare practices, identify the impact on those affected and have the potential to improve healthcare by aiding service planning and provision.

The funding is provided by NIHR Health Services and Delivery. The funding is specifically for the project described. Funding is in place until Dec 31st 2024.

The University of York is the sole data processor. The University of York provide IT support for the purposes of data processing.

Co investigators from other organisations are involved and partly responsible for the delivery of the study. These include individuals from the University of Leeds, Banger University, Cardiff University Hospital, Bradford Royal Infirmary and Manchester University Hospital. The co-investigators will not be accessing or processing the NHS England data. The contract and responsibility of delivering the study is the University of York and they are sole data controllers.

12 paediatric oncology and haematology experts (consultants and nurses) were consulted in an advisory capacity to identify the definitions of high intensity care via a zoom event.

Organisations represented as part of the study steering committee include, NHS England, The Paediatric Intensive Care Society, Association for Paediatric Palliative Medicine and Together for Short Lives.

Data processing will be only carried out by substantive employees of the University Of York who have been appropriately trained in data protection and confidentiality.

A parent advisory group was consulted regarding the collection of the data for the purposes described above. 6 parents as part of the parent advisory panel contributed to the design, management and dissemination of the overall study. Additionally, one of the co-investigators on this study is a bereaved parent.

The study design and use of patient identifiable data without consent was discussed with four bereaved parents of children who had cancer, these parents were from the Martin House Research Centre Family Advisory Board (www.york.ac.uk/mhrc) and PORT (Paediatric Oncology Reference Team). PORT is a team of parents in the UK who have direct experience of children's cancer. These parents were all supportive of the use of their child’s data without consent for linkage of these datasets.

“yes it is acceptable, no parent/carer would object a study that would lead to any service improvements. Personally, I would not want to be contacted now to request permission to use my child's data.”

“for me, it comes down to a question of whether the research needs to be done and, if so, what is the least invasive/ least amount of data sharing that can be done in order to do it. So, as someone whose child has died from cancer, I think that this research is important and so, for me, what option would be most acceptable for me and other people in this situation, given that (child’s name) data would be in this. I think that what you have suggested - that the data is identifiable only in order to link the datasets, sits ok with me. I would be happy with that. To me, it seems that you have tried to find a way of sharing as little data as possible with as few people as possible and I think that it is ok to do that without consent, because the research team won't have the identifiable data. If they did, I would want to be asked for consent.”

These parents also reviewed the text for the website and suggested some changes which have been made by the research team. There are existing opt out systems for the individual datasets used in this study. These opt outs will be adhered to for this study. A study specific opt out will be advertised via the University of York and Children Cancer and Leukaemia Group (CCLG) websites for 6 weeks prior to data extraction directing the parents to contact NHS England.

This study involves data linkage with PICANet (Paediatric Intensive Care Audit Network) and ICNARC (Intensive Care National Audit & Research Centre) data to achieve its objectives.

PICANet and ICNARC will transfer identifiable intensive care data of both living and deceased children (no clinical data) to NHS England. The data will consist of identifying details (specifically NHS Number, Name, Date of Birth, Postcode, Gender, unique study ID, PID number) to be linked with NHS England data. This flow of identifiable data has approval from CAG.

NHS England will link the data received from PICANet and ICNARC to the relevant Cohort records. NHS England will then only flow back a Unique Study ID and a PICANet or ICNARC serial number. The data flowing back to PICANet and ICNARC from NHS England will contain no direct identifying data items nor will it contain any NHSE data. PICANet and ICNARC will then send their intensive care data containing/comprised of clinical and demographic data and the unique study ID via secure transfer to the University of York.

NHS England will identify the main study cohort using the inclusion criteria from the HES, SACT, RTDS and Cancer Registry. NHS England will generate a unique study ID. A pseudonymised linked dataset will be transferred via a secure electronic file transfer system (SEFT) to the University of York. The data will contain no direct identifying data items but will contain the Unique Study ID which can be used by the study team at the University of York to link the NCRAS data with the record level intensive care data (PICANet and ICNARC) held by the recipient.

The data will be stored on servers at the University of York Department of Health Sciences. No data is cached locally on end user devices.

No data is backed up off site at other locations.

This data sharing agreement permits the University of York to access NHS England data for 3 years. Any subsequent amendments or extensions will be subject to a new application. At the end of the study the University of York will apply to NHS England to seek approval for the data to be archived for 10 years. Data will be stored in the University of York in accordance with GDPR and the University of York guidelines. At the end of the default retention period (10 years) NHS England will issue a data destruction certificate where all data will be confidentially destroyed by a secure method.

The data will be accessed onsite at the premises of the University of York and by authorised personnel via remote access in line with NHS England's Remote access policy. The data will remain on the servers at the University of York at all times.

The data will not leave England/Wales at any time.

Employees or agents of the University of York are only permitted to access pseudonymised data including information derived from NHS England data. Such datasets will adhere to the relevant suppression rules.

PICANet and ICNARC will only flow identifiers from their intensive care datasets to NHS England and receive only a PID and study ID back from NHS England. PICANet and ICNARC are not permitted to access the pseudonymised record level study cohort derived from NHS England disseminated to the University Of York.

NIHR Health Services and Delivery (funder) and the institutions listed as key influencers are not permitted to access the data.

All personnel accessing the data have been appropriately trained in data protection and confidentiality.

The data provided to the University of York will be combined with data derived from the NHS England datasets provided to PICANet and ICNARC.

All analyses will use the pseudonymised dataset. There will be no requirement and no attempt to reidentify individuals when using the pseudonymised dataset.

Analysts from the University of York will process the data for the purposes described above.

Once data linkage has been undertaken an assessment of data quality and completeness will be conducted for all the key clinical and demographic variables of interest. Data conflicts will be solved using

1) Qualitative assessment based on demographic data - removal of non-reliable/inconsistent data;

2) Count of missing data,

3) Use the most common recorded value (either exact or nearest for dates).

An assessment of missing data will be undertaken once the data are linked and multiple imputation using chained equations will be used where appropriate. If imputed datasets are used, then a sensitivity analysis comparing complete case and imputed analyses will be undertaken.

Derivation of Key variables:

Some of the key demographic variables will be obtained by combining different data sources e.g. ethnic group, deprivation score. In this situation if any conflict between data sources occurs, the study team will assign the most commonly recorded ethnic group (census 2011 categories) assuming that is not ‘unknown’. The study team will use standard small number suppression rules when publishing data (no cell size <10).

Descriptive Statistics:

Appropriate summary statistics, e.g. frequencies and proportions for categorical variables and mean (with standard deviation) or median (with interquartile range) for continuous variables will be produced for all the key variables to describe any variation.

Demographic (age, sex, ethnicity, deprivation score), and treatment information (geographical, hospital/trusts) data will be used for describing sample (objectives 2 and 4) and independent variables in the regression models (objective 3).

Primary outcome: any one of the following high intensity treatments: intravenous chemotherapy < 14 days from death (yes/no); more than one emergency department visit (yes/no); and more than one hospitalization or intensive care unit admission < 30 days from death (yes/no) (21).

Secondary outcomes: mechanical ventilation < 14 days from death, place of death (hospital, home, hospice, other) and any additional data items identified in the stakeholder workshop (see below) and considering the most updated definitions used in the scientific literature. The additional items identified in the stakeholder workshop were: Radiotherapy in last 14/30 days; any ICU admission in the last 30 days of life; >1 hospital admission in the last 30 days of life; HSCT in the last 100 days of life; >1 ED visit in the last 14 days of life; Oral or IV chemotherapy in the last 14 days of life; Mechanical ventilation in last 14 days of life; Hospital death.

Sensitivity analyses are planned to assess potential differences on the definition of the outcomes.

Analyses will evaluate and compare outcomes used in different End of Life care models (identified in WS1 of the study) using appropriate regression models. Each analysis will account for the multiple confounding factors in this population (age, underlying diagnoses, comorbidities, outpatient attendance, socioeconomic status (Index of multiple deprivation) identified using causal inference methods.

The health economic analyses will embed the estimation of the resource use of each package of care into the regression analyses of the retrospective data. The findings of these regressions will be used to inform a full costings analysis by combining with estimates of the unit costs of each resource use element. This analysis will explore the variation in the cost of the End of Life care models through extensive sensitivity and scenario analyses.

The expected outputs of the processing will be:

• Submissions to peer reviewed journals such as Archives of Disease in Childhood, BMJ Supportive & Palliative Care, Palliative Medicine [approx. 12 months after receipt of data].

• Presentations at World Research Congress of European Association for Palliative Care; European Congress on Paediatric Palliative Care; Health Services Research UK Conference.

• Publication of reports on the Martin House Research Centre (MHRC) website (https://www.york.ac.uk/healthsciences/research/public-health/projects/martinhouse).

The outputs will not contain NHS England data and will only contain aggregated information with small numbers suppressed as appropriate in line with the relevant disclosure rules for the dataset(s) from which the information was derived.

All outputs will be aggregated data with small numbers suppressed, in line with the HES analysis guide. Conference and journal outputs hope to be available to clinicians, academics and members of the public.

The outputs will be communicated to relevant recipients through the following dissemination channels:

• Open-source frameworks such as being published on online webpages open to all,

• Email alerts to the key stakeholders,

• Briefing documents provided to the clinical leads in all the paediatric oncology centres, in the UK,

• Social media,

• Zoom-meetings,

• Journals,

• Public events such as scientific conferences.

All the study outputs will be available via the study website and via links from other websites. As well as email alerts to the key stakeholders a copy of the research briefing will be sent to the clinical leads in all the paediatric oncology centres, in the UK.

There may also be communication via emails, social media (@UoYmhrc) and/or virtual Zoom meetings.

The study team have identified the following local, national and international priority audiences for this study:

1. Parents

2. Clinicians

3. Healthcare managers and commissioners

4. Clinical membership bodies

The key influencers are:

1. Royal College of Paediatrics and Child Health

2. Paediatric Intensive Care Society

3. British Association of Perinatal Medicine

4. Association for Paediatric Palliative Medicine

5. Together for Short Lives

The key decision makers for this clinical area are NHS England and NICE (both part of the study steering committee).

The communication channels described in the following section will be used to update progress on the whole study and to disseminate the expected research outputs which will all be available to download from the study website (based on overall project results):

1. Logic model of EoL care for children.

2. Infographic representation of the typology of models of EoL care for children.

3. Research briefing for clinicians, setting out key findings and implications for practice and training+ animation.

4. Recommendations on future routine data collection.

5. Summary for commissioners provided to each ICS/STP.

6. Summary of findings for parents for distribution via parent facing organisations e.g. Together for Short Lives.

7. Estimates of the cost of EoL care in a paediatric population produced in this work will be submitted to the PSSRU Unit Cost of Health and Social Care Volume.

8. The wider clinical and academic audiences will be reached via conference presentations and academic articles.

9. Final report for the HS and DR journal.

10. Minimum of six journal papers (open-access).

The outputs are expected from early 2024. Final outputs will be towards the end 2024 – the peer review is expected to take up to 12 months so that is 3 years in total.

The findings of this research study are expected to contribute to evidence-based decision-making for policy-makers, local decision-makers such as doctors, and patients to inform best practice to improve the care, treatment and experience of health care users relevant to the subject matter of the study.

The services provided to clients are expected to identify improvement opportunities which the client may then exploit by making changes to systems, processes, resources or infrastructure in order to improve patient experience and patient care.

Overall, the results from this study will hopefully yield new knowledge about inequalities in access. These results will feed into the revised NICE guidelines for End of Life (EoL) care for the study population and shape delivery of EoL care in order to utilise finite resources to maximise impact. This will ensure that there can be genuine progress in the ability of researchers, decision makers, the children and their families to contribute to an understanding of how we can ensure the limited funding for EoL care can be used for greatest benefit for the children at the end of their lives.

The study is expected to lead to recommendations that have an impact on the following areas:

1. Impact on children and families

The study will describe the use of high intensity treatments, assess, and changes over time in the use of these ‘high intensity’ treatments. It will also assess whether the use of these ‘high intensity’ treatments varied depending on any clinical or demographic characteristics of the children or young person used by the treatment centre. This knowledge is essential for recognising children and family needs, which has the potential to support healthcare provision. Ultimately, it may help to improve the quality of life for the children and their families, in particular the groups identified as receiving more high intensity treatment at EoL.

2. Impact on healthcare services and palliative care teams

Currently, the provision of this care varies across the country and little is known about how this variation impacts on children and their families. There are growing numbers of specialist palliative (end of life) care services and children’s hospices in the UK, but there is little evidence to tell us how these services should be developed and what their role should be in supporting children and young people at the end of life. By assessing whether the use of ‘high intensity’ treatments varied depending on the model of End of Life care the study will be able to provide better understanding on this topic.

Results will be valuable to inform service provision and planning, as it will provide better understanding of the population who may benefit from the services and the impact of the model of care received in their EoL.

3. Impact on commissioning: service and economic implications for the NHS

The hospital based paediatric palliative care services are NHS funded, with the majority of children's hospice services being provided by the voluntary sector. The data provided from this project hopes to ensure that the future provision of and planning of both NHS and voluntary sector services is based on robust data.

By assessing the resource use of the different models of End of Life care in children with cancer the study will facilitate understanding of what budget is needed to offer different forms of care, and importantly the role of inequality. Also, it will allow us to understand the likely benefits of additional funding in EoL care in terms of patients’ outcomes for the first time, facilitating a clear indication to budget setters.

Yielded Benefits is not a requirement for new applications.