BRIGHTLIGHT: Do specialist cancer services for teenagers and young adults (TYA) add value?

NHS Digital previously supplied Hospital Episode Statistics (HES) data for the purpose of the BRIGHTLIGHT teenage and young adult (TYA) cancer cohort study.

The Teenage and Young Adult Cancer Specialism Scale (TYA CSS) has been developed from the HES data. The preliminary analysis of the data proved to be more complex than had been anticipated so the team need to conduct additional analysis in order to explain this complexity. As place of care is central to the aim of the study (do specialist services for teenagers and young adults add value), retaining HES data to inform this analysis is critical.

UCLH require HES data to undertake analysis to refine the Cancer Specialism Scale. The research team previously developed this based on the proportion of admitted care episodes in ‘specialist’ TYA principal treatment centres (PTC). Participants were therefore defined as receiving all, some or no specialist care. However, not all TYA PTCs were equal so additional work has been undertaken based on the qualitative aspects of the BRIGHTLIGHT grant to define what specialist or age-appropriate care is. The plan will then be to redefine hospitals as specialist based on this definition and recalculate the proportion of care based on this. This will also extend this to include outpatient as well as admitted patient care.

UCLH have already received HES admitted patient care, outpatient and A&E data for 2012 to 2016, which has been used in the previous analysis. The original expectation was that the data would be required until December 2018 but it is now required until 2025 to enable the research team to get the resource to be able to redefine the term specialist and re-do all the analysis with this re-categorised data.

Young people often receive care in multiple hospitals and have many admissions, so the aim was to discover their access to ‘specialist’ care. Getting this through self-report would have resulted in under/erroneous reporting. Similarly, it is not possible to get this information accurately from their clinical teams because the study was not open to recruitment in every hospital in England (only 107). This was the only accurate way of determining this information.

Recruitment into the cohort was problematic so it was extended an additional 12 months from December 2013 to December 2014. As a longitudinal study, this had a knock on effect so data collection was not complete until March 2018. When analysis began in May 2018 it became evident that the study needed to account for severity of illness so the team needed to conduct an additional study to develop a bespoke severity of illness scale (young people have multiple cancer types and cancer grades are not comparable). The paper reporting a description of the cohort and the development of these measures is published in BMJ Open. Development of the scale was not completed until September 2018. Analysis for the primary outcome is completed and published in BMJ Open but the flaws in the scale has resulted in more questions than answers. Additional funding has been awarded by the NIHR for a follow-on study using the same methodology in a single time point. The survey data will be analysed at the National Cancer Registration and Analysis Service using the syntax and analytical codes developed in BRIGHTLIGHT. The aim is to complete this complex analysis in readiness for the analysis of the new dataset in 2023/24.

The legal basis for processing the personal data is Article 6(1)(e): the controller uses HES data is to enable analysis to continue on a research project that aims to improve the quality of care for teenagers and young adults with cancer. The legal basis also relates to Article 9(2)(j): is necessary for archiving purposes, scientific or historical research purposes or statistical purposes, the team have health and ethnicity data for all participants in the cohort as these are central to the aims of the study, which focuses on cancer care and equality in the delivery of care (based on socioeconomic, geographical and ethnicity). Data are pseudonymised and held against study numbers only. This meets the public interest because young people currently have the poorest outcomes in comparison to children and older adults, yet their potential worth to society is huge in terms of becoming a ‘useful citizen’. Providing specialist services that deliver high quality cancer care but in addition, support returning to work and education, providing peer support and psychological care has been implemented to mitigate the outcome differences. However, this is at a cost not only to the NHS by providing these specialist units and healthcare professionals but also to young people who often need to travel a distance to access them. BRIGHTLIGHT aims to quantify the value of specialist services and understand the barrier to accessing these.

Cancer is the second leading cause of death for young people accounting for 11% of deaths in Teenagers and young adults (TYA) aged 15-24. While potentially curable for many patients, there is evidence that outcomes for some cancers have not improved in line with those achieved for children and older adults. The needs of TYA are poorly met by the well-developed cancer services traditionally tailored towards the needs of children and those for older adults with cancer.

A number of issues advocate special attention for young people with cancer and there is a critical need for a robust evidence-base to support current and future healthcare policies. For example, a unique spectrum of cancer types occur in TYA which are distinct from those affecting younger children and older adults. A cancer diagnosis in TYA has an acute and unique impact on a critical and complex stage of life development, disrupting physical health, social and educational goals as well as psychological wellbeing.

Concern has arisen that traditional cancer services are insufficient for TYA. Young people frequently fall between children’s and adult cancer services, into what has been described as 'the grey zone' or 'no man's land'. The consequence of this is realised when lesser improvements in outcomes for young people are observed compared to children and some older adult cancers.

There are several potential causes of the shortfall in outcome (survival) improvements which include but are not limited to:

• delays in cancer diagnosis;

• unfavourable tumour biology as increasing age;

• inconsistent use of molecular diagnostics that may be central to optimal care;

• limited access to clinical trials;

• lack of concordance with treatment protocols; and

• a lack of specialist supportive care.

Young people themselves describe unsatisfactory experiences of care which include:

• lack of recognition of their autonomy;

• failure to maintain their need to continue to meet normal life goals during treatment;

• lack of peer support;

• care by staff with little experience of young people; and

• inappropriate care environments.

Psychosocial and healthcare needs of this specific population is increasingly highlighted in the international literature. Place of treatment and cancer care, in terms of both disease and age appropriate specialist settings is increasingly acknowledged as potentially significant to the outcome for TYAs with cancer.

In recent years there has been a rapid expansion in the availability of dedicated services for TYA in the UK. It is now accepted that young people should have access to specialist cancer care. Thirteen principal treatment centres are currently in place. Key components of services include tumour site-specific expertise delivered in conjunction with meeting the broader psychosocial needs of young people to support successful navigation of critical life transitions.

The NICE Improving Outcomes Guidance does not direct that care of all 13-24 year olds will take place in specialist centres. Instead, it recommended that all patients aged less than 19 years are referred to principal treatment centres for their treatment. Those 19 years and over should be offered 'unhindered access to age-appropriate care'. This division resulted from: a requirement to be consistent with the National Service Framework for Children and Maternity Services; and in recognition of the heterogeneity of medical and personal need in older young people; and finally, an acknowledgment that there was insufficient persuasive evidence to mandate a greater degree of centralisation of care.

Thus, 19-24 year olds should be offered choice of place of care, either referral to principal treatment centres or more local, adult cancer services. Initiatives to support information giving to assist young people to decide on a place of care have begun but their effectiveness is as yet unknown (www.nhs.uk/young-cancer-care/pages/cancer-care-choices.aspx). Other settings of non-specialist care include 'shared care centres', usually the closest local hospital to an individual patient’s home, where management of acute complications of treatment, and other aspects of care, may occur either in children’s or adult services.

BRIGHTLIGHT is a cohort survey of 1,114 young people newly diagnosed with cancer. University College London Hospitals NHS Foundation Trust (UCLH) required HES data in order to derive a measure of specialist care for the BRIGHTLIGHT teenage and young adult (TYA) cancer cohort study, specifically the proportion of overall hospital care taking place in dedicated teenage cancer centres or specialist cancer centres within the first 6 months of diagnosis. This information will then continue to be used to evaluate the added benefit to patients of being treated at a specialist unit in relation to their general health and wellbeing, and the cost-effectiveness of specialist care in comparison to other types of cancer care.

UCLH calculated the proportion of specialist care based on admitted patient care data based on the analysis undertaken in 2008 by a PhD student. Are subsequent critique of this method concluded that for some diagnoses, none of the treatment young people received would be recorded in APC but outpatient care so UCLH potentially have under/over reported access to specialist care. UCLH now want to revise this analysis included OPD data as well.

Data were pseudonymised based on the dates of admission because UCLH need to link to time of diagnosis (obtained from NCRAS).

The cohort were diagnosed between July 2012 and December 2014 so UCLH hold HES data that covers up to a minimum of 12 months post diagnosis for all cohort members.

The cohort were recruited from across England. UCLH only have data that links to the cohort and reflects where they were treated. There was no alternative way for accurately getting this information.

UCLH only hold the data required to identify treating hospital, reason for admission (to determine that the admission was related to their cancer and not for some other reason, such as pregnancy), and dates of admissions.

University College London Hospitals NHS Foundation Trust are the sole data controller who also process data. University College London is a data processor. The Professor of Teenage and Young Adult Cancer Research at Leeds Teaching Hospitals and the University of Leeds was a co-applicant on the BRIGHTLIGHT grant. He is using some of the patient-reported outcome and experience data in a subsequent research study funded through the Economic and Social and Research Council. The co-applicant has no involvement in decision-making and none of the data being shared is from the HES dataset, and is patient-report data only.

Young people have been extensively involved in BRIGHTLIGHT from inception to the current day. Details of our first 10 years have been published showing the breadth of involvement of the Young Advisory Panel (YAP): https://researchinvolvement.biomedcentral.com/articles/10.1186/s40900-018-0135-x

UCL have continued to work with the YAP and in 2020 undertook a series of virtual workshops to get an interpretation of the programme grant results, to refine the original survey to reflect care in 2021 and to inform secondary analysis of survey data on the route to diagnosis. The YAP continue to be involved in studies subsequent to this and applicant will be holding engagement events in 2022/23 as part of the dissemination strategy. All future plans for analysing any of the BRIGHTLIGHT dataset will involve their collaboration and/or involvement as co-researchers.

The only data being processed at the University of Leeds is BRIGHTLIGHT Survey data. No data from NHS Digital will be processed outside of UCL/UCLH.

The data was processed for the analysis to support the original outputs 1, 2, 6 and 7 (details of which can be found in the 'Expected Output' section) and shall also continue to include the study of the variation in diagnostic pathways (e.g. diagnosis following emergency admission to hospital) and diagnostic intervals (e.g., time from symptom to diagnosis) for TYA cancer patients, and their predictors and consequences (e.g., survival, experience of subsequent care). The analysis will also continue to inform the health economic evaluation of BRIGHTLIGHT and to explore the diagnostic pathway for each participant in the study.

UCL will undertake an economic evaluation to estimate the short-term costs and benefits associated with specialist cancer care. Unit costs (NHS Reference Costs) will be applied to the admitted patient, outpatient and Accident and Emergency (A&E) data based on HRG, diagnosis and operative procedure codes. Costed episodes (admitted patient data) and contacts (outpatient and A&E data) will be collapsed to the patient level using patient level identifiers and linked across datasets and years to produce a patient-level dataset of NHS hospital costs for BRIGHTLIGHT patients. Data on patient-level primary care contacts, drugs and costs, out-of-pocket expenses, survival, health-related quality of life (measured using the EQ-5D) and quality-adjusted life years (QALYs) from the BRIGHTLIGHT database will be added to produce a patient level dataset of costs and QALYs. To analyse these data regression analysis, regressing patient level costs and QALYs will be used against the level of specialist cancer care each patient received, adjusting for patient characteristics including age, gender, ethnicity, deprivation, comorbidities and diagnosis. Predictive margins will be calculated and these will provide a measure of the short-term costs and benefits associated with different levels of specialist care. These short-term costs and outcomes will then be used to predict the lifetime cost and QALYs of each patient data from published sources, which will be used to calculate the lifetime incremental cost per QALY gained of different levels of specialist care.

To study the variation in diagnostic pathway a variable for diagnosis through ‘emergency presentation’ (or not) will be derived from the start and end dates of HES episodes (inpatient/outpatient/A&E data) and variables for the method of admission (inpatient data) and source of referral (outpatient data). This emergency presentation variable will be considered an exposure in analyses to address the following objectives:

1. Variation of diagnostic intervals, according to diagnostic route (e.g. diagnosis shortly after an emergency admission to hospital – representing ‘emergency presentation’ route to diagnosis), and use of certain pre-diagnostic investigations (e.g., endoscopies).

2. Clinical (stage, survival) and patient-reported (experience) outcomes associated with different diagnostic pathways and intervals.

3. Quality of life outcomes (e.g., education, relationships) associated with different diagnostic pathways and intervals.

4. The representativeness of the sample with respect to survivorship bias and non-response patterns

The original processing activities are:

Each study participant will have linked in-patient (including day case) HES admissions data recording their NHS number, HES_ID, episode start and end dates, episode order, consultant code, and provider code. HES data will also be linked to the scores on the EQ5D, a health status utility score, to enable health economic analysis to be calculated. This has been undertaken and completed at University College London (UCL).

UCLH has provided NHS Digital with study ID, NHS number and date of birth for linkage to HES. NHS Digital will provide the linkage to HES APC, OP and A&E.

HES cleaning has been undertaken by UCL to ensure no duplicate episodes, no multiple admissions less than 2 days apart with the same HES_ID, no admission entries allowed after date of death (if deceased); admissions limited to any occurring from and which overlap the date of diagnosis.

Consultant and provider codes will be checked and linked against a TYA lookup database compiled by the BRIGHTLIGHT study team, and each episode coded as TYA or not. This will be repeated for any site specialist care. This was completed as originally planned but based on the qualitative aspect of the study, age-appropriate care has been defined so we need to reclassify hospitals as 'specialist TYA' care based on criteria. This is work that hasn't been able to progress due to the pandemic. When they have been reclassified then a new lookup table will be developed.

Once the data is received by UCLH, UCLH will use the data to produce look up tables and lists of hospitals that provide specific specialist services. The data along with a hospital code will be sent to UCL to process. The outputs will not identify at individual hospital level, the outputs will identify the specialist service as a whole.

The look-up tables and lists of hospitals that provide TYA and tumour-specific specialist care will be compiled from knowledge of current service configuration. After this has been checked by key professionals working within cancer services in England, this information, with hospital/Trust code, will be sent to UCL to link to HES data. The outputs will not identify outcomes at individual hospital level, the outputs will identify the specialist service as a whole.

Overall and TYA/site specific in-patient activity will be calculated for each participant by aggregating length of stay across all episodes extending up to 6 months after diagnosis.

All data will be processed within the Department of Applied Health Research, University College London and University College London Hospitals NHS Foundation Trust . All individuals with access to the data are employed by UCL or UCLH no other third party will have access to the data.

There was a minor update to a previous HES extract in order to rerun the linkage and

1) obtain the 3 and 5 digit provider codes,

2) include amended details for 37 individuals for whom no linkage match could be found in the previous run

This Agreement relates to original outputs 1, 2, 6 and 7 below that are still to be produced.

The original outputs were -

1) a study specific measure of specialist care for each individual cancer patient, This has been completed and a number of publications reported using this metric (All references related to this study are in the final report published in the NIHR journal library: https://www.journalslibrary.nihr.ac.uk/pgfar/pgfar09120/#/abstract

2) research articles in peer review journals (anticipated journals include Lancet Oncology, British Journal of Cancer, European Journal of Cancer), This has been completed and a number of publications reported using this metric (All references related to this study are in the final report published in the NIHR journal library: https://www.journalslibrary.nihr.ac.uk/pgfar/pgfar09120/#/abstract

3) patient newsletters UCL have moved away from newsletters to using Twitter and blogs on the BRIGHTLIGHT website – www.brightlightstudy.com

4) funding reports to the NIHR. See the link to the final published report https://www.journalslibrary.nihr.ac.uk/pgfar/pgfar09120/#/abstract

5) Assessment of what proportion of the Teenage and Young Adults (TYA) cancer population received care at specialist centres between 2012-2014 and whether there was equitable access according to socio-demographic characteristics, e.g. socioeconomic group, ethnic group; UCL only recruited 20% of the diagnosed population so we were unable to do the analysis related to inequalities of care. Details are in the final report https://www.journalslibrary.nihr.ac.uk/pgfar/pgfar09120/#/abstract. There are two additional manuscripts we are in the process of submitting for publication in Cancers and Journal of Adolescent and Young Adult Oncology

6) Determine the benefit to patients receiving care at specialist TYA units in comparison to those who received care at general cancer units through evaluation of quality of life measures;

7) Assess the cost-effectiveness of specialist TYA care compared to other models of care. Other models of care will be general adult care and cancer-specific centres (e.g. bone, sarcoma). UCL have completed this and reported in full as appendices in the final report https://www.journalslibrary.nihr.ac.uk/pgfar/pgfar09120/#/abstract

There are a number of publications that are now listed on the study webpage: www.brightlightstudy.com (output 2) and patients have been kept updated through the study newsletters (output 3), although this was restricted to only those who gave their email addresses in the final point of data collection to comply with GDPR. All contact with participants has now ended through email addresses and information is posted only on the website and through social media. The study funding reports have been submitted on time to the NIHR (output 4) and the final report was submitted to the NIHR in April 2020. This has been peer reviewed and is currently with the editors for publication. Data provided by PHE from NCRAS indicated that the teenage and young adult BRIGHTLIGHT cohort only includes 20% of the population so it is not possible to conduct the analysis for output 5.

The additional outputs this will afford the study are:

i) Publications from the cohort using patient reported outcome data, clinical process data and hospital activity data (the first description of the cohort, primary and secondary outcome papers are all published in BMJ Open).

ii) The publication related to health economics including the HES data is through the NIHR final report.

iii) Presentation of the final results was restricted as many of the conferences were cancelled in 2020. UCLH are presenting them at the forthcoming TYAC national conference in September 2021, the Christie TYA Research Symposium in October 2021 and aspects of the study will be presented at the Global AYA Cancer Congress in December 2021.

iv) The revised scale and how it was developed will be published in an academic journal and submitted as a presentation for the 5th Global AYA Congress (anticipated in May/June 2023).

v) The syntax and coding for developing the revised scale will be used in the analysis of other NIHR funded research.

There have been considerable changes to service delivery since the start of the study. The Clinical Reference Group for Children and Young People’s Cancer submitted a proposed service specification to guide commissioning of services, which recommends joint care. This is contrary to the key finding from the cohort that care delivered jointly through a specialist and non-specialist unit is associated with poorer quality of life, other patient outcomes and is more expensive. The qualitative study on the other hand indicated that the change in culture over the duration of the study (7 years) may negate this result. Retaining the HES data to enable more exploratory analysis will enable UCLH to refine the metric UCLH developed as the exposure variable. UCLH have additional NIHR funding to re-run the survey in 2021/22 so UCLH will be able to apply this new metric to as the exposure variable to test the theory that culture has changed.

Additional analyses of data are also planned and/or underway as the BRIGHTLIGHT survey contained a comprehensive range of questions reflecting the range of issues encountered as a result of a cancer diagnosis. This includes exploring the pathways and routes to diagnosis to enable interventions to be developed for earlier diagnosis (paper published in JAMA and a second under review in the British Journal of Cancer); evaluation of factors influencing mental health problems after diagnosis to support the development of interventions to promote emotional well-being (analysis being undertaken by students at UCL); and work exploring caregivers information needs so they are better able to support young people (analysis now complete by a nurse at Birmingham University). UCLH are meeting Professor Hough, Chair of the CYP Cancer CRG to discuss the gaps in evidence and the potential use of BRIGHTLIGHT data to address this. It is anticipated that this will include more health economic analysis, which is integral to the HES dataset.

UCLH anticipate that results from all this additional work will be presented widely at national and international conference throughout 2022/23, to as varied an audience as possible. This includes multi-disciplinary cancer conferences (National Cancer research Institute (NCRI), Teenage and Young Adults Cancer (TYAC), European Society for Medical Oncology (ESMO), International Society for Paediatric Oncology (SIOP) annual conferences), psychology (IPOS), nursing (RCN International Nursing Research Conference), and quality of life (ISOQOL).

Finally, UCLH have recently had a grant approved by the NIHR specifically to disseminate BRIGHTLIGHT result. This includes working with the Policy Lab at King’s College London to work with key stakeholders to develop policy recommendations for TYA cancer services.

All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

The key benefit afforded through this agreement will be to undertake more detailed exploration of the scale developed using HES data received from NHS Digital. The appendix to a BMJ Open paper describing the cohort, outlines the development of the scale. This is what was described at the onset of the study but over the last 10 years the team have a more detailed understanding of the delivery of cancer services to young people so this scale is potentially too simplistic. Rather than using classifications to broadly define services as specialist or not, the extension will enable classifications to be made at Trust and hospital level.

The Clinical Reference Group for Children and Young People’s Cancer submitted a proposed service specification to guide commissioning of services, which recommends joint care. This is contrary to the key finding from the cohort that care delivered jointly through a specialist and non-specialist unit is associated with poorer quality of life, other patient outcomes and is more expensive. The results from this study will have an important contribution towards future consultation as they will indicate the model of care that will be most beneficial to young people but cost effective to the NHS.

This study is producing evidence that can contribute to efforts to improve diagnostic timeliness through better recognition and appraisal of symptoms by patients and clinicians. The James Lind Alliance’s recent priority setting exercise for TYA with cancer highlighted the need for psychological interventions to support young people as the key priority. BRIGHTLIGHT data is the largest national dataset that will be able to provide the descriptive observational analysis to underpin the development of these interventions. Central to these analyses will be the HES data on location of care and the cost data that has been calculated from this.

UCLH were unable to progress this through their previous extension request due to the impact of the pandemic (the researcher undertaking this was redeployed). UCLH will measure the benefit through the influence of the results on future policy. UCLH are working with the Chair of the CRG to ensure their evidence is informing future commissioning and likewise that their future research is designed to inform the delivery of care.

To date detailed understanding has been gained of recruitment of young people to research, which has contributed to the NIHR investing in the infrastructure to specifically improve the involvement of young people in research. Also a detailed understanding has been gained of the culture of teenage and young adult cancer care delivery from a focused ethnography study, which will help guide additional analysis of the HES data in refining the scale (referred to in output 1).

There have been a number of benefits and positive outputs from what has been learnt about conducting a cohort study with young people. First, a process of working with young people as co-researchers and collaborators in research has been established (Taylor et al. 2011; Gibson et al 2012; Fern et al. 2013; Taylor et al. 2015a; Taylor et al. 2016; Taylor et al. 2017; Taylor et al. 2018). The latter referenced paper presents an overview of this experience with guidance to other researchers; this guidance has been presented and shared in conferences and other research teams nationally and internationally contact the research team for support in user involvement (also see a blog written for international clinical trials day for the NIHR: https://www.nihr.ac.uk/blogs/does-patient-and-public-involvement-and-engagement-ppie-in-research-really-make-a-difference/6264). UCLH have diversified how they work as a result of the pandemic to include more virtual meetings, which has increased young people’s involvement in their work. Papers reporting these experiences and results are currently in development.

A further benefit relates to the recruitment of young people to research. Recruitment of young people to clinical trials is well established as being problematic. The model of recruitment into the cohort had been tested during feasibility work and was thought to overcome a number of the challenges previously reported. However, despite national buy-in and much publicity, recruitment was not as successful as the team had planned, taking much longer to recruit the cohort. Additional adjunct studies and reviews were undertaken in order to better understand the pitfalls and to make recommendations for improving young people’s recruitment to cancer research (Fern et al. 2014; Kenten et al. 2017; Fern and Taylor 2018). This was accompanied by meetings with key stakeholders who could influence recruitment, including the Association of the British Pharmaceutical Industry, NIHR, NHS England, and the National Cancer Research Institute. Recruitment to trials has become a key objective in the recent cancer strategy (recommendation 45) and based on the BRIGHTLIGHT experience, members of the team are central to work that is commencing in 2019 looking at re-configuring the existing research workforce to achieve the target to recruit 50% of young people to clinical trials by 2025.

UCLH have had applications submitted for additional funding to develop the clinical guidelines and develop interventions based on these experience. To date these have not been successful but UCLH have a further grant being submitted in November 2022, which is being supported by London Leadership to undertake some safe to fail experiments. The results from these single centre results will strength the application for a national study. UCLH have also contributed to consultation exercises by NHS DigiTrials on the acceptability of methods to young people.

UCLH understand more now about the impact of specialist care on patient outcomes (Taylor et al. 2020, Fern et al, 2021) and also on the unmet needs of carers (Martins et al. 2019). Other studies in the programme grant have identified the education and training needs of the workforce caring for young people (Taylor et al. 2016, Lea et al. 2021) and the culture in which care is delivered (Lea et al. 2022). Finally, additional analysis of cohort data have explored the symptoms prior to diagnosis and association with the timeliness of diagnosis (Koo et al. 2020) and for the first time UCLH have shown a relationship between prolonged routes to diagnosis and outcome (Forster et al. 2022).

Here is a link to the study website for further information: www.brightlightstudy.com