Cancer Survivorship Studies

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

[54 paragraphs unchanged] Where individuals have opted out of disease registration by the National Disease Registration Service (NDRS), their data has been permanently removed from the registry and therefore will not be disseminated under this Data Sharing Agreement (DSA). https://digital.nhs.uk/ndrs/patients/opting-out

Processing activities

All organisations party to this agreement must comply with the data sharing framework contract requirements, including those regarding the use (and purposes of that use) by “personnel” (as defined within the data sharing framework contract i.e. employees, agents and contractors of the data recipient who may have access to that data). [54 paragraphs unchanged]

Unchanged: Expected output, Expected measurable benefits, Benefits reported.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: + NDRS Cancer Registrations

Objective for processing

[8 paragraphs unchanged] The British Childhood Cancer Survivor Study (BCCSS) is a national population-based cohort [62 words unchanged] years from diagnosis, in Britain. This cohort was flagged with NHS Digital (now NHS England) back in 2008/09 and legislation in place at the time allowed NHS [47 words unchanged] with a Scottish GP and forward this information routinely to the CCCSS. [32 paragraphs unchanged] The combined BCCSS and TYACSS cohorts will provide a unique national resource [73 words unchanged] level, cohort data and the results of the requested linkages with NHS Digital England with any third party. Any potential future data sharing of the combined [11 words unchanged] involved to undertake the appropriate data sharing applications and procedures with NHS Digital England and stakeholders involved, including Legal Services and Research Governance at the University [40 words unchanged] agreements in place, prior to this an approved sublicense agreement with NHS Digital England would also need to be in place. [1 paragraph unchanged] The University of Birmingham is the sole Data Controller as it determines [8 words unchanged] personal data. The University of Birmingham also process data disseminated by NHS Digital England for the purposes of this research. [10 paragraphs unchanged] Where individuals have opted out of disease registration by the National Disease Registration Service (NDRS), their data has been permanently removed from the registry and therefore will not be disseminated under this Data Sharing Agreement (DSA). https://digital.nhs.uk/ndrs/patients/opting-out

Processing activities

[1 paragraph unchanged] The University of Birmingham, CCCSS, will provide the patient identifiers for the cohort, including a unique identifier, to NHS Digital England via Secure Electronic File Transfer (SEFT). NHS Digital England will link the cohort to the requested datasets, then remove the patient [8 words unchanged] CCCSS, the unique identifier plus all requested pseudonymised fields for the datasets. NHS Digital England will disseminate the data to the University of Birmingham, CCCSS, via SEFT. The BCCSS cohort is an extract of cancer survivors from the National [23 words unchanged] as ‘index number’ – this is the ‘study reference number’ in NHS Digital England records). The names, date of birth, NHS number for individuals in the cohort have been used to link the cohort to national registers of cancer and death. The individuals in the TYACSS are identified from a cancer registration cohort [21 words unchanged] as ‘index number’ – this is the ‘study reference number’ in NHS Digital England records). The names, date of birth, NHS number for individuals in the cohort have been used to link the cohort to national registers of cancer and death. [1 paragraph unchanged] The single combined cohort, described above, will be used by the CCCSS [61 words unchanged] other national datasets of outcomes in addition to those held by NHS Digital. England. Further details are provided in Section 5a “Objectives for processing” of this [34 words unchanged] and any additional application processes having been completed and approved by stakeholders. [13 paragraphs unchanged] Personal identifiable information required for linkage purposes will be stored separately from the pseudonymised linked data provided by NHS Digital. England. As described above each cancer survivor has been randomly assigned a unique [139 words unchanged] College of Medical and Dental Sciences (COLMDS) that satisfies University Security requirements. [2 paragraphs unchanged] The CCCSS and NHS Digital England will be informed by COLMDS-IT Services prior to any changes to the [16 words unchanged] Any such changes must be agreed between the CCCSS and COLMDS-IT Services. The unique identifier plus prefix ONLY will be used to identify data disseminated by NHS Digital England for each cancer survivor in the combined cohort. [11 paragraphs unchanged] All CCCSS staff who will have access to the NHS Digital England data are substantive employees of University of Birmingham or CCCSS PhD students [64 words unchanged] responsibilities. Study Centre Staff will never disclose identifiable data to third parties. [4 paragraphs unchanged] In addition to the data disseminated by NHS Digital, England, CCCSS shall undertake individual patient electronic record linkage between each of the BCCSS and TYACSS cohorts and several national outcome registers/databases including: [13 paragraphs unchanged]

Changed only in punctuation, spacing or capitalisation: Expected measurable benefits.

Unchanged: Expected output, Benefits reported.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: + Cancer Registration Data; + Civil Registrations of Death; + Demographics; + Emergency Care Data Set (ECDS); + Hospital Episode Statistics Accident and Emergency (HES A and E); + Hospital Episode Statistics Admitted Patient Care (HES APC); + Hospital Episode Statistics Critical Care (HES Critical Care); + Hospital Episode Statistics Outpatients (HES OP); + Mental Health Minimum Data Set (MHMDS); + Mental Health Services Data Set (MHSDS); + Mental Health and Learning Disabilities Data Set (MHLDDS)

Objective for processing

The objectives are to establish a national population-based cohort of 34000 individuals diagnosed with cancer before aged 15, between 1940 and 2005 inclusive, in Britain, and surviving at least 5 years from diagnosis. Investigate observed and expected risks of specific causes of death, subsequent primary cancers and other serious non-cancer morbidity using existing registries and databases including the national death and cancer registries, Hospital Episode Statistics for England, the Patient Episode Database for Wales, the Information Services Division linked database for Scotland and the Myocardial Ischaemia National Audit Project for England and Wales. There is no comprehensive national system to monitor adverse health and social outcomes among the entire population of survivors of childhood, teenage and young adult cancer in Britain. However, there already exists two established national population-based cohorts of such survivors, which would enable such a comprehensive monitoring system to be created. The study aim will be to establish a system to monitor the risks of adverse health outcomes and related healthcare activity and cost among these survivors, and to determine how observed risks and costs compare with those expected from the general population to determine subgroups of survivors who experience substantially increased risk and those service users who require more support. Adverse health outcomes and associated costs would be obtained from electronic record linkage of the cohorts with existing national registries/databases. The report by the Independent Cancer Taskforce “Achieving World-Class Cancer Outcomes –A Strategy for England 2015-2020” https://www.england.nhs.uk/publication/achieving-world-class-cancer-outcomes-a-strategy-for-england-2015-2020 emphasised the importance of risk stratification of cancer survivors in relation to their risk of developing serious adverse health conditions to ensure that the intensity of clinical follow-up care is in proportion to such risk. It is an unfortunate fact that individuals who have survived cancer experience greater risks of adverse health conditions, and greater risks of dying, than is expected from rates of these events in the general population. The National Cancer Research Institute, NHS-England and UK Health Security Agency (UKHSA) have each produced cancer strategies which also emphasise the importance of such risk stratification. It is estimated that by 2030 there will be 4 million individuals living with the long-term consequences of cancer and its treatment, but unfortunately there is to date very little research on the problems which they experience, the causes and how they might be prevented or reduced in the future. The purpose and legal basis for the processing of these cohorts by the University of Birmingham to carry out research in their Centre for Childhood Cancer Survivor Studies (CCCSS) does so in the public interest and processing is necessary for archiving purposes in the public interest, scientific or historical research purposes. In particular the legal basis for the processing of these cohorts is covered under GDPR article 6.1(e): Public task: “the processing is necessary for you to perform a task in the public interest or for your official functions, and the task or function has a clear basis in law”. Also since the university are processing special category data (health data) this is covered by GDPR 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”. The British Childhood Cancer Survivor Study (BCCSS) is a national population-based cohort of almost 35,000 individuals who were diagnosed with cancer under the age of 15 years, between 1940 and 2006, in England, Wales or Scotland, and who survived at least 5 years from diagnosis. The original childhood cohort (MR787) relates to a cohort of individuals diagnosed with cancer between 1940 and 2006, aged under 15 at diagnosis and who survived at least 5 years from diagnosis, in Britain. This cohort was flagged with NHS Digital back in 2008/09 and legislation in place at the time allowed NHS Digital to liaise with Information Services Division (ISD) / NHS Central Register (NHSCR) Scotland to obtain details of cancers, deaths and demographic (NHS status) information for those cohort members who were diagnosed with cancer in Scotland, died in Scotland or who were domiciled in Scotland and registered with a Scottish GP and forward this information routinely to the CCCSS. The Teenage and Young Adult Cancer Survivor Study (TYACSS) is a national population-based cohort almost 201,000 individuals diagnosed with cancer when aged 15 to 39 years inclusive, between 1971 and 2006, in England or Wales and who survived at least 5 years from diagnosis. The BCCSS and TYACSS cohorts are maintained by the CCCSS at the University of Birmingham, established in 1998 with charity funding from Cancer Research UK and the Kay Kendall Leukaemia Fund. The CCCSS proposes to combine the BCCSS and TYACSS cohorts. This will build the cohorts into one cohort that is as up to date as possible. CCCSS are aware that most data provided under this agreement will relate to England only. The CCCSS is developing other agreements to enable coverage of Britain as a whole. The CCCSS understands that equivalent permissions and agreements need to be in place to accommodate the equivalent legislation in Scotland, for the Scottish component of the combined survivor cohort. The CCCSS is currently completing an application to the NHS Scotland Public Benefit and Privacy Panel for Health and Social Care (HSC-PBPP) via the Electronic Data Research and Innovation Service (eDRIS). As part of this application the CCCSS will be requesting data and linkage for the Scottish equivalent population of Teenage and Young Adult cancer survivors, which is estimated be comprise around 25,000 individuals, so that the combined cohort can be accurately referred so as relating to Britain. The CCCSS is seeking to establish a comprehensive system to monitor the risks of adverse health and social outcomes among survivors of childhood, teenage and young adult cancer in Britain by combining the BCCSS and TYACSS cohorts and undertaking national population-based record linkage based investigations of the absolute and excess risks of : Specific causes of death experienced by 5-year survivors of childhood, teenage and young adult cancer, using civil registration data from the national death registries. Subsequent primary cancers experienced by 5-year survivors of childhood, teenage and young adult cancer, using the cancer registration data from national cancer registries. Hospitalisation for non-neoplastic conditions among 5-year survivors of childhood, teenage and young adult cancer, using the national Hospital Episode Statistics database for England. The uptake of the full spectrum of mental health services available within the community among 5-year survivors of childhood, teenage and young adult cancer using the national Mental Health Services Datasets. Further linkage will be undertaken with additional national datasets relating to education, employment, cardiac outcomes and treatment history (chemotherapy and radiotherapy) subject to the appropriate data sharing agreements and ethical approvals being obtained. By identifying and monitoring the risks of adverse health and social outcomes, and the related healthcare activity and cost among these cancer survivors, the CCCSS can identify how observed risks and costs compare with those expected from the general population and determine subgroups of survivors who experience substantially increased risk and who may require further services. Such large-scale population-based investigations have so far provided the most reliable and unbiased evidence available, below is a summary of the impacts which research and publications produced (or anticipated) by the Centre for Childhood Cancer Survivor Studies (CCCSS) have had historically (and anticipated in the future) on the following: survivors of cancer themselves and their families; health care professionals caring for cancer survivors; policy setting of national bodies e.g. NHS England Service Specifications. counselling, educating and empowering survivors; developing plans for the follow-up and care of survivors; providing educational material for health care professionals including GPs; evaluating risks and benefits of proposals for types of treatment in the future; providing advice to national authorities in relation to specific groups of survivors who may require further follow-up, surveillance or clinical intervention, and those groups of survivors who could be discharged from follow-up; providing economic evaluation of the expected and actual costs to the health services of treating and supporting cancer survivors In order to identify and compare the observed adverse health and social outcomes in the cohort with the general population the CCCSS will require numbers and types of the equivalent adverse health and social outcomes identified from amongst the general population. The tabulated data (with small numbers not suppressed) required for this purpose is the subject of a separate application (DARS-NIC-461060-D7X5H). All attempts have been made to minimise the data requested in accordance with the research questions/aims outlined above. The CCCSS only requires limited data to identify cancer survivor's current status within the NHS (e.g. currently registered with a GP) and whether a survivor has left the NHS (e.g. Embarked). This information is required to determine whether a survivor is likely to contribute to NHS data (e.g. Hospital Episode Statistics) and to identify censorship (Embarkation, Death) from CCCSS's survival analyses. The CCCSS will minimise the data required to identify all underlying and contributing causes of deaths (for planned mortality analyses), whether the cause of death is subject to scrutiny (coroner), date of death (for censorship in survival analyses), employment at death (the CCCSS plans to examine employment history) and place of death (if hospitalised may be reflected in Hospital Episode Statistics (HES) data). The CCCSS requires data relating to administration and care to provide an insight on the costs and impact on mental health services experienced by the cancer survivor cohort. Demographics data may add a further level of specificity in this area of research. Clinical data will provide an insight into the adverse mental health outcomes and problems experienced by the cohort as compared to the general population. The least amount of the available HES and Emergency Care Data Set (ECDS) has been identified as required for the intended processing purposes and outputs from the CCCSS. The CCCSS has carefully reviewed all fields in these datasets with reference to intended processing purposes and outputs and is requesting a limited number of the available fields. The CCCSS has only requested the minimum number of fields required to generate standardised hospitalisation ratios which will measure differences in secondary care use between a cohort of cancer survivors and the general English population. As this will be done for an extensive range of hospitalisations determined by diagnosis, co-morbidity and procedures the CCCSS has had to request most clinical fields. The majority of the data requested is taken from the sections relating to admissions, periods of care, episodes and spells (admissions and discharge), patient related data (including maternity, psychiatric, diagnosis and clinical history). Data for cost analysis is taken from Healthcare Resource Group (HRG) data. The CCCSS also needs to capture data to determine healthcare costs and explore geographical variation. The CCCSS is seeking to determine how the cancer survivor cohort uses outpatient services in comparison with the English population more generally. The CCCSS has only requested the minimum amount of fields required to generate standardized Emergency Department attendance ratios comparing how the cohort of cancer survivors use EDs compared to the English population adjusted for case mix indicators and sufficient data to determine costs. The CCCSS has undertaken a high degree of minimisation by excluding a large number of fields that relate to administration, coding versions, patient pathway and management, socio-economic status, organisational and geographical data, data quality indicators and flags. The CCCSS only requires a limited number of the variables in relation to cancer registration - anniversary date, registration year, diagnostic information (site, type and behaviour of cancer), centre and registry to not only identify subsequent primary neoplasms (SPNs) in the cancer survivor cohort but to also identify duplication, multiple registration and recurrence of existing cancers in the cohort (not SPNs). Confidentiality Advisory Group (CAG) Section 251 approval has previously been obtained for both BCCSS and TYACSS cohorts and has been approved for the combined cohort. Individual consent has not been sought as the cohort includes over 240,000 survivors of cancer. It would not be feasible to obtain consent for this number of patients. This is a national registry of survivors of cancer diagnosed under age 40 years. A principal strength of the proposed study is that it is population-based thereby avoiding selection bias. The combined BCCSS and TYACSS cohorts will provide a unique national resource in that it will provide a national register of survivors of both childhood cancer and of teenage and young adult cancer. The CCSSS has been approached by independent senior researchers in the UK who have shown an interest in using data from such a national cohort as a starting point for their own research investigations. The CCCSS acknowledges that this current agreement precludes the CCCSS from sharing the results of any pseudonymised, record level, cohort data and the results of the requested linkages with NHS Digital with any third party. Any potential future data sharing of the combined cohort with third parties will require the CCCSS and the parties involved to undertake the appropriate data sharing applications and procedures with NHS Digital and stakeholders involved, including Legal Services and Research Governance at the University of Birmingham. The third parties would be required to produce evidence of independent funding and a peer-review of the proposed projects that include data from the combined cohort. No data sharing would take place until applications have been ratified and agreements in place, prior to this an approved sublicense agreement with NHS Digital would also need to be in place. There is the possibility that results from analysis and data processing undertaken by the CCCSS on the combined cohort may produce anonymised data (with aggregation and small numbers suppressed as per HES Analysis Guidance) that generate additional important scientific and clinical insights with potential benefits for cancer survivors, health care professionals and those national bodies setting policy. In such instances these aggerated and suppressed results may be shared with third parties. The University of Birmingham is the sole Data Controller as it determines the purposes and means of the processing of personal data. The University of Birmingham also process data disseminated by NHS Digital for the purposes of this research. The Brain Tumour Charity, Children with Cancer, Dutch Cancer Society and UKHSA are funding the study and have no influence on the study design or study outcomes. Patient and Public Involvement (PPI) PPI representatives provided initial feedback on the study protocol, and they are keen to be involved with the project and support the dissemination of the study results. Research Priority Setting Partnership overseen by the James Lind Alliance. The research proposed here has been identified as being among the top-ten research priorities in three Research Priority Setting Partnership initiatives overseen by the James Lind Alliance. Survivors are central to identifying such research priorities in James Lind Alliance led initiatives. Specific PPI group established for research funded by The Brain Tumour Charity. In the development of the successful grant application to The Brain Tumour Charity, CCCSS consulted with their Research Involvement Network (RIN) which is comprised exclusively of survivors of a brain tumour. Twelve members of the RIN fed-back in detail on research proposals and this impacted the research. CCCSS have agreed that two members of the RIN will join the Project Steering Group and therefore be involved over the entire lifetime of the project. Each of the 12 members of the RIN completed a detailed questionnaire which provided their views on the priorities to be given to the various research questions proposed by CCCSS. This impacted which research questions were selected to be taken forward and the priority to be given to different research questions. There have been other national research priority setting initiatives undertaken by the James Lind Alliance www.jla.nihr.ac.uk and these were also carefully taken into account. There are two members of the RIN who will be joining the Project Steering Group for their input into the outputs (particularly the website material) of the research on the combined cancer survivor cohort. Local group of survivors of childhood cancer. The Consultant Paediatric Oncologist at the Birmingham Children’s Hospital has established a local group of about 20 survivors to be consulted on for both clinical and research questions. The Consultant Paediatric Oncologist has indicated that this group may be asked to provide guidance to the survivorship research relating to survivors of childhood cancer. Birmingham Cancer Research UK Centre PPI Group. There is a Cancer Research UK funded Senior Research Nurse at the Birmingham Cancer Research UK Centre who maintains a group of about 20 survivors of a variety of cancer diagnosed at a range of adult ages. CCCSS have consulted with, and will continue to consult with, this group of survivors. Pancare. The CCCSS Director is a founding member of PanCare, an organisation for childhood cancer survivors and their families; clinicians caring for those with, or cured of, cancer; researchers addressing the needs of those with, or cured of childhood cancer. This pan-European organisation meets twice each year and has a significant survivor membership. This provides an international forum to seek input from survivors into research being undertaken or planned. The National Cancer Research Institute –Teenage and Young Adult & Germ Cell Tumour –Clinical Studies Group (NCRI-TYA&GCT-CSG). The CCCSS Director is a member of this national group and there are also two survivors of TYA cancer among the membership. This provides a national forum for input from survivors into survivorship research studies.

Processing activities

Not stated in the previous version; added here.

All organisations party to this agreement must comply with the data sharing framework contract requirements, including those regarding the use (and purposes of that use) by “personnel” (as defined within the data sharing framework contract i.e. employees, agents and contractors of the data recipient who may have access to that data).

The University of Birmingham, CCCSS, will provide the patient identifiers for the cohort, including a unique identifier, to NHS Digital via Secure Electronic File Transfer (SEFT). NHS Digital will link the cohort to the requested datasets, then remove the patient identifiers and send back to University of Birmingham, CCCSS, the unique identifier plus all requested pseudonymised fields for the datasets.

NHS Digital will disseminate the data to the University of Birmingham, CCCSS, via SEFT.

The BCCSS cohort is an extract of cancer survivors from the National Register of Childhood Tumours identified from national cancer registration. Each individual in the cohort has a unique random integer identifier (hereby referred to as ‘index number’ – this is the ‘study reference number’ in NHS Digital records). The names, date of birth, NHS number for individuals in the cohort have been used to link the cohort to national registers of cancer and death.

The individuals in the TYACSS are identified from a cancer registration cohort provided by the Office of National Statistics. Each individual in the cohort has a unique random integer identifier (hereby referred to as ‘index number’ – this is the ‘study reference number’ in NHS Digital records). The names, date of birth, NHS number for individuals in the cohort have been used to link the cohort to national registers of cancer and death.

The BCCSS and TYACSS cohorts will be combined into a single cohort. The unique integer identifiers (‘index numbers’) will be checked for duplication between original cohorts. The original unique identifiers will have an additional character prefix to distinguish them from the original source cohort, and to avoid the possibility of duplication. Both cohorts now have Section 251 support under CAG Reference 21/CAG/0078.

The single combined cohort, described above, will be used by the CCCSS to create the first national monitoring system for adverse health and social outcomes among the entire population of survivors of childhood, teenage and young adult cancer in Britain, the CCCSS is developing other agreements to enable coverage of Britain as a whole.. In order for this monitoring system to be most effective the CCCSS plans to link the cohort with limited other national datasets of outcomes in addition to those held by NHS Digital. Further details are provided in Section 5a “Objectives for processing” of this application. Linkage to these additional national datasets will be undertaken independently using data already held by the CCCSS and only in accordance with the appropriate data sharing agreements and ethical approvals being in place and any additional application processes having been completed and approved by stakeholders.

The CCCSS will obtain national cancer incidence rates and mortality rates in the general population from UKHSA and the Office for National Statistics. Tabulated data not available in the public domain that is required for this purpose is the subject of a separate, complementary, application (DARS-NIC-461060-D7X5H).

Risks of specific causes of death:

Each individual enters risk at the date of 5-year survival and contributes person-years until the exit date (first of emigration date or date of end of ascertainment). Standardised Mortality Ratios (SMRs) and Absolute Excess Risks (AERs) will be calculated as O/E and [(O-E)/py]*10000 where O and E are the observed and expected numbers of deaths, respectively, and ‘py’ is the person-years at risk accumulated. To investigate variation in SMRs and AERs across levels of risk stratification factors. Poisson regression models will be utilised. Poisson regression is the modelling of count data and is used to predict a dependent variable that consists of "count data" given one or more independent variables. Cumulative incidence of death from a specific cause will be estimated treating other causes of death as competing risks.

Risks of subsequent primary neoplasms (SPNs):

Similar statistical methodology to that described for deaths would be used to determine subgroups of survivors at substantially excess risk of specific SPNs, but the summary measures would be Standardised Incidence Ratios (SIRs) and Absolute Excess Risks (AERs defined as for deaths, in terms of observed (O) and expected (E) numbers of SPNs of a particular site/type, as in the previous work by the CCCSS). SIRs is the ratio of the observed number of cancer cases to the expected number of cases multiplied by 100. An SIR of 100 indicates that the number of cancer cases observed in the population evaluated is equal to the number of cancer cases expected in the comparison or “normal” population.

Risk of non-neoplastic adverse health and social outcomes:

The period at risk begins from the start date for ascertainment of the specific outcome and ends at the current end date of ascertainment provided the survivor does not exit through emigration or death before this end date. The risk of a specific adverse health outcome would be compared over the period at risk using Poisson regression in relation to the risk stratification factors. In this way it would be possible to identify particular subgroups at greatest risk (Internal analysis).

The CCCSS will be requesting tabulated data (with small numbers unsuppressed) from general population data (DARS-NIC-461060-D7X5H), for example, the general population hospitalisation, community mental health events classified by age, sex and calendar year. By dividing the number of events in each cell by the general population estimate of those at risk provides an expected rate for the derivation of expected numbers. Again, Poisson regression would be used to compare the observed and expected number, Standardized Hospitalization Ratio (SHR), over the levels of a particular risk factor adjusting for others (external analysis).

Cost analysis:

Unit costs will be assigned to healthcare activity using standard Department of Health (DoH) guidance. Healthcare Resource Group (HRG) codes will be assigned to secondary care episodes using the DoH Grouper software with sensitivity analysis between year-specific and common-base-year assignment. Prescriptions will be costed using the national tariff for branded medicines and the electronic medicines compendium for generic medicines. On this basis, the cost of subsequent primary neoplasms, the cost of non-neoplastic adverse events and the overall cost-profile of survivors will be described.

Comparison of overall costs and cost profiles will be made descriptively through cost-profile visualisation. In parallel with the risk analysis described above, internal analysis will identify predictors of cost using generalised linear models. External analysis based on the same cells as the risk analysis will estimate observed and expected cost compared with the general population.

There will be no requirement or attempt to re-identify individuals, except under two limited circumstances. The first being, the CCCSS has been approached by specific members of the BCCSS cohort in the past to request medical records CCCSS may hold relating to treatment for a specific cancer as these have been lost by the NHS. This has been critically important on several occasions in avoiding the risk of serious complications to a major organ through appropriate restriction of the cumulative life-time dose of radiation from radiotherapy. If a member of the combined cohort has a similar request, then that individual’s record may be identified. The second circumstance will be to identify those members of the cohort who wish to be removed from the cohort. There is no other circumstance where an individual’s record will be re-identified.

Security

Personal identifiable information required for linkage purposes will be stored separately from the pseudonymised linked data provided by NHS Digital. As described above each cancer survivor has been randomly assigned a unique integer identifier (‘index number’). This identifier is propagated through all database systems. For each survivor across the database systems, identified by index number, all personally identifiable information (names, NHS number, addresses) will be removed. All dates recorded in the database systems for each survivor, will be adjusted by a randomly generated integer offset value. The integer offset value and identifiable information for each survivor (identified by index number) will be recorded on a database that will be stored on three external hard drives encrypted to the University standards. Two of the hard drives will be stored in a fire-proof safe in a locked room within the CCCSS Study Centre which requires a swipe key and two physical locks to access. The third hard drive will be stored in a fire-safe in a separate location within the University of Birmingham College of Medical and Dental Sciences (COLMDS) that satisfies University Security requirements.

Only members of the CCCSS will have access to the encryption keys for the external hard drives.

Both a manual and electronic log will be kept to record details of access to, and changes to, the databases on the encrypted drives. Any adjustments to identifiable data will be replicated across each version of the database on each separate encrypted drive. The databases will be checked against each other for consistency. A manual log will be kept at the Study Centre along an electronic log with a description of changes, the dates the changes were made and who made the changes that will be kept on each encrypted drive and on the Study Centre network share.

The CCCSS and NHS Digital will be informed by COLMDS-IT Services prior to any changes to the location or storage situation of the encrypted drive stored in COLMDS outside of the Study Centre. Any such changes must be agreed between the CCCSS and COLMDS-IT Services.

The unique identifier plus prefix ONLY will be used to identify data disseminated by NHS Digital for each cancer survivor in the combined cohort.

The database of linked data will be stored and maintained according to the Information Security and Management Policy (ISMP) of the University of Birmingham.

The ISMP sets out the security arrangements under which sensitive data should be identified, processed and stored, while the Data Protection Policy deals with data protection and privacy.

Further information relating to University of Birmingham standards and guidance can be found at: http://itsecurity.bham.ac.uk/policy.

In accordance with the ISMP security practices shall be deployed to protect people technology and processes from deliberate attack based on a risk assessment within the framework of an Information Security Management System (ISMS) compatible with ISO27001 and related standards.

Data shall be stored on centrally managed stores (file servers) and archived in a manner to facilitate secure disposed of when no longer required. The file servers are located in locked, secure, specialised rooms with appropriate heating and ventilation, on the University of Birmingham Campus. For projects undertaken at the CCCSS the file servers reside in the College of Medical and Dental Sciences (COLMDS). Access to server rooms is limited to UBHAM Information Technology personnel only with swipe card systems.

Multiple tape back-ups of the servers occur daily (incremental) and weekly (full) and are stored in fireproof safes in separate locked locations within COLMDS. Access to rooms is limited to UBHAM Information Technology personnel only with swipe card systems.

File, database or disk encryption must be used except where compensating controls can be shown to provide an equivalent level and protection.

Data transferred over data communication networks shall be encrypted or otherwise protected to University Standards (at least AES 256 compliant). This includes email and other types of electronic messaging.

Removable media and mobile devices shall be encrypted to University Standards (at least 128-bit encryption).

Access to University systems, IT infrastructure and facilities will only be to authorized persons with a valid ID card, a University assigned username and a password that conforms to University standards (at least 8 characters long, one uppercase letter, at least one lower case letter, at least one digit, at least one non-alphanumeric character, enforced renewal every 180 days).

While all members of CCCSS staff have access to a personal file-store account a separate network share will be made available to all Study Centre staff only, for file sharing purposes and storage of the linked data.

All CCCSS staff who will have access to the NHS Digital data are substantive employees of University of Birmingham or CCCSS PhD students and have completed the mandatory training on Information Security Awareness, Data Protection and Freedom of Information. All Study Centre staff members are required to sign an undertaking of confidentiality which, if broken, would likely lead to dismissal. All Study Centre staff with access to patient identifiable information have received appropriate training and receive ongoing supervision and support to ensure they are aware of their responsibilities. Study Centre Staff will never disclose identifiable data to third parties.

There are currently two full time PhD students registered with the CCCSS. While they have not signed a specific contract, they have to declare that they will comply the policies of the University.

There is a compulsory on-line course which explains students’ legal responsibilities, especially data protection including research data. “Data Protection and Information Security for students”.

In addition, the course refers to the data protection policy: "All staff, students, honorary and associate members of staff and any other University of Birmingham Data Users must comply with this (Data Protection) Policy, and disciplinary action can be taken against those who do not comply, particularly in cases when there has been deliberate, wilful or negligent disregard of the Policy and University requirements" It is clearly stated that disciplinary procedure can be taken in case of non-compliance with the rules.

Personal data will be processed according to a valid lawful basis and conditions for processing.

In addition to the data disseminated by NHS Digital, CCCSS shall undertake individual patient electronic record linkage between each of the BCCSS and TYACSS cohorts and several national outcome registers/databases including:

• National causes of death registry in Scotland

• National cancer registry in England, Scotland and Wales

• National Institute of Cardiovascular Outcomes Research for England and Wales

• National NHS GP prescription database maintained by UKHSA

• National educational outcomes database

• British Society of Blood and Marrow Transplantation and Cellular Therapy

•Birmingham Women's and Children's NHS Foundation Trust

a. West Midlands Regional Children’s Tumour Registry (17/CAG/0103)

• NHS Wales Informatics Service, (NWIS)

a. Patient Episode Database for Wales (PEDW)

• National employment status database.

CCCSS aim to exploit the advantages which Britain has compared with other parts of the world, with the exception of the Nordic countries, because of the long-standing existence of nationwide population-based cancer registration. CCCSS have established the two largest population-based cohorts available relating to childhood cancer survivors and teenage and young adult cancer survivors – the BCCSS and TYACSS, respectively. These cohorts benefit from the longest follow-up time currently available to any similar cohorts as a result of national population-based cancer registration being established in Britain in the early 1960s.

CCCSS shall extend as specified above the electronic record linkage between each of the BCCSS and TYACSS cohorts and: the national death and cancer registries, the national hospitalisation databases, the national cardiovascular databases, the national GP prescription databases and the national community mental health services databases. This will provide ascertainment of all: deaths with details of underlying cause; subsequent primary neoplasms diagnosed; outpatient, inpatient and emergency care episodes at NHS hospitals; cardiovascular events and procedures, GP prescriptions dispensed to survivors; mental health services used within the community. Linkage to national databases relating to educational attainment and employment status will be undertaken.

Expected output

Not stated in the previous version; added here.

The CCCSS shall investigate the variation in absolute and excess risk of deaths, subsequent primary neoplasms and adverse health and social outcomes by comparing the observed cohort with the general population.

All outputs will contain only data that is aggregated with small numbers suppressed in line with the appropriate data set minimisation guidelines.

Manuscripts will be prepared for publication in international peer-reviewed journals. Authorship eligibility will be determined to be consistent with the requirements of the leading international general medical journal including New England Journal of Medicine, The Lancet, the Journal of the American Medical Association and the British Medical Journal.

There are already over 100 such manuscripts published using the BCCSS cohort, and about 10 published using the TYACSS cohort.

Funding bodies will need to be acknowledged within peer-reviewed publications, but they have no rights to alter, delay or prevent such publications.

The CCCSS intends to establish a website for the study which will contain information regarding progress and accessible summaries of the research in the form of regular newsletters. The Study Protocol, a complete list of publications and previous Newsletters will be available on the Study Website.

Senior members of the CCCSS regularly speak at the Annual Education Day organised by the Late Effects Group of the Children’s Cancer and Leukaemia Group. This is well attended (100 to 150 attendees) by those responsible for the care (doctors and nurses) of survivors of childhood cancer at Centres throughout the UK.

The CCCSS Director is a full member of both the Children’s Cancer and Leukaemia Group (CCLG) and the Late Effects Group of the CCLG.

The various Patient and Public Involvement (PPI) groups (as described in the Objectives of Processing section) are keen to be involved with the project and support the dissemination of the study results. These PPI groups are;

- Research Priority Setting Partnership overseen by the James Lind Alliance

- The Brain Tumour Charity Research Involvement Network (RIN)

- Local group of survivors of childhood cancer at the Birmingham Children’s Hospital

- Birmingham Cancer Research UK Centre PPI Group

- Pancare.

- The National Cancer Research Institute –Teenage and Young Adult & Germ Cell Tumour –Clinical Studies Group (NCRI-TYA&GCT-CSG).

Expected measurable benefits

Not stated in the previous version; added here.

It is estimated that by 2030 there will be 4 million individuals living the in UK who have been previously diagnosed with cancer. Unfortunately, there has been very little research on such survivors’ risks of potential side effects of the cancer and/or its treatment. The research that has been carried out so far relates mostly to survivors of childhood cancer and very little research is available relating to survivors of cancer diagnosed in adulthood. The necessary first step is to quantify the risks of potential side-effects experienced by survivors. Then more detailed studies may follow to understand the causes/mechanisms underlying the increased risks and this is likely to lead to insights for strategies to reduce, or ideally prevent, the occurrence of such side-effects developing among future survivors. Such risks also enable the risk stratification of the existing population of cancer survivors so that NHS resources are concentrated on those most at risk with a view to developing interventions aimed at preventing, or reducing the risk of, side-effects before they develop, or where this is not possible identifying the side-effects as early as possible with a view to improving the prognosis.

Three serious limitations of previous research relating to British survivors of cancer diagnosed when young relate to the absence of national databases relating to cardiovascular conditions, GP prescriptions and the use of mental health facilities accessed within a community setting – each of these limitations can now be overcome with the current proposal.

There has been very little previous research concerned with the healthcare costs relating to British survivors of cancer diagnosed when young and how these compare with those expected from the general population. This can now be addressed for England and Scotland because the data and expertise will be available to the project.

Uses of evidence produced for changing clinical practice and health policy:

Such large-scale population-based investigations of the risks of a comprehensive spectrum of fatal and non-fatal adverse health outcomes has provided the most reliable and unbiased evidence available for:

• feeding back to, counselling, educating and empowering survivors;

• developing evidence-based clinical follow-up guidelines;

• preparing “survivorship care plans”;

• providing educational material for health care professionals including GPs;

• evaluating risks as well as benefits of proposals for future treatment protocols;

• advising national health authorities in relation to subgroups of survivors at particularly high risk for consideration of potential recall for counselling, surveillance or other intervention;

• identification of low risk groups for potential discharge from hospital based follow-up;

• provide risk stratification information to national health authorities, particularly NHS England, to guide the evidence-based levels of intensity of clinical follow-up needed by different specific subgroups of survivors;

• provide health economic evaluations from financial information recorded in hospital activity registers to compare the observed and expected costs relating to survivors.

Impact and benefits for patients, healthcare professionals and policy makers over the next 3 years:

• The large-scale population-based design of CCCSS studies ensures that the risks produced are as comprehensive and accurate as it is possible to produce in practice.

• Every survivor of childhood, teenage and young adult cancer in England at the completion of treatment should be assessed by a clinician concerning their long-term risk of serious adverse health outcomes and whether the risk is high, medium or low. This is now stated in latest NHS England Service Specifications for survivors of childhood cancer and separately for survivors of teenage and young adult cancer. The risk stratification tool to be used was developed by CCCSS as part of the National Cancer Survivorship Initiative.

• The CCCSS plans with substantial input from survivors and healthcare professionals to establish a website were findings from the BCCSS and TYACSS cohorts will be summarized in lay terms. The CCCSS has a specific grant to fund meetings to develop the material for these websites amongst other research.

• The evidence which the CCCSS produces in relation to specific adverse health outcomes has been, and will be, included in the development of clinical follow-up guidelines relating to such specific adverse health outcomes through strong links the CCCSS has with the International Late Effects of Childhood Cancer Guideline Harmonization Group (www.ighg.org) led by key European and North American investigators. These guidelines are increasingly used in follow-up clinics throughout the world.

• Each survivor of cancer should be medically reviewed at the end of treatment and an individual survivorship care plan prepared which should be shared with the survivor and their GP. The risk stratification tool which the CCCSS developed, as mentioned above, and clinical follow-up guidelines which the CCCSS contributes to are both key elements in developing such care plans.

• The CCCSS will maintain its close links with NHS England so that the latest evidence is reflected in their Service Specifications relating to cancer survivors.

• The CCCSS will maintain its close links with the national UK professional organization of clinicians caring for children with cancer: the Children’s Cancer and Leukaemia Group (CCLG), The CCCSS Director and a Senior Colleague are full members of both the CCLG overall and its Late Effects Group. This facilitates the transfer of new evidence produced into clinical practice.

• The CCCSS will maintain its close links with PanCare (https://www.pancare.eu) a European-wide organization of healthcare professionals, survivors of childhood cancer and their families and researchers. This provides an international forum to seek input from survivors and inform survivors of new research findings relevant to them. The CCCSS Director is a founding member of PanCare.

• Continue to feedback to survivors, healthcare professionals and the NHS new evidence emerging through the various networks described above.

Benefits reported

Yielded Benefits is not a requirement for new applications. Proven impact of previous publications: The evidence which had previously been produced has had an impact in a number of policy and clinical practice areas: • The comprehensive risk stratification evidence produced already in relation to survivors of childhood, teenage and young adult cancer concerning their risk of serious adverse health outcomes in relation to their type of cancer, type of cancer treatment received, treatment era, age at treatment, years from cancer diagnosis, attained age and gender has had impact. Our initial publication concerned with the long-term risk of the total burden of serious adverse health outcomes carried out as part of the National Cancer Survivorship Initiative, identified subgroups of survivors a high, medium and low risk. This publication has recently been used as key evidence in review which NHS England has undertaken of its Service Specifications. In the future every survivor of childhood, teenage or young adult cancer will be assessed in relation to their long-term risk of developing serious adverse health outcomes using our risk stratification tool at the end of treatment and this will inform clinical decisions regarding the intensity of clinical follow-up necessary ranging from survivor self-management with easy and rapid access back into the NHS system at one end, to regular hospital consultant led multi-disciplinary team care at the other. https://www.engage.england.nhs.uk/consultation/childrens-cancer-services/user_uploads/service-specification-childrens-networks-and-principle-treatment-centres.pdf https://www.engage.england.nhs.uk/consultation/teenager-and-young-adults-cancer-services/user_uploads/service-specification-tya-principal-treatment-centres-and-networks.pdf • The recently established International Late Effects of Childhood Cancer Guideline Harmonization Group (www.ighg.org) led by key European and US investigators aims to produce standard clinical follow-up guidelines for survivors which are as evidence-based as possible and acceptable to clinical communities throughout the world. The CCCSS Director and Senior Colleagues have been closely involved in this initiative and in particular in the development of several recent and on-going international guidelines. • Advising national health authorities of subgroups of survivors with particularly high risks of specific outcomes for potential recall or other intervention. There is on-going work concerning the introduction of screening (colonoscopy/faecal occult blood sampling) for bowel cancer in survivors of childhood cancer who received external beam radiotherapy to the abdominopelvic region; also on-going work into understanding whether there are any ways to reduce the substantial risks of stroke in survivors of childhood, teenage and young adult cancer who received external beam radiotherapy for an intracranial tumour. • The CCCSS Director and Senior Colleagues regularly speak at the Annual Education Day organised by the Late Effects Group of the Children’s Cancer and Leukaemia Group. This is well attended (100 to 150 attendees) by those responsible for the care (doctors and nurses) of survivors of childhood cancer at Centres throughout the UK. • The CCCSS Director and the Senior Lecturer are each full members of both the Children’s Cancer and Leukaemia Group (CCLG) and the Late Effects Group of the CCLG. • The BCCSS receives formal clinical input into its research plans from a CCLG/BCCSS liaison/advisory committee jointly organised by the Late Effects Group of the CCLG and the Cancer Survivorship Group. • The TYACSS receives formal clinical input into its research plans from the NCRI-TYA&GCT-Group of which the CCCSS Director is a full member, and who also chairs the Survivorship Subgroup of the main Group.