Evaluating variation in special educational needs provision for children with Down syndrome and associations with emergency use of hospital care.

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

Datasets: + Civil Registrations of Death; + Emergency Care Data Set (ECDS); + HES:Civil Registration (Deaths) bridge; + 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)

Objective for processing

There is considerable variation across England in both health and education services for children with chronic conditions and special educational needs. This The initial request from University College London (UCL) relates (which received approval under a previous version of this Data Sharing Agreement) related to the first stage of a larger project to evaluate variation in service the provision from education, children’s of educational, social care and healthcare services for children with Down syndrome and to determine the impact of this on emergency use of hospital care. This The first stage will involve involved data from the National Down Syndrome Cytogenetic Register (part of the National [9 words unchanged] by Public Health England (PHE)) and Hospital Episode Statistics for England (HES). The patient identifiable HES data provided by NHSD to PHE is linked [21 words unchanged] and Rare Diseases Register (https://www.gov.uk/guidance/the-national-congenitalanomaly-and-rare-disease-registration-service-ncardrs). PHE has Section 251 support to collect information data on cases of CARDs individuals with congenital anomalies or rare diseases in England and to link this to other sources of information on the diagnosis, treatment (such as HES) and outcomes of these individuals. A planned second stage of the research will include linkage to the National Pupil Database but this is not permitted under this Data Sharing Agreement. Approval for the second stage will be sought as an amendment when the Department for Education has acquired the necessary certification for their information governance procedures. Objectives relating to the first stage are as follows: This research is jointly funded by the Economic Social Research Council (ESRC) and the Administrative Data Research Network (ADRN). The ESRC fund the Administrative Data Research Network (ADRN) (soon to be the Administrative Data Research Centre for England (ADRC-England)). The ADRN will be the public facing website for this study. ADRC - England is led by the University of Southampton and runs in collaboration with (amongst others) UCL, who are the data controllers for this request. ADRN have a prominent role in cross-sectoral linkages hence why the study information is being published on the ADRN website. Objective A1: Identify a cohort of children with and without Down syndrome using linked data from the National Down Syndrome Cytogenetic Registry (NDSCR) and Hospital Episode Statistics (HES). ESRC is the funder for ADRN/ADRC and support for the study will also be provided by the Children’s Policy Research Unit (CPRU) at UCL, which is funded by the Department for Health. The ADRN facilitates requests to access administrative data for research by independently reviewing applications and liaising with stakeholders. The ADRN review panel approved this study and published a summary on their website. The ADRN and ADRC-E are multicentre collaborations involving researchers from many institutions. This project is being led by ADRC-E researchers at University College London only and does not involve any researchers from the University of Southampton. The Institute of Education is an Institute of University College London but was a separate institution at the time of preparing the protocol. Lorraine Dearden, who has an appointment with the (UCL) Institute of Education is a co-investigator and contributed to production of the protocol. Only substantive employees of UCL will have access to the data specified in this Agreement. A project summary has been published on https://adrn.ac.uk/research-impact/research/project165/. A summary will also be published on the CPRU website http://www.ucl.ac.uk/children-policy-research. Objectives relating to this first stage are: Objective A1: Identify a cohort of children with and without Down syndrome using linked data from HES and NDSCR. [1 paragraph unchanged] Public Health England will be supporting the research by linking their NDSCR data to the copy of HES that they already hold, to identify the study cohort. PHE will generate pseudo-identifiers for the cohort and provide pseudonymised NDSCR data to researchers at UCL. PHE will also attach the pseudo-identifiers to a file containing matching variables from HES for the cohort, and send this file to NHS Digital for further linkage. The initial request was approved and these objectives are in progress, with the first stage of data processing nearing completion. Using the matching variables provided by PHE, NHS Digital will link the cohort to the Personal Demographic Service to provide updated postcode histories in preparation for stage 2 (stage 2 will involve sending identifiers to the Department for Education for linkage but approval for that sharing does not form part of this application). NHS Digital will then attach encrypted HESID pseudonyms to the cohort using the encryption key for the HES dataset already held by UCL researchers (NIC-393510-D6H1D) and send to UCL a file containing only cohort ID, HESID, ONS mortality data, and indicators of linkage quality. This project is being led by researchers at University College London only and does not involve any other researchers. The Institute of Education is an Institute of University College London but was a separate institution at the time of preparing the protocol. Only substantive employees of UCL will have access to the data specified in this Agreement. UCL will use the incoming file from NHS Digital to extract a subset of the data already provided under NIC-393510-D6H1D, for use in this project under a new data sharing agreement (just as if NHS Digital has provided a new project-specific extract, but avoiding an additional data flow). They will combine the HES data extracted from the existing project with the incoming NDSCR data from PHE and carry out analyses on the linked data in pursuit of the objectives stated above. Legal basis: The legal basis for processing falls under GDPR Article 6(1)( e) and [74 words unchanged] and without Down syndrome, and how they vary over time and region. It will also provide evidence on the effect of the COVID-19 pandemic on health outcomes and use of healthcare services among children and young people with Down syndrome. The legal basis for NHS patient identifiable information being processed by NHS Digital is given by Section 251 of the NHS Act 2006 and the legal basis for disseminating pseudonymised data to UCL Data Safe Haven and ONS Secure Research Service (SRS) is given by Section 261 (1) and Section 261 (2)(b) of the Health and Social Care Act 2012. The legal basis for the sharing of data by the Department for Education is given by: section 537A and 537B of the Education Act 1996, section 83 of the Children Act 1989, section 99 of the Childcare Act 2006 and section 253A of the Apprenticeships Skills Children and Learning Act 2009. AMENDMENT REQUEST The second stage of planned work, which was not covered by the initial request but which UCL request approval for in this amendment, aims to address the following additional objectives by linking NDSCR-HES data for the study cohort to the National Pupil Dataset (NPD) held by the Department for Education (DfE). Objective B1: Link the cohort of children to the NPD, including looked after children data. Objective B2: Evaluate the methods of accurate linkage of children with chronic health conditions and their education administrative records. Objective B3: Assess the variation in type of school and special educational needs provision and ‘Looked After Child’ status in children with Down syndrome vs the general population, and the association with subsequent unplanned hospital admissions. The initial request for the first stage of the project proposed linkage of the NDSCR-HES extract to the Personal Demographic Service (PDS) by NHS Digital. The purpose of linkage to PDS was to obtain updated postcode histories and names for the study cohort for onward linkage to NPD data by NHS Digital sending the DfE an ‘identifier’ file including the cohort-IDs, updated postcode histories and names for the study cohort. The intention was that DfE would then use this, in the second stage of the project, to link to the NPD and prepare an extract of de-identified NPD data for subsequent linkage to the NDSCR-HES extract. Approval for this second stage was to be requested as an amendment when the DfE provided the necessary certification for their information governance procedures. However, it is no longer necessary to request that NHS Digital prepares and sends an ‘identifier’ file to the DfE. Instead, this amendment requests permission to utilise a subset of data from an existing agreement (NIC-381972-Q5F0V, the “DHSC-ECHILD-COVID” study). NIC-381972-Q5F0V covers a dataset of all children and young people born from 1.9.1995 and appearing in HES records. An objective of NIC-381972-Q5F0V is to analyse and compare health, education and social care outcomes for vulnerable and non-vulnerable children and young people before, during and post-COVID-19. As children and young people with Down syndrome would be a subpopulation of the vulnerable children identified for this project it is possible to use a subset of these data to address objectives B1-B3 of this project. This has the benefit of not requiring NHS Digital to provide the DfE with identifier variables (e.g. names and post-code histories) for linkage to the NPD and is consistent with the principle of data minimisation. Funding: This research was initially jointly funded by the Economic Social Research Council (ESRC) and through the Administrative Data Research Network (ADRN). Continuation of funding is from the NIHR GOSH Biomedical Research Centre and the NIHR Children and Families Policy Research Unit (CPRU) at UCL (which is funded by the Department for Health).

Processing activities

Data Processing: 1. Processing at PHE (agreed in previous iteration of this Dara Sharing Agreement and completed): 1 - Processing at PHE: 1A) PHE linked NDSCR to HES (both datasets held by PHE) for children born between 1.4.97 and 31.03.14. a) Link NDSCR to HES (both datasets held by PHE) for children born between 1.4.97 and 31.03.14. 1B) For each child identified as having Down syndrome through either a registration in NDSCR or clinical codes in HES, 9 controls without Down syndrome but with the same week and year of birth and local authority of residence were selected at random from HES birth admissions. PHE assigned a ‘cohort-ID’ pseudo-identifier to each cohort member (case or control). There were approximately 117,000 controls and 13,000 cases. This equated to 9 controls per 1 case. b) For every child identified as having Down syndrome through either a registration in NDSCR or clinical codes in HES, select 9 controls without Down syndrome from HES birth admissions. Assign ‘cohort ID’ pseudo-identifier to each cohort member (case or control). The control cohort are selected at random from the same week and year of birth and local authority of residence to children with Down syndrome in the NDSCR-HES cohort, using the HES data only. There will be approximately 117,000 controls and 13,000 cases. *** This equates to 9 controls per 1 case. *** 1C) PHE sent UCL a pseudonymised NDSCR-HES extract including cohort-ID, Down syndrome registration flag, diagnosis date, and linkage quality indicators. c) Send to UCL a file containing cohort ID, Down syndrome registration flag, diagnosis date, and linkage quality indicators. 1D) PHE sent NHS Digital an extract of NDSCR-HES containing cohort-IDs and conventional identifiers obtained from HES (Epikey for birth record, Financial year for birth record, NHS number, DOB, Postcode, Sex, Ethnic group, Birthweight, Birth order). d) Send to NHS Digital a file containing cohort ID and conventional identifiers (Epikey for birth record, Financial year for birth record, NHS number, DOB, Postcode, Sex, Ethnic group, Birthweight, Birth order) obtained from HES. 2. Processing at NHS Digital (agreed in previous iteration of this Dara Sharing Agreement and ongoing): 2 - Processing at NHS Digital (NHS Digital is in the process of sending the cohort-IDs and UCL-specific HES-IDs for the study cohort (using the encryption key for the HES dataset already held by UCL researchers as part of a previous agreement: NIC-393510-D6H1D) to UCL researchers. UCL will use this information to extract a subset of the data already provided under NIC-393510-D6H1D for use in this project and in pursuit of objectives A1-A2 (just as if NHS Digital had provided a new project-specific extract, but avoiding an additional data flow)). a) Match cohort to PDS, using matching variables provided by PHE (derived from HES). 2A) NHS Digital used identifying variables from the NDSCR-HES extract sent by PHE (NHS number, DOB, postcode) to identify UCL-specific HES-IDs for each cohort member. b) Prepare file containing matching variables (Study ID, PDS sequence number, First name, surname, other names, DOB, postcode, ethnic group) for future sharing with Department for Education pending approvals for stage 2 (to be disseminated under a future amendment). This process is taking place now (rather than at the amendment stage) to ensure that the file created uses the same cohort, if this is done at a later date it could be that the cohort is slightly different, therefore, the results may be incorrect. 2B) NHS Digital to prepare and share file with UCL containing cohort-ID and UCL-specific HES-ID (using the same encryption key as used for NIC-393510-D6H1D). c) Prepare file for sharing with UCL now, containing cohort ID, HESID (using the same encryption key as used for NIC-393510-D6H1D) and indicators of match quality to PDS. 3 - Processing at UCL (agreed previous iteration of this Dara Sharing Agreement and ongoing): d) NHS Digital to send file to UCL. 3A) The NDSCR-HES file provided by PHE (step 1C above) will be linked to the file provided by NHS Digital (step 2B above) in order to provide the UCL-specific HES-IDs for the study population in NDSCR-HES and enable linkage to an existing HES-ONS extract held by UCL (NIC-393510-D6H1D). 3 - Processing at UCL 3B) UCL researchers will categorise the resulting NDSCR-HES-ONS cohort according to underlying chronic conditions, major congenital anomalies and other comorbidities (based on clusters of diagnostic and procedure codes in HES), local authority, gender, maternal age at birth, deprivation quintile and ethnicity. a) UCL will use the file provided by NHS Digital to create an extract of cohort HES data from project NIC-393510-D6H1D for separate storage, access and processing by the researchers named on this (Down syndrome) project, and link this extract with the files provided by NHS Digital and Public Health England. 3C) UCL researchers will also undertake descriptive analyses to determine variation in unplanned admissions and health outcomes (e.g. mortality) by region of residence, adjusting for child characteristics in the cohort with Down syndrome compared with controls. a) UCL researchers will categorise the NDSCR-HES cohort according to underlying chronic conditions, major congenital anomalies and other comorbidities (based on clusters of diagnostic and procedure codes in HES), local authority, gender, maternal age at birth, deprivation quintile and ethnicity. b) UCL researchers will undertake descriptive analyses to determine variation in unplanned admissions and health outcomes (e.g. mortality) by region of residence, adjusting for child characteristics in the cohort with Down syndrome compared with controls. [5 paragraphs unchanged] AMENDMENT REQUEST 4 –Augment NDSCR-HES-ONS cohort with NPD data from an existing agreement (NIC-381972) UCL researchers request permission to augment the NDSCR-HES-ONS cohort (containing children with Down syndrome and controls), by linking pseudonymised NPD data from an existing agreement: NIC-381972-Q5F0V (“DHSC-ECHILD-COVID” study). 4A) UCL researchers request that the NDSCR-HES-ONS file be transferred to ONS SRS. 4B) In the ONS SRS, the NDSCR-HES-ONS file will be linked to a NPD-HES-ONS extract containing de-identified information from education and social care records from an existing agreement (NIC-381972-Q5F0V). The dataset in ONS SRS will be accessible to UCL researchers and on the relevant DSA with the Department for Education (DfE DSA approval panel number: DS00435; DfE Data Request reference number: DR200604.02B). 4C) Descriptive analyses will be carried out to assess the variation in type of school, special educational needs provision, access and utilisation of educational services, and ‘looked after child’ status in children with Down syndrome compared with the general population, and the association with subsequent unplanned hospital admissions. The de-identified linked HES-NPD attribute data will be held on the ONS SRS and will only be accessible remotely from the UCL Data Safe Room which has restricted and monitored access. No record level data can be removed from the ONS SRS and statistical disclosure controls are applied by ONS staff. Access will be restricted to named users, who are part of the study team and are accessing the data for the purposes outlined in this DSA. Access to the data is via the ONS SRS environment. Office of National Statistics (ONS) and UCL have signed and maintain an organisational agreement to use the ONS Secure Research Statistics (SRS) service for the purposes of secure statistical research, signed on 21/03/2019 with an indefinite expiry date. The only HES data stored in the ONS SRS are the 4 one-year cohorts of HES, plus the anonymised PMRs for those records that link to NPD. The HES data will include month of death and month of birth so no identifying data. For security and resource reasons the SRS is a Managed Service. Equiniti Ltd (based in Belfast) maintains the system, on behalf of the ONS SRS. They do so through encrypted (TLS1.2) VPN tunnel and Remotely Access (RA) the SRS. All Equiniti Ltd administrators are SC cleared and have no access to any data. ONS SRS Research Support “Admin” staff only have permissions to carry out such tasks as creating users, updating patches, testing and installing software applications, arranging DR, ITHC for the SRS environment, closing SRS sessions down, i.e. all the SRS environment Admin maintenance - essentially they are “power users”. There have been no data infractions by Equiniti Ltd staff in the last 5 years of them maintaining the environment, they have been very professional. The high level security document that Equiniti Ltd provided states: 6.3. Service Management support for the SRS Service is provided from Equiniti offices in Belfast, all staff are SC cleared. The office hosting the SRS Desk is IS0/IEC27001 2018 certified. Equiniti Ltd nor any their staff process the data. Therefore Equinity Ltd is not considered to be a Data Processor. The ONS SRS environment is an isolated system. It has no connectivity to the internet other than using it as a bearer to pass TLS1.2 encrypted image packages for a virtual desktop infrastructure (VDI), hosted on an accredited cloud server hosted by UKCloud Ltd on the mainland UK. UKCloud Ltd merely host the environment, they have no access to data. Therefore CloudUK Ltd is not considered to be a Data Processor. CLOUD SECURITY NHS Digital security has provided assurance regarding the use of the Office of National Statistics' Secure Research Statistics service (ONS SRS), hosted by CloudUK Ltd in this application. The Office of National Statistics has submitted a selection of security documentation to support the use of cloud storage. NHS Digital Security have reviewed the documentation and provided relevant feedback, where necessary. NHS Digital are satisfied that the documentation demonstrates the level of security and governance in place. The Office of National Statistics have supplied evidence to support: • The use of the Data Risk Model to assess the Risk Profile Class. • Risk Management of the use of the Cloud for this data, taking into consideration Confidentiality, Integrity and Availability. • The use of Pseudonymisation. • Board level involvement in the Risk Management Process evidenced through Minutes of these meetings. • Understanding of the Shared Responsibility Model The Office of National Statistics have a very good understanding of the security controls available to them to provide the appropriate controls to secure data in the Cloud. Using the Cloud, benefits from the inherited controls that cannot practically be replicated locally such as Physical Controls, Resilience of Systems, Power Supplies, Communications and Geographically dispersed Data Centres within a region. Elasticity in provisioning is also a consideration that benefits organisations in managing workloads. The Cloud provider, CloudUK, will use UK Data Centres only. The IT System provider based in Belfast will not handle any data.

Expected output

The first output of this study is to create a dataset comprising [31 words unchanged] children with Down syndrome vs the general population, over time and by region. Findings will be disseminated through peer-reviewed academic journals, conferences and social media. Findings will also be shared with region (as agreed in the Down's Syndrome Association. The researchers will specifically target a number of journals and conferences. These include the PLoS One, BMJ Open, Social Science and Medicine, Journal of Epidemiology and Community Health, Journal of Public Health, Archives of disease in childhood and paediatrics, Public Health Science Conference, and the International Conference on Congenital Anomalies and Pathology. initial request). The researchers will publish a series of methodological papers in peer reviewed journals reviewing the linkage and validating the data from the two data sources. The targeted journals will include the International Journal of Epidemiology, and PLoS One. Findings will be disseminated through peer-reviewed academic journals, conferences and social media. Findings will also be shared with the Down's Syndrome Association. The researchers will specifically target a number of journals and conferences. These include the PLoS One, BMJ Open, Social Science and Medicine, Journal of Epidemiology and Community Health, Journal of Public Health, Archives of disease in childhood and paediatrics, Public Health Science Conference, and the International Conference on Congenital Anomalies and Pathology. The researchers will publish a series of methodological papers in peer reviewed journals reviewing the linkages and validating the data from the different data sources. The targeted journals will include the International Journal of Epidemiology, and PLoS One. [1 paragraph unchanged] ~ Health outcomes (as indicated by health service use and mortality) for [6 words unchanged] these have changed over time and how they compare with the general population population. ~ Methodological issues in the construction of the linked dataset. datasets. The first of these reports covering each topic are expected to be [10 words unchanged] detailed analysis and reporting to be conducted over the next three years. Methodological outputs are expected to be lead directly to translation into practice at Public Health England. Findings about the health outcomes will be shared with stakeholder groups such as the Down's Syndrome Association, to support dissemination to patients and to receive feedback to guide the design of future research applications for the linked data (subject to approvals). Findings will also be disseminated directly through seminars at Public Health England, which help to develop plans for translating this demonstration project into the establishment of routine ongoing linkage systems for population health monitoring, both for Down syndrome and other disease registers maintained at Public Health England. Findings will also be presented to clinicians at meetings organised by the North Thames CLARHC, GOSH and the Down’s Syndrome Medical Interest group who disseminate findings across the UK https://www.dsmig.org.uk/. As funding for ongoing work will be through the CPRU findings will be disseminated to policy makers at DH and through the CPRU network of stakeholders in children’s health. Methodological outputs are expected to lead directly to translation into practice at Public Health England. Findings about the health outcomes will be shared with stakeholder groups such as the Down's Syndrome Association, to support dissemination to patients and to receive feedback to guide the design of future research applications for the linked data (subject to approvals). All outputs will contain aggregate level data only and all small numbers will be supressed in compliance with ADRC-E statistical output controls and the HES analysis guides. No potentially disclosive outputs will be shared or published. The projects in this application are expected to finish three years after obtaining the data. Findings will also be disseminated directly through seminars at Public Health England, which help to develop plans for translating this demonstration project into the establishment of routine ongoing linkage systems for population health monitoring, both for Down syndrome and other disease registers maintained at Public Health England. Findings will also be presented to clinicians at meetings organised by the North Thames Collaboration for Leadership in Applied Health Research and Care (CLARHC), Great Ormond Street Hospital (GOSH) and the Down Syndrome Medical Interest group who disseminate findings across the UK https://www.dsmig.org.uk/. As funding for ongoing work will be through the CPRU, findings will be disseminated to policy makers at Department of Health and Social Care (DHSC) and through the CPRU network of stakeholders in children’s health. All outputs will contain aggregate level data only and all small numbers will be supressed in line with the HES analysis guides. The projects in this application are expected to finish three years after obtaining the data. AMENDMENT REQUEST The second output, which is part of this amendment, is to use a subset of linked NPD-HES-ONS data from project NIC-381972-Q5F0V (“DHSC-ECHILD-COVID” study) to identify a cohort of children with and without Down syndrome and assess/compare the variation in the provision of educational and social care services, the association with health outcomes and the use of healthcare services such as unplanned hospital admissions. The methods used to link the health records and education records of children with Down syndrome, and their accuracy, will also be assessed. Additional output due to amendment: ~ Educational achievement and service provision outcomes (e.g. access and utilisation of special educational needs services) for people with Down syndrome compared with the general population, and how these are associated with use of health services (e.g. unplanned hospital admissions and health outcomes).

Expected measurable benefits

[1 paragraph unchanged] Results will also be useful for the development of policies and strategies [31 words unchanged] the data providers as the assessment of data quality and consistency between sources sources, and the development of optimal strategies for linking large administrative datasets will be an integral part of the study. [2 paragraphs unchanged] ~ A feasible method for routine linkage of disease registers to hospital episode statistics and mortality data at Public Health England will be established. A routine method for further linkages to education data will also be established. [1 paragraph unchanged] Public and Patient Information: During the design of the study, UCL have actively involved three separate parent support groups (Down Syndrome Association, Contact a Family, Council for Disabled Children). Discussions with the heads of research leads of these organisations about the project plan were met with unequivocal support for linkage between the registry, healthcare and NPD data. The problem of variation in service provision by local authority and over time was well recognised. All organisations expressed the view that, in their experience, parents of children Down syndrome would welcome linkage between these datasets for research. Parents already experience extensive transfer of their identifiers between multiple service providers. They recognise the potential benefits of research that investigates the interrelationship between services and would not view use of pseudo-anonymised linked data as a threat to their privacy. Following study completion, UCL will engage with the support groups for parents of children with Down syndrome to gain their perspective on the interpretations of the data during analysis. UCL have well-established mechanisms for patient and public involvement through the CPRU (Children, Young People and Families Policy Research Unit funded by the Department of Health Policy Research Programme). AMENDMENT REQUEST The second stage of the study, covered in this amendment, aims to link HES-ONS data for individuals with Down Syndrome to National Pupil Database data and evaluate the accuracy of linkage. The initial benefit will be to act as an exemplar for future studies aiming to link cohorts of children with rare, complex diseases to education and social care data. Comparing the provision of educational and social care services in children with Down syndrome to those without, and assessing the interplay with health outcomes, will inform policy-makers, advocacy groups, clinicians and families of areas where further support may need to be targeted in order to improve outcomes for children and young people with Down syndrome.

Benefits reported

Data has not yet been disseminated so no benefits have yet been realised. Results from initial analyses, assessing the quality of stand-alone and linked HES and NDSCR, have already been published (Doidge et al, 2020 - Prevalence of Down's Syndrome in England, 1998–2013 - https://doi.org/10.23889/ijpds.v5i1.1157). A key message of this work was that HES and NCDSCR data both indicated an increasing prevalence of Down’s syndrome among live births in England, but linked data with quantitative bias analyses provided more robust estimation and demonstrated that prevalence was stable. This initial work demonstrated the value of routine linkage of administrative and registry data, and is being used by co-authors from Public Health England. The main findings so far allude to the quality of data linkage rather than any expected benefits.