NHS Digital Data Release Register - reformatted

Cardiff University projects

233 data files in total were disseminated unsafely (information about files used safely is missing for TRE/"system access" projects).

🚩 Cardiff University was sent multiple files from the same dataset, in the same month, both with optouts respected and with optouts ignored. Cardiff University may not have compared the two files, but the identifiers are consistent between datasets, and outside of a good TRE NHS Digital can not know what recipients actually do.

STEADFAST Modelling the associations between wider health and social characteristics and diabetes-related health - Identifier data flows to DfE for linkage in ONS — DARS-NIC-674735-Z0H6K

Type of data: information not disclosed for TRE projects

Opt outs honoured: Identifiable, Yes (Section 251 NHS Act 2006)

Legal basis: Health and Social Care Act 2012 - s261(5)(d); National Health Service Act 2006 - s251 - 'Control of patient information'.

Purposes: No (Academic)

Sensitive: Non-Sensitive

When:DSA runs 2022-12-23 — 2023-12-22 2023.08 — 2023.08.

Access method: One-Off

Data-controller type: CARDIFF UNIVERSITY

Sublicensing allowed: No

Datasets:

  1. Demographics

Objectives:

On 1 February 2023, NHS Digital merged with NHS England. NHS England has assumed responsibility for all activities previously undertaken by NHS Digital. The merger was completed by a statute change. Any reference made to NHS Digital within this Data Sharing Agreement is in reference to the merged organisation known as NHS England.

Cardiff University requires NHS Digital data for the research study: “STEADFAST - The personal cost of health conditions in childhood”.

Cardiff University’s overall aim is to quantify the links between educational outcomes and diabetes-related health outcomes, for example, how educational settings influence blood glucose levels and the time to onset of diabetes-related complications. This linkage of diabetes and education data for England and Wales arose from work by Cardiff University modelling whether rates of hospital admissions for young people living with diabetes were higher than for their peers without diabetes.

The purpose of this study is to provide a greater understanding of the interrelationship between diabetes-related health and education. The substantive motivation for this study is grounded in the evidence that most of the costs arising from diabetes come not from the day-to-day care and medications, but the complications arising from elevated blood glucose over the life course. There is currently limited evidence on the causes of less optimal diabetes management and the potential mechanisms for interventions to improve this. Thus, this study hopes to broaden the evidence base beyond the purely clinical factors to investigate the wider health and social determinants of diabetes-related health, using linked administrative data and focussing on education.

Children of school and university age with diabetes are most frequently living with type 1 diabetes, accounting for approximately 98% of cases, so the research initially focuses on this population. Education and health outcomes for children with type 2 diabetes (and other rarer forms of diabetes) are equally important however, and the analysis is replicated for each of these groups.

Under linked Data Sharing Agreement (DSA) DARS-NIC-158283-T2Q2D versions 0-2, diabetes health data covering England and Wales was shared for people born from the 1983-1984 academic year to the 2009-2010 academic year. Data linkage of the diabetes health data with school and university records was completed for a subset of 2080 people who studied at a Welsh educational institution (school, university, college). Diabetes health data included those linked from the national paediatric diabetes audit data (controlled by the Healthcare Quality Improvement Partnership (HQIP)) for individuals born 1992-1993 to 2009-2010, and the adult national diabetes audit data (from NHS Digital) for individuals born 1983-1984 to 2001-2002.

This Agreement, DARS-NIC-674735-Z0H6K-v0, is a new Agreement to request the equivalent diabetes health and education data for people living in England and Wales. This Agreement should be read in parallel with DARS-NIC-669808-V6T0M-v0 and DARS-NIC-669962-W1F6D-v0. Each DSA covers a different element of the required data flows for this study. Linked agreement DARS-NIC-158283-T2Q2D-v2 is being retained to cover the historic data flows described above, and may be amended in future to request further diabetes audit years to be linked with Welsh educational records.

The substantive dataset requested from NHS Digital is the National Diabetes Audit (NDA) which provides information regarding diabetes-related health for all people with diabetes. NDA data for linkage in the Secure Anonymised Information Linkage (SAIL) databank was disseminated and processed under linked agreement DARS-NIC-158283-T2Q2D versions 0-2. Data was disseminated to SAIL databank at the University of Swansea with a pseudonymised identifier only.

NDA data for people living in England and Wales is requested under linked agreement DARS-NIC-669962-W1F6D-v0. The NDA data that is required relate to characteristics of the diabetes diagnosis (diabetes type, age at diagnosis) and measures of diabetes-related health (e.g. blood glucose levels, levels of protein in urine), and care processes (e.g. retinopathy screening for damaged blood vessels in the eye; foot exams to assess nerve or blood vessel damage). Diabetes-related health is primarily measured using blood glucose levels, directly or by using ‘HbA1c’, a proxy for blood glucose management control measured regularly as part of diabetes clinical management and recorded in the diabetes audit data. This data will be disseminated to the Office for National Statistics (ONS) – Secure Research Service (SRS) with a pseudonymised identifier only.

Cardiff University are also requesting that NHS Digital provide identifiable information of relevant individuals in the NDA to the organisations who hold their educational records.

Under linked agreement DARS-NIC-158283-T2Q2D versions 0-2, identifiable information for people with diabetes living in England and Wales was disseminated to Digital Health and Care Wales (DHCW), in order that the relevant primary and secondary education (provided to DHCW from the Welsh Government), and higher education (provided to DHCW from the Higher Education Statistics Agency (HESA)) records could be identified. The identifiers were provided alongside the same pseudonymised identifier provided to SAIL databank with the NDA data. The education data was made available for analysis by Cardiff University in a de-identified format in the SAIL databank, linked to the applicable diabetes health record via the pseudonymised identifier.

Under this Agreement, DARS-NIC-674735-Z0H6K-v0, NHS Digital are requested to provide identifiable information of relevant individuals in the NDA who were living in England and Wales to the Department for Education (DfE) in order that their primary, secondary and higher education records can be identified. The identifiers will be provided alongside the same pseudonymised identifier provided to the ONS-SRS with the NDA data. The education data will be made available for analysis by Cardiff University in a de-identified format in the ONS-SRS, linked to the applicable diabetes health record via the pseudonymised identifier.

Under linked agreement DARS-NIC-669808-V6T0M-v0, NHS Digital are requested to supplement the identifiable information of individuals in the National Paediatric Diabetes Audit (NPDA), provided by the Royal College of Paediatric and Child Health (RCPCH), with their names. This would be done using NHS Digital’s ‘Demographics’ dataset. Names can then be used by the DfE alongside other identifiers supplied by the RCPCH to retrieve the relevant educational records of individuals in the NPDA. These would be made available for analysis by Cardiff University in a de-identified format, linked to the applicable diabetes health record.

The NDA and identifiable data from NHS Digital will, when combined with the associated education data, allow the study to model how characteristics of a person’s diabetes impact their education and, simultaneously, how their education affects their diabetes-related health.

Educational outcomes are recorded from the time a child enters school until they leave university, including measures of attendance, attainment, and broader characteristics of the educational experience such as special educational needs and school exclusions.

The mechanisms driving the relationships between education and health can be broken down into three pathways. Firstly, diabetes-related health may affect educational outcomes, for example, through biological mechanisms, including the effects of excess glucose on the brain structure and social mechanisms such as adjusting management routines to fit in with a university lifestyle. Secondly, education may affect diabetes-related health. For example, continuing education beyond compulsory schooling might provide structure and support that facilitate better management. Thirdly, individual characteristics (observed and unobserved) may directly affect both education and diabetes-related health, such as motivation and intelligence.

To help unpick which of these are happening in the data, Cardiff University uses repeated measures of educational outcomes and health outcomes for an individual to tease out the ordering of events. For example, if a person with less optimal blood glucose levels has high rates of school absence, it would be possible to look back and see if that person had high rates of absence before they were diagnosed with diabetes. In practice, the statistician will look at many thousands of individuals at once and consider many such differences simultaneously, but the principle is the same.

The data request covers people born from 01/09/1983–31/08/2002 with diabetes, living in England and Wales, who are included in the diabetes audits (NDA and/ or NPDA) from 2003 onwards. This is anticipated to include 30,000 patient records, with linkage anticipated for 28,000 individuals. Individuals born prior to this have been excluded since their full education data, including data on their first year of university, is not available. Individuals born after 2002 will be recorded in the NPDA, which is not provided by NHS Digital. As the study is primarily interested in the effect of diabetes during education, cases are restricted to those diagnosed with any form of diabetes prior to age 24 or younger. Whilst university cohorts are typically 18-21, due to the high prevalence of delays in starting time at university (gap year, changing university, changing course etc.) and the varying length of courses (sandwich degrees, placements, conversion courses), the request will include all ages up to the government definition of the end of youth education, i.e. up to age 24.

The predicted ages of school students include early years (nursery and reception) aged 3-5, compulsory schooling aged 5-16, and key stage 5 ages 16-18. Some students start school early or leave later than normal, however there will be no request for additional cohorts for this contingency and will instead make this an amendment if it transpires to be an issue.

The DfE do not hold data for children educated at private schools or educated home, thus for these cases (~7%) the diabetes audit data would not be assigned a linkage ID by DfE. Unlinked cases would not form part of the core analysis, however Cardiff University would carefully examine why different cases did not link. For example, there may be noticeable patterns in children who did not link for technical linkage reasons, such as lower matching rates for children from certain minority ethnic groups where the matching algorithm may not be as efficient for those names. Cardiff University would also consider if the diabetes health outcomes are different for those children who do not link, since that may motivate a change in monitoring, for example for children who are educated in settings other than at school (home educated).

There should be no NDA cases in the requested extract that will not meet the conditions of the study. It would be expected all cases to have attended school, and those that do not attend university are still important as controls for comparison with those that did. There will be no request for controls from NHS Digital.

Although linkage with Welsh educational data was undertaken under linked agreement DARS-NIC-158283-T2Q2D versions 0-2, data is requested for individuals resident in both England and Wales for the current request because individuals can be located across borders for their medical care or educational settings, and individuals move across borders over time. The study will combine estimates from data linked to Welsh educational records in the SAIL databank, and data linked to English educational records in the ONS-SRS, to create a single estimate for England and Wales.

Cardiff University request individual-level data for modelling changes over time for individuals, e.g. how individual differences in education experiences affect diabetes management and vice versa. Cardiff University requests identifiers only for the purpose of data linkage; however, the analysis will be carried out on de-identified data.
Cardiff University request data from 2003/04 to 2017/18 to best model trajectories of HbA1c and the time to the first onset of early complications.
Cardiff University requests data for England and Wales to ensure maximum power and generalisability of the results.

Cardiff University confirm there are no alternative, less intrusive ways of achieving the purpose. A high degree of data linkage is required for this study in order to follow the entire life course trajectories of both diabetes related health and education through the combined linkage of both the paediatric and adult diabetes audits, and education data from primary school to university. Patient identifiers are sent to different organisations than the substantive clinical data to minimise the risk of disclosure.

The request has been restricted to only the essential clinical measures and associated metadata such as date of measure, location of measure (clinic).
Cardiff University are the research Sponsor and sole data controller for this DSA.

The personal data for this study is processed under UK GDPR Article 6(1)(e) - Public Task - for academic medical research carried out as a task in the public interest. The processing is necessary for Cardiff University (as a public authority for the purposes of data protection legislation) to perform a task in the public interest. The task has a clear basis in law.

The special category personal data for this study is processed under UK GDPR Article 9(2)(j) 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 right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject. The public interest lies in the improved evidence on the links between educational measures (school settings, absence, attainment) and health measures (HbA1c levels) for young people with diabetes. The study aims to improve care for all young people with diabetes in the education system by informing clinicians and commissioners of variation and outcomes and complications to support work to improve and standardise treatment selection choices. For these reasons, the processing also meets the conditions of Schedule 1 Part 1 paragraph 4 of the Data Protection Act 2018.

Cardiff University has approval from the Health Research Authority for data providers to set aside the common law duty of confidentiality with respect to the sharing of person identifiers in health datasets, known as a ‘Section 251 exemption’.

SAIL and DHCW are data processors for the Welsh subsets of data as described above. The DfE and ONS are data processors for the English subsets of data as described above.

The Welsh Government and HESA provide education data which is linked with NHS Digital data.

The RCPCH provide NPDA identifiable data to NHS Digital, and NPDA substantive data to the ONS-SRS. The Healthcare Quality Improvement Partnership (HQIP) are the data controller of the NPDA.

The Welsh Government, HESA, RCPCH and HQIP do not process NHS Digital data.

None of the organisations listed above, except Cardiff University, determine the purposes or the means of the processing of NHS Digital data. They are not therefore considered joint data controllers of NHS Digital data under this DSA.

The Medical Research Council (MRC) and UK Research & Innovation (UKRI) have provided funding for this study. The funders do not determine the purposes or the means of the processing of NHS Digital data and do not process the data. They are not therefore considered joint data controllers or data processors of NHS Digital data under this DSA.

Yielded Benefits:

The study has been lucky in terms of the timings of the opportunities to contribute to the policy debate. Cardiff University were able to provide evidence, based on analysis of the Welsh data, to the new Additional Learning Needs Bill for Wales and the Special Educational Needs and Disabilities Green Paper in England, though these are yet to report their findings so have not yet resulted in any change in practice. To support public involvement and engagement, the study team has given presentations of the results obtained so far to practitioners (e.g., the Royal College of Paediatrics and Child Health annual conference, the Brecon Annual Meeting of type 1 diabetes practitioners in Wales), patients (e.g., the Yorkshire and Humber Young People’s Diabetes Network Annual Meeting), and researchers (The Type 1 Diabetes Consortium at the Diabetes UK conference). The findings have been well received by clinicians, who report that they feed back some of the findings, particularly the ‘good news story’ that overall, children with diabetes attain in school as well as their peers without diabetes. Practitioners have also reported that the quantification of expected absence (average and range) has also been useful to provide a benchmark for children with diabetes.

Expected Benefits:

Although laws exist in England and Wales to support people with health conditions in schools and universities, the legislation (and associated guidance and resulting practice) does not always meet the child's needs. In schools, there is a gap between support for special needs (which stems from the Children and Families Act 2014 and traditionally focused on support for ‘learning conditions’, e.g. dyslexia) and support for children with health conditions (which stems from the Equalities Act 2010 and traditionally focuses on support through reasonable adjustments). In universities, students lack clarity on how financial support for health conditions works, particularly for conditions like diabetes, where there is wide variation in the required amount and type of support. Students will also be adjusting to having less clinical support than they are accustomed to, having only recently transitioned from paediatric to adult services before moving to a university campus far from their ‘home’ clinical team.

The high-level goal of this study is to ensure that children living with diabetes (and their families) can (i) manage their condition to stay healthy and (ii) fulfil their potential in education. By combining health data (including HbA1c) and education data (including educational outcomes and support), the project hopes to generate robust evidence of variation in outcomes, which may motivate changes in legislation and practice. Although the primary route for an impact on health is through policy change (and the associated improvement in guidance and implementation); the project will continue to inform practitioners of the range of outcomes and thus show the scope and potential benefit from improved support, both in terms of educational outcomes and diabetes-related health outcomes, and the link between the two.

Study outcomes for children in schools: This will focus on how rates of absence and attainment trends are associated with diabetes management. The longitudinal (annual measures) nature of the data means that researchers can compare outcomes over time; for example, annual measures of school absence can help show the changes in missed schooling before vs after a diagnosis of diabetes, in addition to comparing absence rates with similar children who were not diagnosed with diabetes. Similarly, regular measurements of HbA1c allow researchers to identify changes in blood glucose management as a child moves from primary school to secondary school. If differences exist, clinical teams may choose to give additional support based on non-clinical (educational) challenges that may influence diabetes management.

Study outcomes for children leaving school: Students with diabetes who are leaving compulsory schooling will transition to A-levels, vocational courses, or the labour force, and the study will model how these choices subsequently impact the trajectories of HbA1c. For example, the study may find that those who choose not to continue formal education beyond age 16 have less structure, affecting their diabetes management. If differences exist, clinical teams may choose to provide extra support or delay the transition from paediatric to adult diabetes care until after the transition to employment is complete.

Study outcomes for higher education students: The work focusing on universities will look for changes in trajectories of diabetes management, for example, comparing those that live at home with those who move away to attend university, identifying whether there is a significant change in management and the timings of that change. Differences may be seen in HbA1c during the first term as students try to reconcile good management routines with university social life, or perhaps HbA1c may not alter until the stress of the final year of university with impending high-stakes exams. If differences exist, clinical teams may choose to be more proactive in supporting the transition between home and university care and arrange follow-up appointments, both at home and at university, scheduled at times that fit around academic term dates.

Study outcomes for young adults: The research on young adults focuses on how childhood HbA1c levels and educational outcomes determine early adult health outcomes. This study focuses on the ‘double whammy’ effects of a history of sub-optimal HbA1c levels and lower educational outcomes. Relevant determinants might include individual factors (e.g. special educational needs or behavioural issues), school factors (e.g. lack of appropriate support), or family-level factors (e.g. socioeconomic status). If differences do exist, clinical teams may choose to flag individuals earlier who are struggling with diabetes management and education as being particularly vulnerable and receive additional support in the adult care setting, with potential health and social benefits to the individuals (who are most likely to develop early diabetes-related complications).

In partnership with Diabetes UK the project held a series of public workshops which reported the expected measurable benefits from the perspective of patient and public involvement participants (https://www.adruk.org/fileadmin/uploads/adruk/Documents/Diabetes_Education_Public_Workshop_Report_Aug_2021_01.pdf, and included: more understanding by others of what they are going through, more support with managing diabetes in schools, and earlier intervention at times of high stress. There was a focus on diabetes-education interactions during transition from paediatric to adult diabetes care – this comes amongst many other stressful events that teenagers go through, not least of which were the GCSE exams.

Participants focused on the benefit of bridging the gap in educational experiences between children with type 1 diabetes versus those without – “ensuring an equal playing field”. Participants reported a lack of knowledge around the impact of type 1 diabetes on education, which they felt was likely to be large. They felt that a study of this kind could influence policy, and ultimately the education setting, which in turn could result in a better experience for young people. A related point was the hope for increased general awareness and understanding of how type 1 diabetes might affect individuals beyond direct medical outcomes.

Outputs:

The expected outputs will be academic journal papers modelling the associations between education outcomes and trajectories of diabetes-related health.
The nature of data-based outputs would be descriptive statistics and regression coefficients, all of which would be aggregate data with small numbers (less than 10) suppressed; this is a requirement of taking any results out of the ONS-SRS and is rigorously checked by ONS analysts.

The original aim was to finish the draft of the outputs using data in SAIL disseminated under linked agreement DARS-NIC-158283-T2Q2D versions 0-2, between 21/01/2019 to 21/08/2020, though the first of these papers was only accepted for publication on 01/08/2022 by the journal ‘Diabetes Care’. It is likely that future analysis will combine estimates from data deposited in both SAIL databank and ONS-SRS, and is expected to be published before 2025.

Findings from the first paper have been presented at an invited address at the Royal College of Paediatric and Child Health annual conference and a presentation to clinicians at the Brecon Group Annual meeting.

The first research paper is based upon how measures of diabetes-related health influence educational trajectories, and shows that (i) children with diabetes have higher absence rates, but similar attainment and progression to university as their peers and (ii) HbA1c tracks with educational outcomes. Cardiff University provide tentative evidence that this link is likely to be the result of external factors affecting both education and health outcomes. Due to issues with the NPDA data, Cardiff University have not yet been able to undertake any analysis pertaining to how education settings affect HbA1c trajectories, or how external factors affect both diabetes-related health trajectories and educational trajectories.

Any further study findings will be shared with clinicians through academic publications (e.g., Diabetes Care) and presentations at national forums (e.g., Diabetes UK Conference & Brecon Scientific Meeting). The findings will be shared with the public more broadly using a plain English version of the study; this will hopefully be published in the diabetes audit report and presented by the project lead at patient forums. Further engagement work with the results will be coordinated with Cardiff University’s coproduction group of young people.

An agreement is in place with Royal College of Paediatric and Child Health to include a patient level report on associations between diabetes related health and education for their annual report. Updates and summaries will be published on the project website (https://www.cardiff.ac.uk/research/explore/research-units/childhood-health-and-education?).

A dissemination plan has been produced in partnership with Diabetes UK. This includes two papers supported by plain English reports produced by the Diabetes UK policy team. The research team and Diabetes UK have begun to meet with policymakers from England and Wales, but it is too early to say how these connections will be exploited to communicate results. All outputs will belong to Cardiff University but will be freely available to Diabetes UK to use as agreed under Cardiff University’s heads of terms. Information about how the study has already consulted young people are provided on the webpage for the first 2021 patient and public engagement workshop: https://www.adruk.org/our-work/the-personal-cost-of-health-conditions-in-childhood-engaging-the-public/public – ADR UK and the workshop report: https://www.adruk.org/fileadmin/uploads/adruk/Documents/Diabetes_Education_Public_Workshop_Report_Aug_2021_01.pdf

In the second workshop, the young people produced a video explaining how researchers use data relating to young people with diabetes. Although the video and report produced are not yet released, there is a holding version on YouTube: https://www.youtube.com/watch?v=oimfnSoENxo.

Several of the representatives from Cardiff University’s patient and public involvement sessions have become the coproduction group, helping steer the project and the next iteration of the public engagement.

Processing:

This study uses a split file process to transfer individual-level data. The objective is to ensure that, as far as possible, every person’s health data is processed separately from the identifiers that would link that health data to a person.

Each clinical data provider assigns a study-specific pseudonymised identifier (study ID) to each participant. They then split the whole dataset into an identifiers dataset (containing variables such as NHS number, name, date of birth, postcode, and gender) and a substantive dataset (containing de-identified clinical/ education data).
The substantive data is transferred directly to the repository.

The identifiers file is shared with a trusted third party (details below) who uses the identifiers to match individuals’ ‘study ID’ (pseudonymised identifiers) to their ‘linkage ID’, and then deletes all the real-world identifiers such as names and dates of birth, before transferring the ‘study ID’ and ‘linkage ID’ into the repository where the ‘study ID’ enables re-joining of the ‘linkage ID’ to the substantive data, and the ‘linkage ID’ enables linkage to the other datasets which have been processed in the same way.

Under linked agreement DARS-NIC-158283-T2Q2D versions 0-2, the data flows were as follows:

NHS Digital identified people born from 01/09/1983 to 31/08/2002 (ie the 1983/84 to 2001/02 academic years) who appeared in the National Diabetes Audit (NDA) dataset and assigned them a unique study ID.

From this cohort, NHS Digital provided a one-off drop of data covering 2003/04 to 2017/18 audit years, containing the substantive diabetes-related data, accompanied by study ID only, to the Secure Anonymised Information Linkage (SAIL) databank at the University of Swansea. This file contained no identifying information other than the study ID.

From this cohort, NHS Digital also provided a one-off drop of data covering 2003/04 to 2017/18 ‘audit years,’ containing study ID, patient name, NHS number, date of birth, gender, and postcode only to Digital Health and Care Wales (DHCW).

DHCW created the ‘linkage ID’, a hashed version of the NHS Number, referred to by DHCW and SAIL as the Anonymised Linkage Field (ALF), then destroyed all the identifying information (name, NHS number, date of birth, gender, postcode). After this process, the identifiers file contains only the study ID and linkage ID, along with less disclosive versions of the demographic data (gender, week of birth, lower super output area). This file was onwardly flowed to SAIL.

Within SAIL, the substantive NDA file was re-joined to the identifiers linkage field file using the study ID so that each case of the substantive data got a linkage ID.
The same process as outlined above for NDA data also happened for the other de-identified datasets that were placed in SAIL, including the National Paediatric Diabetes Audit (NPDA), Higher Educational Statistics Agency dataset (HESA), and Welsh Dataset Education Records (WED, also known as the National Pupil Database or NPD). The NPDA, HESA, and WED data were linked to the NHS Digital NDA data using the linkage ID.

Data accessible to Cardiff University via SAIL therefore are:
- NDA data covering England and Wales
- NPDA data covering England and Wales
- HESA (higher education) data covering England and Wales
- Primary and secondary education data for Wales only (WED)

Under this Agreement, DARS-NIC-674735-Z0H6K-v0.0, the data flows are as follows:

NHS Digital identify individuals based in England and Wales and born between the 01/09/1983 to 31/08/2022 (ie the 1983/84 to 2001/02 academic years), in the NDA dataset under DARS-NIC-669962-W1F6D and assign them a unique study ID.

Using the NHS numbers held in the NDA, NHS Digital retrieves the names, date of birth, latest postcode, and gender from the Personal Demographics Service (NHS number was included for DHCW, but is not used here as Department for Education (DfE) are unable to process it).

These identifiers are sent to the DfE alongside the unique study ID. This will be a one-off deposit of data, though permission for annual refreshes of data will be sought in the future. DfE, acting as the trusted third party, will use the identifiers to match individuals’ ‘study ID’ to the ‘linkage ID’ (referred to by DfE and the Office for National Statistics (ONS) as the Pupil Matching Reference (PMR)), then destroy all the identifying information (name, date of birth, gender, postcode). After this process, the identifiers file contains only the study ID and linkage ID. (Unlike for the SAIL process, less disclosive versions of the demographic data [gender, week of birth, lower super output area], are not retained in the identifiers file). This modified identifiers file, containing only study ID and linkage ID is onwardly flowed to ONS-SRS repository where the ‘study ID’ enables re-joining of the ‘linkage ID’ to the substantive data, and the ‘linkage ID’ enables linkage to the other datasets which have been processed in the same way.

In parallel, under linked agreement DARS-NIC-669962-W1F6D-v0, substantive deidentified NDA health data for the 2003/4 to 2017/18 audit years for the above-described cohort will be sent to the ONS Secure Research Service (SRS), with the unique study ID only. This will be a one-off deposit of data, though permission for annual data refreshes will be sought in the future.

Under linked agreement DARS-NIC-669808-V6T0M-v0, NHS number, date of birth, postcode and gender will be provided to NHS Digital by the Royal College of Paediatric and Child Health (RCPCH) for individuals born in the academic birth cohorts from 1983/4 to 2001/2 who are in the NPDA. NHS Digital will retrieve the names of these individuals and send identifiable Demographics data (names, date of birth, latest postcode, and gender) to the DfE alongside a unique pseudonymised identifier provided by the RCPCH. This will be a one-off deposit of data, though permission for annual data refreshes will be sought in the future.

The project includes four further data flows which are not covered by the above-listed DSAs:

1. RCPCH send substantive deidentified NPDA data directly to the ONS-SRS, containing the same unique pseudonymised identifier as provided to NHS Digital under DARS-NIC-669808-V6T0M-v0
2. DfE send substantive deidentified education records (compulsory education) directly to the ONS-SRS, alongside the linkage ID
3. HESA send substantive deidentified education records (higher education) directly to the ONS-SRS, alongside the linkage ID.
4. After the DfE ‘hashing service’ use the identifiers supplied by NHS Digital to identify the correct individual against their records and retrieve their PMR, the original study ID supplied by NHS Digital, alongside the PMR retrieved by the DfE, are onwardly flowed into the ONS-SRS.

DfE will destroy the real-world identifiers as soon as the PMR field is retrieved and the necessary pseudonymised identifiers provided to the ONS-SRS. They will provide NHS Digital with a data destruction certificate.

ONS-SRS staff will work with the Cardiff University project lead to link the data records within the ONS-SRS environment. Substantive NDA and NPDA data will be linked with a PMR via the original study IDs, where the DfE have been able to retrieve a PMR.

NDA, NPDA, and education data will then be linked using the PMRs.

ONS-SRS staff will support Cardiff University with the linkage evaluation for the extract to be used for research, determining what percentage of clinical records have been successfully linked together (NPDA to NDA) and linked to associated education records (NPDA to NPD, NPDA to HESA, NDA to NPD, NDA to HESA).

Unlinked clinical data will remain in the ONS-SRS environment but will not be further processed for the purposes of these Data Sharing Agreements (DARS-NIC-674735-Z0H6K-v0.0, DARS-NIC-669808-V6T0M-v0, and DARS-NIC-669962-W1F6D-v0). Other linkage mechanisms may be explored in future to improve the success of the linkage, and data on individuals not in school or university may give rise to important information for diabetes management.

The linked de-identified data-sets will be made available for multilevel modelling analysis of the associations between education and health by Cardiff University on the ONS-SRS. The DfE will provide educational data for controls who do not have diabetes, alongside the educational data for those with diabetes, to enable Cardiff University to model the differences in outcomes.

The combining of health data with education data increases the likelihood that a person may appear as unique in the dataset; however, given that there are 30,000+ cases of young people with diabetes in England and Wales, and the high-level nature of the variables, it is unlikely that this dramatically increases the risk of re-identification. There will be no requirement or attempt to re-identify individuals for the purposes of the study or any other reason. The primary protection against reidentification is the creation of the linkage ID along with the destruction of the real-world identifiers prior to the data being put in the repository, meaning the data accessed by Cardiff University is de-identified. Secondly, ONS-SRS and Cardiff University will check data at the outset for any identifiability risks in the raw data or linked data before using it for analysis. This checking is routinely conducted by experienced analysts at ONS who use a mature process to identify small numbers, unique/rare cases, or other identifiability risks. Thirdly, researchers from Cardiff University who are accessing the data are trained not to re-identify data.
NHS Digital data is not being linked to any publicly available data.

Data processing is only conducted by substantive employees of the data processors who have specific teams and staff trained and explicitly employed for this purpose. Data will be accessed for analysis through secure remote gateways into the ONS-SRS. No data will be accessed outside the UK. Person identifiers (such as name and date of birth) will not be shared with third parties beyond those required to create the pseudonymised linked datasets (i.e. only DfE).

Public engagement focused on both the research questions and the data processing has been carried out throughout the project, going back to 2014 before the study was funded. The largest public involvement work has been the sessions with young people with type 1 diabetes in 2020, in partnership with Diabetes UK and the MRC Regulatory Support Centre. This work is summarised in a website (https://www.adruk.org/news-publications/news-blogs/public-views-on-the-use-of-personal-identifiers-for-linking-diabetes-and-education-data-for-research-439/) with links to the full report towards the bottom of the page. Further public engagement was carried out in 2022 across 19 focus groups, the link to the work is https://dareuk.org.uk/sprint-exemplar-project-steadfast, and the report summarising this work should be released in early 2023.

STEADFAST Modelling the associations between wider health and social characteristics and diabetes-related health - Cohort Matching Service for NPDA identifier data flows to DfE for linkage in ONS — DARS-NIC-669808-V6T0M

Type of data: information not disclosed for TRE projects

Opt outs honoured: Identifiable (Section 251 NHS Act 2006)

Legal basis: Health and Social Care Act 2012 - s261(5)(d); National Health Service Act 2006 - s251 - 'Control of patient information'.

Purposes: No (Academic)

Sensitive: Non-Sensitive

When:DSA runs 2022-12-23 — 2023-12-22

Access method: One-Off

Data-controller type: CARDIFF UNIVERSITY

Sublicensing allowed: No

Datasets:

  1. Demographics

Objectives:

On 1 February 2023, NHS Digital merged with NHS England. NHS England has assumed responsibility for all activities previously undertaken by NHS Digital. The merger was completed by a statute change. Any reference made to NHS Digital within this Data Sharing Agreement is in reference to the merged organisation known as NHS England.

Cardiff University requires NHS Digital data for the research study: “STEADFAST - The personal cost of health conditions in childhood”.

Cardiff University’s overall aim is to quantify the links between educational outcomes and diabetes-related health outcomes, for example, how educational settings influence blood glucose levels and the time to onset of diabetes-related complications. This linkage of diabetes and education data for England and Wales arose from work by Cardiff University modelling whether rates of hospital admissions for young people living with diabetes were higher than for their peers without diabetes.

The purpose of this study is to provide a greater understanding of the interrelationship between diabetes-related health and education. The substantive motivation for this study is grounded in the evidence that most of the costs arising from diabetes come not from the day-to-day care and medications, but the complications arising from elevated blood glucose over the life course. There is currently limited evidence on the causes of less optimal diabetes management and the potential mechanisms for interventions to improve this. Thus, this study hopes to broaden the evidence base beyond the purely clinical factors to investigate the wider health and social determinants of diabetes-related health, using linked administrative data and focussing on education.

Children of school and university age with diabetes are most frequently living with type 1 diabetes, accounting for approximately 98% of cases, so the research initially focuses on this population. Education and health outcomes for children with type 2 diabetes (and other rarer forms of diabetes) are equally important however, and the analysis is replicated for each of these groups.

Under a linked Data Sharing Agreement (DSA), DARS-NIC-158283-T2Q2D versions 0-2, diabetes health data covering England and Wales was shared for people born from the 1983-1984 academic year to the 2009-2010 academic year. Data linkage of the diabetes health data with school and university records was completed for a subset of 2080 people who studied at a Welsh educational institution (school, university, college). Diabetes health data included those linked from the national paediatric diabetes audit data (controlled by the Healthcare Quality Improvement Partnership (HQIP)) for individuals born 1992-1993 to 2009-2010, and the adult national diabetes audit data (from NHS Digital) for individuals born 1983-1984 to 2001-2002.

This Agreement, DARS-NIC-669808-V6T0M-v0, is a new Agreement to request the equivalent diabetes health and education data for people living in England and Wales. This Agreement should be read in parallel with DARS-NIC-674735-Z0H6K-v0 and DARS-NIC-669962-W1F6D-v0. Each DSA covers a different element of the required data flows for this study. DARS-NIC-158283-T2Q2D-v2 is being retained to cover the historic data flows described above, and may be amended in future to request further diabetes audit years to be linked with Welsh educational records.

The substantive dataset requested from NHS Digital is the National Diabetes Audit (NDA) which provides information regarding diabetes-related health for all people with diabetes. NDA data for linkage in the Secure Anonymised Information Linkage (SAIL) databank was disseminated and processed under DARS-NIC-158283-T2Q2D versions 0-2. Data was disseminated to SAIL databank at the University of Swansea with a pseudonymised identifier only.

NDA data for people living in England and Wales is requested under linked agreement DARS-NIC-669962-W1F6D-v0. The NDA data that is required relate to characteristics of the diabetes diagnosis (diabetes type, age at diagnosis) and measures of diabetes-related health (e.g. blood glucose levels, levels of protein in urine), and care processes (e.g. retinopathy screening for damaged blood vessels in the eye; foot exams to assess nerve or blood vessel damage). Diabetes-related health is primarily measured using blood glucose levels, directly or by using ‘HbA1c’, a proxy for blood glucose management control measured regularly as part of diabetes clinical management and recorded in the diabetes audit data. This data will be disseminated to the Office for National Statistics (ONS) – Secure Research Service (SRS) with a pseudonymised identifier only.

Cardiff University are also requesting that NHS Digital provide identifiable information of relevant individuals in the NDA to the organisations who hold their educational records.

Under linked agreement DARS-NIC-158283-T2Q2D versions 0-2, identifiable information for people with diabetes living in England and Wales was disseminated to Digital Health and Care Wales (DHCW), in order that the relevant primary and secondary education (provided to DHCW from the Welsh Government), and higher education (provided to DHCW from the Higher Education Statistics Agency (HESA)) records could be identified. The identifiers were provided alongside the same pseudonymised identifier provided to SAIL databank with the NDA data. The education data was made available for analysis by Cardiff University in a de-identified format in the SAIL databank, linked to the applicable diabetes health record via the pseudonymised identifier.

Under linked agreement DARS-NIC-674735-Z0H6K-v0, NHS Digital are requested to provide identifiable information of relevant individuals in the NDA who were living in England and Wales to the Department for Education (DfE) in order that their primary, secondary and higher education records can be identified. The identifiers will be provided alongside the same pseudonymised identifier provided to the ONS-SRS with the NDA data. The education data will be made available for analysis by Cardiff University in a de-identified format in the ONS-SRS, linked to the applicable diabetes health record via the pseudonymised identifier.

Under this Agreement, DARS-NIC-669808-V6T0M-v0, NHS Digital are requested to supplement the identifiable information of individuals in the National Paediatric Diabetes Audit (NPDA), provided by the Royal College of Paediatric and Child Health (RCPCH), with their names. This would be done using NHS Digital’s ‘Demographics’ dataset. Names can then be used by the DfE alongside other identifiers supplied by the RCPCH to retrieve the relevant educational records of individuals in the NPDA. These would be made available for analysis by Cardiff University in a de-identified format, linked to the applicable diabetes health record.

The NDA and identifiable data from NHS Digital will, when combined with the associated education data, allow the study to model how characteristics of a person’s diabetes impact their education and, simultaneously, how their education affects their diabetes-related health.

Educational outcomes are recorded from the time a child enters school until they leave university, including measures of attendance, attainment, and broader characteristics of the educational experience such as special educational needs and school exclusions.

The mechanisms driving the relationships between education and health can be broken down into three pathways. Firstly, diabetes-related health may affect educational outcomes, for example, through biological mechanisms, including the effects of excess glucose on the brain structure and social mechanisms such as adjusting management routines to fit in with a university lifestyle. Secondly, education may affect diabetes-related health. For example, continuing education beyond compulsory schooling might provide structure and support that facilitate better management. Thirdly, individual characteristics (observed and unobserved) may directly affect both education and diabetes-related health, such as motivation and intelligence.

To help unpick which of these are happening in the data, Cardiff University uses repeated measures of educational outcomes and health outcomes for an individual to tease out the ordering of events. For example, if a person with less optimal blood glucose levels has high rates of school absence, it would be possible to look back and see if that person had high rates of absence before they were diagnosed with diabetes. In practice, the statistician will look at many thousands of individuals at once and consider many such differences simultaneously, but the principle is the same.

The data request covers people born from 01/09/1983–31/08/2002 with diabetes, living in England and Wales, who are included in the diabetes audits (NDA and/ or NPDA) from 2003 onwards. This is anticipated to include 30,000 patient records, with linkage anticipated for 28,000 individuals. Individuals born prior to this have been excluded since their full education data, including data on their first year of university, is not available. Individuals born after 2002 will be recorded in the NPDA, which is not provided by NHS Digital. As the study is primarily interested in the effect of diabetes during education, cases are restricted to those diagnosed with any form of diabetes prior to age 24 or younger. Whilst university cohorts are typically 18-21, due to the high prevalence of delays in starting time at university (gap year, changing university, changing course etc.) and the varying length of courses (sandwich degrees, placements, conversion courses), the request will include all ages up to the government definition of the end of youth education, i.e. up to age 24.

The predicted ages of school students include early years (nursery and reception) aged 3-5, compulsory schooling aged 5-16, and key stage 5 ages 16-18. Some students start school early or leave later than normal, however there will be no request for additional cohorts for this contingency and will instead make this an amendment if it transpires to be an issue.

The DfE do not hold data for children educated at private schools or educated home, thus for these cases (~7%) the diabetes audit data would not be assigned a linkage ID by DfE. Unlinked cases would not form part of the core analysis, however Cardiff University would carefully examine why different cases did not link. For example, there may be noticeable patterns in children who did not link for technical linkage reasons, such as lower matching rates for children from certain minority ethnic groups where the matching algorithm may not be as efficient for those names. Cardiff University would also consider if the diabetes health outcomes are different for those children who do not link, since that may motivate a change in monitoring, for example for children who are educated in settings other than at school (home educated).

There should be no NDA cases in the requested extract that will not meet the conditions of the study. It would be expected all cases to have attended school, and those that do not attend university are still important as controls for comparison with those that did. There will be no request for controls from NHS Digital.

Although linkage with Welsh educational data was undertaken under linked agreement DARS-NIC-158283-T2Q2D versions 0-2, data is requested for individuals resident in both England and Wales for the current request because individuals can be located across borders for their medical care or educational settings, and individuals move across borders over time. The study will combine estimates from data linked to Welsh educational records in the SAIL databank, and data linked to English educational records in the ONS-SRS, to create a single estimate for England and Wales.

Cardiff University request individual-level data for modelling changes over time for individuals, e.g. how individual differences in education experiences affect diabetes management and vice versa. Cardiff University requests identifiers only for the purpose of data linkage; however, the analysis will be carried out on de-identified data.
Cardiff University request data from 2003/04 to 2017/18 to best model trajectories of HbA1c and the time to the first onset of early complications.
Cardiff University requests data for England and Wales to ensure maximum power and generalisability of the results.

Cardiff University confirm there are no alternative, less intrusive ways of achieving the purpose. A high degree of data linkage is required for this study in order to follow the entire life course trajectories of both diabetes related health and education through the combined linkage of both the paediatric and adult diabetes audits, and education data from primary school to university. Patient identifiers are sent to different organisations than the substantive clinical data to minimise the risk of disclosure.

The request has been restricted to only the essential clinical measures and associated metadata such as date of measure, location of measure (clinic).

Cardiff University are the research Sponsor and sole data controller for this DSA.

The personal data for this study is processed under UK GDPR Article 6(1)(e) - Public Task - for academic medical research carried out as a task in the public interest. The processing is necessary for Cardiff University (as a public authority for the purposes of data protection legislation) to perform a task in the public interest. The task has a clear basis in law.

The special category personal data for this study is processed under UK GDPR Article 9(2)(j) 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 right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject. The public interest lies in the improved evidence on the links between educational measures (school settings, absence, attainment) and health measures (HbA1c levels) for young people with diabetes. The study aims to improve care for all young people with diabetes in the education system by informing clinicians and commissioners of variation and outcomes and complications to support work to improve and standardise treatment selection choices. For these reasons, the processing also meets the conditions of Schedule 1 Part 1 paragraph 4 of the Data Protection Act 2018.

Cardiff University has approval from the Health Research Authority for data providers to set aside the common law duty of confidentiality with respect to the sharing of person identifiers in health datasets, known as a ‘Section 251 exemption’.

SAIL and DHCW are data processors for the Welsh subsets of data as described above. The DfE and ONS are data processors for the English subsets of data as described above.

The Welsh Government and HESA provide education data which is linked with NHS Digital data.

The RCPCH provide NPDA identifiable data to NHS Digital, and NPDA substantive data to the ONS-SRS. The Healthcare Quality Improvement Partnership (HQIP) are the data controller of the NPDA.

The Welsh Government, HESA, RCPCH and HQIP do not process NHS Digital data.

None of the organisations listed above, except Cardiff University, determine the purposes or the means of the processing of NHS Digital data. They are not therefore considered joint data controllers of NHS Digital data under this DSA.

The Medical Research Council (MRC) and UK Research & Innovation (UKRI) have provided funding for this study. The funders do not determine the purposes or the means of the processing of NHS Digital data and do not process the data. They are not therefore considered joint data controllers or data processors of NHS Digital data under this DSA.

Yielded Benefits:

The study has been lucky in terms of the timings of the opportunities to contribute to the policy debate. Cardiff University were able to provide evidence, based on analysis of the Welsh data, to the new Additional Learning Needs Bill for Wales and the Special Educational Needs and Disabilities Green Paper in England, though these are yet to report their findings so have not yet resulted in any change in practice. To support public involvement and engagement, the study team has given presentations of the results obtained so far to practitioners (e.g., the Royal College of Paediatrics and Child Health annual conference, the Brecon Annual Meeting of type 1 diabetes practitioners in Wales), patients (e.g., the Yorkshire and Humber Young People’s Diabetes Network Annual Meeting), and researchers (The Type 1 Diabetes Consortium at the Diabetes UK conference). The findings have been well received by clinicians, who report that they feed back some of the findings, particularly the ‘good news story’ that overall, children with diabetes attain in school as well as their peers without diabetes. Practitioners have also reported that the quantification of expected absence (average and range) has also been useful to provide a benchmark for children with diabetes.

Expected Benefits:

Although laws exist in England and Wales to support people with health conditions in schools and universities, the legislation (and associated guidance and resulting practice) does not always meet the child's needs. In schools, there is a gap between support for special needs (which stems from the Children and Families Act 2014 and traditionally focused on support for ‘learning conditions’, e.g. dyslexia) and support for children with health conditions (which stems from the Equalities Act 2010 and traditionally focuses on support through reasonable adjustments). In universities, students lack clarity on how financial support for health conditions works, particularly for conditions like diabetes, where there is wide variation in the required amount and type of support. Students will also be adjusting to having less clinical support than they are accustomed to, having only recently transitioned from paediatric to adult services before moving to a university campus far from their ‘home’ clinical team.

The high-level goal of this study is to ensure that children living with diabetes (and their families) can (i) manage their condition to stay healthy and (ii) fulfil their potential in education. By combining health data (including HbA1c) and education data (including educational outcomes and support), the project hopes to generate robust evidence of variation in outcomes, which may motivate changes in legislation and practice. Although the primary route for an impact on health is through policy change (and the associated improvement in guidance and implementation); the project will continue to inform practitioners of the range of outcomes and thus show the scope and potential benefit from improved support, both in terms of educational outcomes and diabetes-related health outcomes, and the link between the two.

Study outcomes for children in schools: This will focus on how rates of absence and attainment trends are associated with diabetes management. The longitudinal (annual measures) nature of the data means that researchers can compare outcomes over time; for example, annual measures of school absence can help show the changes in missed schooling before vs after a diagnosis of diabetes, in addition to comparing absence rates with similar children who were not diagnosed with diabetes. Similarly, regular measurements of HbA1c allow researchers to identify changes in blood glucose management as a child moves from primary school to secondary school. If differences exist, clinical teams may choose to give additional support based on non-clinical (educational) challenges that may influence diabetes management.

Study outcomes for children leaving school: Students with diabetes who are leaving compulsory schooling will transition to A-levels, vocational courses, or the labour force, and the study will model how these choices subsequently impact the trajectories of HbA1c. For example, the study may find that those who choose not to continue formal education beyond age 16 have less structure, affecting their diabetes management. If differences exist, clinical teams may choose to provide extra support or delay the transition from paediatric to adult diabetes care until after the transition to employment is complete.

Study outcomes for higher education students: The work focusing on universities will look for changes in trajectories of diabetes management, for example, comparing those that live at home with those who move away to attend university, identifying whether there is a significant change in management and the timings of that change. Differences may be seen in HbA1c during the first term as students try to reconcile good management routines with university social life, or perhaps HbA1c may not alter until the stress of the final year of university with impending high-stakes exams. If differences exist, clinical teams may choose to be more proactive in supporting the transition between home and university care and arrange follow-up appointments, both at home and at university, scheduled at times that fit around academic term dates.

Study outcomes for young adults: The research on young adults focuses on how childhood HbA1c levels and educational outcomes determine early adult health outcomes. This study focuses on the ‘double whammy’ effects of a history of sub-optimal HbA1c levels and lower educational outcomes. Relevant determinants might include individual factors (e.g. special educational needs or behavioural issues), school factors (e.g. lack of appropriate support), or family-level factors (e.g. socioeconomic status). If differences do exist, clinical teams may choose to flag individuals earlier who are struggling with diabetes management and education as being particularly vulnerable and receive additional support in the adult care setting, with potential health and social benefits to the individuals (who are most likely to develop early diabetes-related complications).

In partnership with Diabetes UK the project held a series of public workshops which reported the expected measurable benefits from the perspective of patient and public involvement participants (https://www.adruk.org/fileadmin/uploads/adruk/Documents/Diabetes_Education_Public_Workshop_Report_Aug_2021_01.pdf, and included: more understanding by others of what they are going through, more support with managing diabetes in schools, and earlier intervention at times of high stress. There was a focus on diabetes-education interactions during transition from paediatric to adult diabetes care – this comes amongst many other stressful events that teenagers go through, not least of which were the GCSE exams.

Participants focused on the benefit of bridging the gap in educational experiences between children with type 1 diabetes versus those without – “ensuring an equal playing field”. Participants reported a lack of knowledge around the impact of type 1 diabetes on education, which they felt was likely to be large. They felt that a study of this kind could influence policy, and ultimately the education setting, which in turn could result in a better experience for young people. A related point was the hope for increased general awareness and understanding of how type 1 diabetes might affect individuals beyond direct medical outcomes.

Outputs:

The expected outputs will be academic journal papers modelling the associations between education outcomes and trajectories of diabetes-related health.
The nature of data-based outputs would be descriptive statistics and regression coefficients, all of which would be aggregate data with small numbers (less than 10) suppressed; this is a requirement of taking any results out of the ONS-SRS and is rigorously checked by ONS analysts.

The original aim was to finish the draft of the outputs using data in SAIL disseminated under DARS-NIC-158283-T2Q2D versions 0-2 between 21/01/2019 to 21/08/2020, though the first of these papers was only accepted for publication on 01/08/2022 by the journal ‘Diabetes Care’. It is likely that future analysis will combine estimates from data deposited in both SAIL databank and ONS-SRS, and is expected to be published before 2025.

Findings from the first paper have been presented at an invited address at the Royal College of Paediatric and Child Health annual conference and a presentation to clinicians at the Brecon Group Annual meeting.

The first research paper is based upon how measures of diabetes-related health influence educational trajectories, and shows that (i) children with diabetes have higher absence rates, but similar attainment and progression to university as their peers and (ii) HbA1c tracks with educational outcomes. Cardiff University provide tentative evidence that this link is likely to be the result of external factors affecting both education and health outcomes. Due to issues with the NPDA data, Cardiff University have not yet been able to undertake any analysis pertaining to how education settings affect HbA1c trajectories, or how external factors affect both diabetes-related health trajectories and educational trajectories.

Any further study findings will be shared with clinicians through academic publications (e.g., Diabetes Care) and presentations at national forums (e.g., Diabetes UK Conference & Brecon Scientific Meeting). The findings will be shared with the public more broadly using a plain English version of the study; this will hopefully be published in the diabetes audit report and presented by the project lead at patient forums. Further engagement work with the results will be coordinated with Cardiff University’s coproduction group of young people.

An agreement is in place with Royal College of Paediatric and Child Health to include a patient level report on associations between diabetes related health and education for their annual report. Updates and summaries will be published on the project website (https://www.cardiff.ac.uk/research/explore/research-units/childhood-health-and-education?).

A dissemination plan has been produced in partnership with Diabetes UK. This includes two papers supported by plain English reports produced by the Diabetes UK policy team. The research team and Diabetes UK have begun to meet with policymakers from England and Wales, but it is too early to say how these connections will be exploited to communicate results. All outputs will belong to Cardiff University but will be freely available to Diabetes UK to use as agreed under Cardiff University’s heads of terms. Information about how the study has already consulted young people are provided on the webpage for the first 2021 patient and public engagement workshop: https://www.adruk.org/our-work/the-personal-cost-of-health-conditions-in-childhood-engaging-the-public/public – ADR UK and the workshop report: https://www.adruk.org/fileadmin/uploads/adruk/Documents/Diabetes_Education_Public_Workshop_Report_Aug_2021_01.pdf

In the second workshop, the young people produced a video explaining how researchers use data relating to young people with diabetes. Although the video and report produced are not yet released, there is a holding version on YouTube: https://www.youtube.com/watch?v=oimfnSoENxo.

Several of the representatives from Cardiff University’s patient and public involvement sessions have become the coproduction group, helping steer the project and the next iteration of the public engagement.

Processing:

This study uses a split file process to transfer individual-level data. The objective is to ensure that, as far as possible, every person’s health data is processed separately from the identifiers that would link that health data to a person.

Each clinical data provider assigns a study-specific pseudonymised identifier (study ID) to each participant. They then split the whole dataset into an identifiers dataset (containing variables such as NHS number, name, date of birth, postcode, and gender) and a substantive dataset (containing de-identified clinical/ education data).
The substantive data is transferred directly to the repository.

The identifiers file is shared with a trusted third party (details below) who uses the identifiers to match individuals’ ‘study ID’ (pseudonymised identifiers) to their ‘linkage ID’, and then deletes all the real-world identifiers such as names and dates of birth, before transferring the ‘study ID’ and ‘linkage ID’ into the repository where the ‘study ID’ enables re-joining of the ‘linkage ID’ to the substantive data, and the ‘linkage ID’ enables linkage to the other datasets which have been processed in the same way.

Under linked agreement DARS-NIC-158283-T2Q2D versions 0-2, the data flows were as follows:

NHS Digital identified people born from 01/09/1983 to 31/08/2002 (ie the 1983/84 to 2001/02 academic years) who appeared in the National Diabetes Audit (NDA) dataset and assigned them a unique study ID.

From this cohort, NHS Digital provided a one-off drop of data covering 2003/04 to 2017/18 audit years, containing the substantive diabetes-related data, accompanied by study ID only, to the Secure Anonymised Information Linkage (SAIL) databank at the University of Swansea. This file contained no identifying information other than the study ID.

From this cohort, NHS Digital also provided a one-off drop of data covering 2003/04 to 2017/18 ‘audit years,’ containing study ID, patient name, NHS number, date of birth, gender, and postcode only to Digital Health and Care Wales (DHCW).

DHCW created the ‘linkage ID’, a hashed version of the NHS Number, referred to by DHCW and SAIL as the Anonymised Linkage Field (ALF), then destroyed all the identifying information (name, NHS number, date of birth, gender, postcode). After this process, the identifiers file contains only the study ID and linkage ID, along with less disclosive versions of the demographic data (gender, week of birth, lower super output area). This file was onwardly flowed to SAIL.

Within SAIL, the substantive NDA file was re-joined to the identifiers linkage field file using the study ID so that each case of the substantive data got a linkage ID.
The same process as outlined above for NDA data also happened for the other de-identified datasets that were placed in SAIL, including the National Paediatric Diabetes Audit (NPDA), Higher Educational Statistics Agency dataset (HESA), and Welsh Dataset Education Records (WED, also known as the National Pupil Database or NPD). The NPDA, HESA, and WED data were linked to the NHS Digital NDA data using the linkage ID.

Data accessible to Cardiff University via SAIL therefore are:
- NDA data covering England and Wales
- NPDA data covering England and Wales
- HESA (higher education) data covering England and Wales
- Primary and secondary education data for Wales only (WED)

Under this Agreement, DARS-NIC-669808-V6T0M-v0, NHS number, date of birth, postcode and gender will be provided to NHS Digital by the Royal College of Paediatric and Child Health (RCPCH) for individuals born in the academic birth cohorts from 1983/4 to 2001/2 who are in the NPDA. NHS Digital will retrieve the names of these individuals and send identifiable Demographics data (names, date of birth, latest postcode, and gender) to the DfE alongside a unique pseudonymised identifier provided by the RCPCH. This will be a one-off deposit of data, though permission for annual data refreshes will be sought in the future.

Under linked agreement DARS-NIC-674735-Z0H6K-v0.0, NHS Digital identify individuals based in England and Wales and born between the 01/09/1983 to 31/08/2022 (ie the 1983/84 to 2001/02 academic years), in the NDA dataset under DARS-NIC-669962-W1F6D and assign them a unique study ID.

Using the NHS numbers held in the NDA, NHS Digital retrieves the names, date of birth, latest postcode, and gender from the Personal Demographics Service (NHS number was included for DHCW, but is not used here as Department for Education (DfE) are unable to process it).

These identifiers are sent to the DfE alongside the unique study ID. This will be a one-off deposit of data, though permission for annual refreshes of data will be sought in the future. DfE, acting as the trusted third party, will use the identifiers to match individuals’ ‘study ID’ to the ‘linkage ID’ (referred to by DfE and the Office for National Statistics (ONS) as the Pupil Matching Reference (PMR)), then destroy all the identifying information (name, date of birth, gender, postcode). After this process, the identifiers file contains only the study ID and linkage ID. (Unlike for the SAIL process, less disclosive versions of the demographic data [gender, week of birth, lower super output area], are not retained in the identifiers file). This modified identifiers file, containing only study ID and linkage ID is onwardly flowed to ONS-SRS repository where the ‘study ID’ enables re-joining of the ‘linkage ID’ to the substantive data, and the ‘linkage ID’ enables linkage to the other datasets which have been processed in the same way.

In parallel, under linked agreement DARS-NIC-669962-W1F6D-v0, substantive de-identified NDA health data for the 2003/4 to 2017/18 audit years for the above-described cohort will be sent to the ONS Secure Research Service (SRS), with the unique study ID only. This will be a one-off deposit of data, though permission for annual data refreshes will be sought in the future.

The project includes four further data flows which are not covered by the above-listed DSAs:

1. RCPCH send substantive deidentified NPDA data directly to the ONS-SRS, containing the same unique pseudonymised identifier as provided to NHS Digital under DARS-NIC-669808-V6T0M-v0
2. DfE send substantive de-identified education records (compulsory education) directly to the ONS-SRS, alongside the linkage ID
3. HESA send substantive de-identified education records (higher education) directly to the ONS-SRS, alongside the linkage ID.
4. After the DfE ‘hashing service’ use the identifiers supplied by NHS Digital to identify the correct individual against their records and retrieve their PMR, the original study ID supplied by NHS Digital, alongside the PMR retrieved by the DfE, are onwardly flowed into the ONS-SRS.

DfE will destroy the real-world identifiers as soon as the PMR field is retrieved and the necessary pseudonymised identifiers provided to the ONS-SRS. They will provide NHS Digital with a data destruction certificate.

ONS-SRS staff will work with the Cardiff University project lead to link the data records within the ONS-SRS environment. Substantive NDA and NPDA data will be linked with a PMR via the original study IDs, where the DfE have been able to retrieve a PMR.

NDA, NPDA, and education data will then be linked using the PMRs.

ONS-SRS staff will support Cardiff University with the linkage evaluation for the extract to be used for research, determining what percentage of clinical records have been successfully linked together (NPDA to NDA) and linked to associated education records (NPDA to NPD, NPDA to HESA, NDA to NPD, NDA to HESA).

Unlinked clinical data will remain in the ONS-SRS environment but will not be further processed for the purposes of these Data Sharing Agreements (DARS-NIC-674735-Z0H6K-v0.0, DARS-NIC-669808-V6T0M-v0, and DARS-NIC-669962-W1F6D-v0). Other linkage mechanisms may be explored in future to improve the success of the linkage, and data on individuals not in school or university may give rise to important information for diabetes management.

The linked de-identified data-sets will be made available for multilevel modelling analysis of the associations between education and health by Cardiff University on the ONS-SRS. The DfE will provide educational data for controls who do not have diabetes, alongside the educational data for those with diabetes, to enable Cardiff University to model the differences in outcomes.

The combining of health data with education data increases the likelihood that a person may appear as unique in the dataset; however, given that there are 30,000+ cases of young people with diabetes in England and Wales, and the high-level nature of the variables, it is unlikely that this dramatically increases the risk of re-identification. There will be no requirement or attempt to reidentify individuals for the purposes of the study or any other reason. The primary protection against reidentification is the creation of the linkage ID along with the destruction of the real-world identifiers prior to the data being put in the repository, meaning the data accessed by Cardiff University is de-identified. Secondly, ONS-SRS and Cardiff University will check data at the outset for any identifiability risks in the raw data or linked data before using it for analysis. This checking is routinely conducted by experienced analysts at ONS who use a mature process to identify small numbers, unique/rare cases, or other identifiability risks. Thirdly, researchers from Cardiff University who are accessing the data are trained not to re-identify data.

NHS Digital data is not being linked to any publicly available data.

Data processing is only conducted by substantive employees of the data processors who have specific teams and staff trained and explicitly employed for this purpose. Data will be accessed for analysis through secure remote gateways into the ONS-SRS. No data will be accessed outside the UK. Person identifiers (such as name and date of birth) will not be shared with third parties beyond those required to create the pseudonymised linked datasets (i.e. only DfE).

Public engagement focused on both the research questions and the data processing has been carried out throughout the project, going back to 2014 before the study was funded. The largest public involvement work has been the sessions with young people with type 1 diabetes in 2020, in partnership with Diabetes UK and the MRC Regulatory Support Centre. This work is summarised in a website (https://www.adruk.org/news-publications/news-blogs/public-views-on-the-use-of-personal-identifiers-for-linking-diabetes-and-education-data-for-research-439/) with links to the full report towards the bottom of the page. Further public engagement was carried out in 2022 across 19 focus groups, the link to the work is https://dareuk.org.uk/sprint-exemplar-project-steadfast, and the report summarising this work should be released in early 2023.

Exploring the mechanisms through which specialist home visiting produces health and well-being benefits for families. — DARS-NIC-374907-Q5W5W

Type of data: information not disclosed for TRE projects

Opt outs honoured: Anonymised - ICO Code Compliant (Does not include the flow of confidential data)

Legal basis: Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii), Health and Social Care Act 2012 – s261(2)(b)(ii)

Purposes: No (Academic)

Sensitive: Sensitive, and Non-Sensitive

When:DSA runs 2020-11-01 — 2023-10-31

Access method: One-Off

Data-controller type: CARDIFF UNIVERSITY, UNIVERSITY COLLEGE LONDON (UCL)

Sublicensing allowed: No

Datasets:

  1. Civil Registration (Deaths) - Secondary Care Cut
  2. HES:Civil Registration (Deaths) bridge
  3. Hospital Episode Statistics Accident and Emergency
  4. Hospital Episode Statistics Admitted Patient Care
  5. Hospital Episode Statistics Outpatients
  6. Civil Registrations of Death - Secondary Care Cut
  7. Hospital Episode Statistics Accident and Emergency (HES A and E)
  8. Hospital Episode Statistics Admitted Patient Care (HES APC)
  9. Hospital Episode Statistics Outpatients (HES OP)

Objectives:

(i) Specialist home visiting to support new families:

The Family Nurse Partnership (FNP) was developed and licensed by the University of Colorado as a voluntary, preventive programme for vulnerable young first time mothers. It offers intensive and structured home visiting, delivered by specially trained nurses, from early pregnancy until age two. The three aims are: to improve pregnancy outcomes, improve child health and development and improve parents’ economic self-sufficiency.

A strong and rigorous US evidence base, developed over 30 years, has shown FNP benefits the neediest young families in the short, medium, and long term across a wide range of outcomes helping improve social mobility and break the cycle of inter-generational disadvantage and poverty.

(ii) Work to establish the programme in the UK and the Cardiff evaluations:

The programme was adapted for implementation and introduced in England in 2007. Due to the greatly differing nature of publicly funded health and social care service provision and socio-cultural context between England and the US, the relative benefits of the programme need to be replicated in England and costs determined before wide-spread implementation can be recommended. The study team have conducted two definitive evaluations:

BB:0-2 - Department of Health and Social Care (DHSC) commissioned the ‘Building Blocks’ randomised controlled trial (RCT) from Cardiff University to provide independent evidence on the effectiveness of the FNP programme in improving short term outcomes for young parents and their babies. The trial began in 2009 and the findings which cover the period from pregnancy to the child’s second birthday' were published in October 2015. The FNP described these as “important early findings and add to the evidence the study team have from the US, Netherlands and other early evaluation in England to help improve FNP in England.”

BB: 2-6 - The National Institute of Health Research further funded Cardiff University to undertake a follow up study to examine child and maternal outcomes to age six. This built on the original study to examine the longer-term impact of FNP intervention. Its principal objective was to determine the effectiveness of the FNP programme in reducing objectively measured long-term maltreatment outcomes when compared to usually provided health and social care alone. The main research outcomes were child in need status, child protection registration, referral to social care. Other indicators of programme impact included measures of injuries and ingestions and hospital DNA rates. The study report has been externally peer reviewed and a final version will be published by NIHR in November 2020.

In the BB:2-6 study, Cardiff University followed-up the cohort of mothers and children who took part in the first Building Blocks trial (BB:0-2) by obtaining health and mortality data from NHS Digital, data from Department for Education (DfE) and data from the Department of Health (abortion statistics) and linked with the original trial data.

(iii) Evidence from the Cardiff work and NHS Digital supplied data to date
The original Building Blocks trial (BB:0-2) provided evidence for the short-term effectiveness of the programme (up to 2 years after birth). The Building Blocks: 2-6 (BB:2-6) study has now provided evidence for the medium-term effectiveness and costs of one of the most promising early intervention programmes for reducing risk of child maltreatment in a targeted vulnerable population. Specifically, data requested from NHS Digital provided the basis for key study outcomes which are indicators of maltreatment. Both studies have provided evidence which has informed policy decision-making at both national and local level about whether to continue implementing the programme and also changes to programme content, format and delivery.

(iv) Newly funded work 2019/20:

This new request reflects the receipt of additional funding to analyse the existing pseudonymised study data already held in the data safe haven (which includes NHS Digital supplied data). The newly funded analyses will explore the mechanisms by which specialist home visiting may have its effect for recipient families. Two funders (Nuffield Foundation, European Research Council (ERC)) have provided resource to support the two separate analyses both of which will utilise the same existing data set.

Proposed analyses funded by ERC and Nuffield Trust:
The newly funded analyses will explore the mechanisms by which specialist home visiting may have its effect for recipient families. The analyses to date have determined the impact of the home visiting intervention upon a wide range of family outcomes. These range from the antenatal period through to early in children’s school life (i.e. at key stage 1). The study team have explored how home visiting has benefited some families but been less effective for others. The study team have also assessed the additional benefit that receiving more home visits has had for some of the more important outcomes that have been assessed. The study team have explored the perceptions of both mothers and family nurses (i.e. the professionals who deliver home visits to families) about how the home visiting may exert its effect (and what may be a barrier to its effectiveness). The study team have explored how some outcomes (such as children being designated as In Need by local departments of children’s services) are influenced by a range of potential risk factors (for example, attributes of the mother and child).

The new analysis will take an econometric approach to explore how the benefits of the home visiting intervention are produced. It will explore the mechanisms through which the home visiting intervention is hypothesised to produce benefits (for example, whether expected changes in maternal behaviour occur and whether these then lead to other improvements). It will explore how characteristics of nurses plays any part in observed effects (for example, whether some nurses are associated with more or less improvement. This may be due to differences in clinical experience or training, for example). It will explore further how home visiting dose and quality may be linked to differential family benefits (visit-level data available for each mother in receipt of specialist home-visiting provides detail on how many visits were delivered, their duration and the content addressed using the scheme’s standard monitoring form). Finally, it will also assess how reported use of other public sector services may also have an influence. The existing trial cohort data (including that derived from the NHS via NHS Digital) will be analysed using state-of-the-art economics methods, benefiting from the input of the lead investigator who has secured the additional funding. The analysis will be able to model dynamically the study data to establish more clearly than has been possible previously how the home visiting programme produces benefits for families.

These two analyses (Nuffield and ERC) will both use the existing pseudonymised data set. The first (Nuffield), will involve two similar study cohorts being assessed in the new analyses (the other being from a study in Germany which implemented a similar intervention). The analyses for the two cohorts will be run in parallel (i.e. the data will remain where currently located) but similar research questions will be asked of the two study datasets. No UK data currently held by Cardiff as data controller will be shared or linked to the German data. Only results will be compared when interpreting the outputs from the two analyses. The second funded analysis (ERC), will involve a broader range of data drawn from a variety of sources these will include: distance from / price of / availability of goods and services for children (e.g. price of toys, distance from GP/parks, food prices); and prices of cigarettes, alcohol and local wages. Like the Nuffield funded work, the ERC analyses will involve data from the Building Blocks trial cohort being analysed separately from any other project data but will share a common underlying research question and specialist economics methods. For the analysis of the Building Blocks dataset more detailed exploration of (non-NHS) trial data based on existing post code will be used to determine the role of local context (including policies) upon family outcomes, no attempt will made to identify or link the data.

The two new project grants are led by a member of staff at UCL and the analyses will be undertaken by the research fellow and themselves, both of whom are substantive employees of UCL and hold honorary titles with Cardiff University for the duration of the studies. Access will be as described above, and access will be approved by the study lead at Cardiff University. All analyses will be reviewed and approved by the study lead at Cardiff University. This study lead is the chief investigator and is costed into both grants to provide these approvals and oversee all outputs from the additionally funded work. A statistician based in Cardiff and currently approved to access the cohort dataset will have responsibility to review the statistical analysis plan for both Nuffield and ERC-funded studies.

Cardiff University was the Data Controller (and sponsor) of the original study and Swansea University (and Cardiff University) were Data Processors. This new application will also include UCL as a joint controller and processor. The additional funded work is funded by both the European Research Council and Nuffield Trust. Neither body may make decisions related to how the data are processed and will not have access to the data. The Chair of the Information Governance Review Panel at SAIL has confirmed that as the data being accessed in relation to the newly funded work are appropriately anonymised, no further ethical review is required.

Cardiff University
Article 6:1(e):This project is a task in the public interest because it was commissioned by the Department of Health to evaluate the impact of home-visiting interventions on child maltreatment.

Article 9:2(j):This is scientific research which uses only the data required to answer these important research aims, using a data safe haven to protect the data and outputs this research will directly influence policy making regarding home visiting programmes to benefit the UK public.

University College London
Article 6(1)(e) of the General Data Protection Regulations (GDPR), i.e. “a task carried out in the public interest”. It also falls under Article 9(2)(j), “processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes”.

(v) Clarification about data subjects

The data subjects were 1645 women (and their first-born child(ren)) who met eligibility criteria for participation in the Building Blocks trial of the Family Nurse Partnership (FNP) and were nulliparous (a woman who has never given birth either by choice or for any other reason. This term also applies to women who have given birth to a stillborn baby, or a baby who was otherwise not able to survive outside the womb) women aged 19 or under, living in one of 18 local authority FNP catchment areas in England who were recruited by 24+6 weeks gestation, had conversational level of English and were able to consent to research.

Women were recruited by locally-based research professionals usually employed by local NHS organisations following identification through NHS maternity units. Women provided their written informed consent to participate including for data in their administrative health records to be provided to the research team.

Relevant cohort data provided for trial by NHS Digital were Inpatient, Outpatient, A&E datasets and used to follow-up mothers recruited into the study and their babies, up until the baby’s second birthday. These datasets were the main source for two of the primary outcomes (and several secondary outcomes).

The BB:0-2 trial began in 2009 and the findings which cover the period from pregnancy to the child’s second birthday' were published in October 2015.

The National Institute of Health Research funded BB:2-6 study follow up study examined child outcomes to age six, building on the original study examining the longer term impact of FNP intervention. A subsequent application for data from NHS Digital (DARS-NIC-333498-D1K7G) permitted collection of the same health data for an additional four years, until the mother’s first child(ren) was aged six years old. As identifiable data were required to enable linkage of the original trial cohort data to these new records, section 251 approval was given.

The Building Blocks study originally involved a lay group who advised on aspects of trial delivery and provided input into the design and setup of the study. An independent group of young people also provided advice to the research team and a lay perspective on the general management and delivery of the project and in particular the preparation of the outputs from the study.

Yielded Benefits:

The original Building Blocks trial (BB:0-2) provided evidence for the short-term effectiveness of the programme (up to 2 years after birth). The Building Blocks: 2-6 (BB:2-6) study has now provided evidence for the medium-term effectiveness and costs of one of the most promising early intervention programmes for reducing risk of child maltreatment in a targeted vulnerable population. Both studies have provided evidence which has informed policy decision-making at both national and local level about whether to continue implementing the programme and also changes to programme content, format and delivery.

Expected Benefits:

It is expected that the new analyses will inform the further development of the home visiting programmes in both the UK and in Germany, as well as being more generally informative for the programme developers in the US. To date, the Cardiff-led study of the programme has resulted in national level innovation and improvement in England (e.g. via the FNP ADAPT programme – the final report on this programme was published online in March 2020) with a consequent impact of improved services for families. The newly planned analyses will aim to similarly lead to service development and is part of suite of similar activity being undertaken either led by Cardiff (e.g. in Scotland) and in association with Cardiff (e.g. in England) although which are out with this specific DSA. This body of work will shed light on the extent to which (and how) early interventions are cost-effective in promoting health and reducing inequality.

The primary aim of this new work is to build knowledge on what determines the effectiveness of home-visiting programmes targeting first-time teenage mothers in contexts with existing universal health care and extensive welfare provision. Understanding the mechanisms through which these interventions achieve their observed impacts is crucial for their long-term sustainability and cost-effectiveness of the programme both in England and Germany.

The FNP National Unit recently released the final report of the ambitious ADAPT (Accelerated Design And Programme Testing) project, which introduced and tested several adaptations of the programme following the publication of the original trial results. These additional analyses will provide timely new evidence to inform the future of the programme, which is uncertain.

From the study team’s previous follow-on study (funded by NIHR PHR) Cardiff University have been meeting with DHSC and FNP National Unit to discuss the emerging results (which will be published later in 2020. This has influenced their decisions and recommendations made regarding funding for the FNP National Unit which remain confidential at this stage.

Processing:

All data transfers have now concluded and data now would only be analysed in the data safe haven, there is no requirement and there will be no attempt to reidentify the individuals in the cohort. Any reference to data flows are now historic. Data already supplied by NHS Digital are described under Products for clarity (having been previously supplied under agreement DARS-NIC-333498-D1K7G-v4.

Cardiff University provided the following fields to NHS Digital:
• Study ID
• NHS Number
• Date of Birth
• Sex
• Postcode

NHS Digital linked this information to the HES and mortality data, stripped out the identifiers and returned a pseudonymised output to the data processor, the Swansea University’s Secure Anonymised Information Linkage (SAIL) Databank. Cardiff University also sent SAIL a copy of the original study (BB:0-2) data and separately sent participant identifiers to the Department for Education which supplied linked pseudonymised data from its records to SAIL. Data was requested and approved from the Department of Health abortion statistics team to link to the mothers in this cohort. The study used the abortion data to calculate whether a mother had a subsequent pregnancy and the study created a composite outcome which was used to report the rates of subsequent pregnancies between the two groups (control and intervention).

The only identifier provided to SAIL from each of the 4 sources was the Study ID. SAIL assigned an anonymous linking field (ALF) to each individual to replace the study ID. The study ID-ALF key is encrypted and stored securely by SAIL. The individuals analysing the data will not have access to this key. The key was the same for each datasets so that all data could be linked without using identifiers to do so. The key is retained by SAIL to enable any individual who expresses at a later date a wish to be removed from the study to be removed from the database.

The research team work within the United Kingdom Clinical Research Collaboration (UKCRC) fully registered clinical trials unit - Centre for Trials Research at Cardiff University and at University College London (UCL).

Mitigating risks of re-identification: All data will be maintained in the safe data haven in Swansea [SAIL] within which all analyses will be undertaken. The research database will not be made available to other researchers (this will be a project specific resource). Access to the pseudonymised dataset will be via a secure remote portal. No data will leave the secure environment at SAIL. Approved, named data users can access the portal and defined data views remotely, subject to the appropriate access level being set and secure access keys being provided to them. Data cannot be downloaded from the portal; all exports of data are approved by SAIL who ensure raw data and results with small numbers (and therefore a risk of identification) are not exported. Only graphs, statistical analyses outputs and aggregated tables will be exported out of the secure portal. All staff who access the data have completed the MRC Safe researcher training which covers areas such as data protection and confidentiality.

Following analysis, aggregated results / publishable information can be requested out of the secure environment for wider disclosure (subject to the data file being approved by data guardians at SAIL). Data guardians check for sensitive data, and small numbers that could risk disclosure before approving the file.

Both organisations party to this agreement must comply with Data Sharing Framework Contract requirements at their respective organisations, 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 members of staff holding the honorary title with Cardiff University will be contractually bound to the data sharing agreement between UCL and Cardiff University. This contract refers to both the Framework and project-level Data sharing agreement as well as the User agreement with SAIL which covers disclosure requirements.

Building Blocks Trial - Data Archive — DARS-NIC-313754-G6X4Z

Type of data: information not disclosed for TRE projects

Opt outs honoured: Anonymised - ICO Code Compliant (Consent (Reasonable Expectation))

Legal basis: Health and Social Care Act 2012 – s261(2)(c)

Purposes: No (Academic)

Sensitive: Non-Sensitive

When:DSA runs 2019-04-01 — 2024-03-31

Access method: One-Off

Data-controller type: CARDIFF UNIVERSITY

Sublicensing allowed: No

Datasets:

  1. Hospital Episode Statistics Accident and Emergency
  2. Hospital Episode Statistics Admitted Patient Care
  3. Hospital Episode Statistics Outpatients
  4. Hospital Episode Statistics Accident and Emergency (HES A and E)
  5. Hospital Episode Statistics Admitted Patient Care (HES APC)
  6. Hospital Episode Statistics Outpatients (HES OP)

Objectives:

HES data was supplied to Cardiff University for processing for the purpose of the Building Blocks study.

This study is now completed and closed.

In order to comply with MHRA guidance and policy on good clinical practice, the data will be securely archived to ensure it is possible to verify the conclusions published from this study.

This agreement permits processing of the data for the purpose of secure storage and back up.

This agreement does not permit any further processing that involves analysis or linkage other than for the purpose of verifying findings in line with the original objectives of the study by repeating previous analyses described in this Agreement. Following publication of the study findings, it is possible that the findings will be questioned or challenged by third parties through direct contact with Cardiff University, contact via the publishing journal or an open letter. In such circumstances, Cardiff University may re-run the previously analyses undertaken to verify that the published results were accurate and may write a response to be issued directly to the challenger or published. Cardiff University may not use the data to undertake different analyses to those undertaken during the original analysis.

This agreement does not permit any onward sharing of the data with the exception that the data may be viewed for the purpose of an audit by a regulator such as the Medicines and Healthcare products Regulatory Agency (MHRA). The MHRA maintains a list of all clinical trials registered with them and reserve the right to audit the trial at any point from when the trial is opened until 5 years post trial closure. A sponsoring organisation and/or the data controller itself may also exercise the right to audit.

If any further data processing is required in addition to the above purposes or if the data needs to be moved to a different location/organisation the applicant must submit an amendment request to NHS Digital before data is accessed.

For reference only, the following describes the objective for processing before the data was archived.

The Family Nurse Partnership programme (or FNP) offers young mothers having their first baby support from a Family Nurse. The nurse visits the woman at home during pregnancy and until the baby’s second birthday. As FNP has been shown to help young families in the USA the government have introduced it in England.

FNP aims to help young mums to:
•have a healthy pregnancy
•improve their child’s health and development
•plan their own future

Building Blocks was a study of how well FNP works in England. 1,645 pregnant teenagers joined the study between 2009 and 2010.

Women were split into two groups. One group was offered FNP and one group had usual care from Health Visitors. All women in the study also had maternity care and any other extra support available to young families in their area.

The study looked at how well the mothers and babies were doing up until the baby’s 2nd birthday.

Actions have been taken to pseudonymise the original Building Blocks trial data. These included converting baby’s date of birth to week and month of birth. Cardiff University separately holds the identifiers and trial data for each participant for the purpose of a separate follow up study, ‘Building Blocks 2-6’. The Building Block 2-6 cohort is a long term follow up of the original Building Blocks cohort. Both use the same cohort but they are separately funded projects, with separate ethics applications. The dataset for the original Building Blocks study is securely archived in a data safe haven which prevents the download of individual data and linkage to identifying data items and will not combined with any data from ‘Building Blocks 2-6’.

Yielded Benefits:

The main findings were; •FNP did not reduce the number of women that smoked in pregnancy. In both groups 56 women out of every 100 smoked late in pregnancy. •FNP did not reduce the number of small or premature babies. In both groups the average baby weighed 7lb 1oz. •FNP did not reduce the number of women getting pregnant again within two years. In both groups 66 women out of every 100 were pregnant again within two years. •In both groups nearly 80 out of every 100 children were seen in hospital as an emergency before their second birthday. •FNP was found to be more expensive than usual care. Other findings were; •About four in every 10 mothers in both groups breastfed their baby. •By two years of age around one in 10 children had needed a trip to A&E because of an accident or swallowing something they shouldn’t have. •Social services had been involved with 14% of children allocated to FNP and 8% of children allocated to usual care. •Children in the FNP group had better language skills than children who got usual care only. These findings were used by the department of health to decide how best to support young families in England.

Expected Benefits:

This data will be retained to comply with guidance and policy on good clinical practice and regulations. It also preserves a unique database which could still yield future benefits for young families by looking at the effects on children in the longer term.

Outputs:

This study is now completed and closed.

The following description of outputs is for information only and the processing of the data occurred before this current Agreement.

No new outputs will be produced using the data under this agreement.

The results have already been presented to the funder (in Spring 2015), the main trial paper has been submitted for publication and results have been presented at conferences from September 2015. Further methodological papers and exploratory analyses were published throughout 2016-2017.

The analysis and paper on infections using A&E and Inpatients data and subsequent pregnancies has been drafted and will be published in 2017/18.

To date the conferences in which the trial results were presented include the Public Health Wales Conference on Nov 3rd and the Royal College of Midwifery Conference on 10th/11th November 2017
Journal articles include;
• Corbacho B, Bell K, Stamuli E, Richardson G, Ronaldson S, Hood K, Sanders, Robling M, Torgerson D. Cost-effectiveness of the Family Nurse Partnership (FNP) programme in England: evidence from the Building Blocks trial. Journal of Evaluation in Clinical Practice 2017; 1-8. 2017 doi: 10.1111/jep
• Channon S, Bekkers M-J, Sanders J, Cannings-John R, Robertson L, Bennert K, Butler C, Hood K, Robling M. Motivational interviewing competencies among UK family nurse partnership nurses: a process evaluation component of the building blocks trial. BMC Nursing 2016 15:55. DOI: 10.1186/s12912-016-0176-0
• Robling M, Bekkers M, Bell K, Butler CC, Cannings-John R, Channon S, Corbacho Martin B, Gregory JW, Hood K, Kemp A, Kenkre J, Montgomery AA, Moody G, Owen-Jones E, Pickett K, Richardson G, Roberts ZES, Ronaldson S, Sanders J, Stamuli E, Torgerson D. Effectiveness of a nurse-led intensive home-visitation programme for first-time teenage mothers (Building Blocks): a pragmatic randomised controlled trial. The Lancet, vol. 387, Issue 10014, pages: 146-155. 2016.
• Stamuli E, Richardson G, Duffy S, Robling M, Hood K. Systematic review of the economic evidence on home visitation programmes for vulnerable pregnant women. British Medical Bulletin 2015 115:19-44

• Robling M et al. Evaluating the Family Nurse Partnership Programme: Main Results of The Building Blocks Trial. Presentation to the Welsh Public Health Conference, November 2015, Cardiff.
• Confidential results consultation with Policy, Practice and Academic Stakeholders (England, Scotland, Northern Ireland, US), June 2015
• Public stakeholder meeting on Building Blocks Trial Results, January 2016
• Foundation Years Information and Research: Seminar Series (x3 meetings): Portcullis House, Westminster 2017
• Feedback of main trials findings to Building Blocks Study participants (mailed leaflets), October 2015

• Evaluating the long-term effectiveness, and the cost and consequences of the Family Nurse Partnership parenting support programme in reducing maltreatment in young children. Robling MR et al.
Lugg-Widger F, Cannings-John R, Channon S, Fitzsimmons D, Hood K, Jones KH, Kemp A, Kenkre J, Longo M, McEwan K, Moody G, Owen-Jones E, Sanders J, Segrott J, Robling M. Assessing the medium-term impact of a home-visiting programme on child maltreatment in England: protocol for a routine data linkage study. BMJ Open 2017; 7:e015728. doi:10.1136/bmjopen-2016-015728.
• BABBLE: Establishing the impact of specialist home-visiting on children’s language development and parent-child interaction: development pathways to later adverse outcomes

Processing:

This study is now completed and closed.

This agreement permits processing of the data for the purpose of secure storage and back up.

This agreement does not permit any further processing that involves analysis or linkage other than for the purpose of verifying findings in line with the original objectives of the study by repeating previous analyses described in this Agreement.

This agreement does not permit any onward sharing of the data with the exception that the data may be viewed for the purpose of an audit by a regulator such as the Medicines and Healthcare products Regulatory Agency (MHRA).

If any further data processing is required in addition to the above purposes or if the data needs to be moved to a different location/organisation the applicant must submit an amendment request to NHS Digital before data is accessed.

The clinical data (used for analysis and now to be archived) is stored on CU servers with only a trial ID to identify and link participants across datasets. The data subjects' identifying details and contact details are held separately and access permissions do not allow for any staff member to access both clinical and identifying data files.

In line with Cardiff University’s Standard Operating Procedure (SOP), the archived data will be held on Cardiff University, CTR secure server and made available only to the CTR designated Electronic Archivist (EA). Therefore, all project data will be moved from its current location to this archive area on the shared drive. Access permissions will restrict access only to the EA.

In the event that the data needed to be accessed for the purposes of audit or to enable verification of previous findings, the dataset would be transferred back to the CU servers with authorisation from the Principal Investigator where it can be accessed by the study statistician only for the purposes of verifying results of previous analyses by rerunning analyses that were previously undertaken. Data would be accessible only for as long as is required to enable verification of the analyses and to write a response as appropriate. Responses will adhere to the rules for small number suppression in the HES Analysis Guide.

The data may not be transferred to any other location and may only be accessed by substantive employees of Cardiff University for the purposes described above.

Under the previous Agreement with NHS Digital access to the data was permitted for both Cardiff University and York University. York University is no longer permitted to access the data and has confirmed destruction of any data it accessed under the previous Agreement.

The data originally requested from NHS Digital was for use in the Building Blocks Trial [ISRCTN 23019866]. Inpatient, Outpatient, A&E datasets were used to follow-up mothers recruited into the study and their babies, up until the baby’s second birthday and these were the main source for two of the primary outcomes (and several secondary outcomes).

This archive agreement is required for the following reasons;
(1) The raw datasets (and any associated datasets and syntax files) may need to be accessed for the purpose of audit;
An audit may be required during the period of this agreement. This would involve a systematic and independent review of trial-related activities and documents, to determine whether data were recorded, analysed and accurately reported. Without all the raw datasets this cannot be done.

(2) Further analysis is expected on these datasets in future;
Any further analysis will only take place following an amendment to this Agreement that would allow further processing of the data.

(3) Clinical trial data needs to be retained and accessible for 15 years after trial completion this requirement would be subject to further data extension requests – no data will be retained without an Agreement being in place.

MR1314 : FRAGMATIC: A randomised phase III clinical trial investigating the effect of FRAGMin® Added to standard Therapy In patients with lung Cancer. — DARS-NIC-291941-Z2Q1C

Type of data: information not disclosed for TRE projects

Opt outs honoured: Identifiable (Consent (Reasonable Expectation))

Legal basis: Health and Social Care Act 2012 – s261(2)(c)

Purposes: Yes (Academic)

Sensitive: Sensitive

When:DSA runs 2019-03-01 — 2021-08-20

Access method: Ongoing, One-Off

Data-controller type: CARDIFF UNIVERSITY, VELINDRE UNIVERSITY NHS TRUST

Sublicensing allowed: No

Datasets:

  1. MRIS - Cause of Death Report
  2. MRIS - Cohort Event Notification Report
  3. MRIS - Flagging Current Status Report

Objectives:

Mortality and NHS Registration data were supplied to Cardiff University for the purpose of a research study called “FRAGMATIC: A randomised phase III clinical trial investigating the effect of FRAGMin® Added to standard Therapy In patients with lung Cancer.”

Velindre University NHS Trust and Cardiff University are joint data controllers for the FRAGMATIC study with joint responsibility to determine the purposes or means in which the data will be used. Cardiff University will be the sole organisation processing the data.

The Data Controllers do not perceive any moral or ethical issues regarding dissemination of results. The majority of participants will already be deceased due to the advanced cancer at the time of recruitment. It is in the public interest to disseminate study results and a requirement of most funding streams. For example, CRUK funding expects the results to be published on their website for any study funded or endorsed by them. Results will be published in a format that does not allow individual the participants to be identified.

This study is sponsored by Velindre University NHS Trust. The sponsor is defined by the organisation(s) with legal oversight responsibility for the study. The study is sub-contracted by the Sponsor to conduct some aspects of the study included some data management and data controller activities. FRAGMATIC was developed by the Wales Cancer Trials Unit (WCTU) at Velindre University NHS Trust before the unit moved to Cardiff University in 2008. The FRAGMATIC study came through with support and endorsement from the NIHR Lung Cancer Chemotherapy sub-group and the trial funding application team included representatives of other bodies, e.g. Wales Cancer Trials Network and University of Glamorgan. These organisations are not currently supporting the study. Ongoing and future data processing will not involve these bodies and will be conducted at Cardiff University only. The Cancer Research UK (CRUK) grant was held by WCTU and only funded non-drug resource following standard CRUK practice. WCTU approached Pfizer for an educational grant to run in parallel to CRUK funding to fund the dalteparin (FRAGMIN) drug and its’ distribution costs for free. Following approval of a previous iteration of this agreement, the WCTU merged into the Centre for Trials Research (CTR) at Cardiff University. The Centre of Trials Research now manages the study.

The trial assessed the effect of adding dalteparin (FRAGMIN) for 24 weeks to standard treatment (trial arm) compared to standard treatment alone (control arm). The primary outcome measure of the trial was overall survival. The 1 year survival rate in the control arms was expected to be 25%. To detect an advantage of 5% in overall survival at 1 year (to 30%) a total of 2202 patients were randomised (1101 in each arm - 2110 in England and Wales and 87 in Scotland). Participants were followed up for a minimum of 2 years after the date of randomisation.

If a participant was lost to follow-up (i.e. no longer seeing their FRAGMATIC hospital doctor), the WCTU (now known as the CTR) contacted the participant’s GP to obtain information on the participant’s status. If needed and the participant had given the necessary consent, they were traced via the Medical Research Information Service at NHS Digital (under previous alternations). This was to ensure that any participant lost would still contribute to the analysis.

Recruitment to the trial commenced 31/07/2007 and closed 16/12/2011. Patient follow up, data cleaning and trial closure with research ethics were completed in 2014. For clarity, ‘trial closure’ refers the dates upon which the study was closed with REC and MHRA as per the definition of trial closure under the UK Medicines for Human Use (Clinical Trial) Regulations. MHRA (i.e. when the last patient had completed protocol treatment) ended on 08/06/2012. REC (date of last data capture, e.g. including long term follow up via NHS Digital) ended on 09/10/2014. Correspondence with both regulatory bodies usually continues beyond date of closure and notification of date of closure based on the final reporting requirements for each body. The latest report specifies the date of final analysis as 20/11/2014 (i.e. date database was closed to edit rights, equivalent to hard lock.) The study needs to be retained for a minimum of 15 years from the later of two closure dates - 09/10/2029 based on REC closure being the latter.


As FRAGMATIC is the largest randomised trial in lung cancer patients that addresses this important clinical uncertainty about how patients should be managed, its results are of great importance to national and international health professionals and researchers. The study data has been published in four publications to date as specified in the "Outputs" section.

Future data processing objectives of Cardiff University include completion of ongoing health economic analysis and further analysis of the data set to further investigate and publish other secondary outcome measures (e.g. VTE, Khorana score etc.) included in the original REC approval.

These ancillary studies include processing of the data of the two trial arms which may have different adverse event profiles and outcomes that might translate into different costs, effects and cost-effectiveness. This work will look at trial data collected about on the study case report forms about dalteparin received, other anticancer treatments, management of Severe adverse events (SAE), hospital admission episodes (all patient nights by speciality) and day case visits. It will also analyse data collected via a patient-completed EQ-5D questionnaire which enables scores to be converted into Quality Adjusted Life Years (QALs) to determine the cost per QALY in a Cost Utility Analysis (CUA) framework, and dyspnoea, anxiety and depression via a patient- completed Hospital Anxiety and Depression Scale (HADS) questionnaire. As death is a potential outcome of SAEs and SAEs are being analysed in the context of cost utility, this work might require further processing of mortality data supplied by NHS Digital.

Yielded Benefits:

The main publication concluded that there was no evidence of a difference in overall or metastasis-free survival between the two arms. There was a reduction in the risk of VTE in the LMWH arm and no difference in major bleeding events but evidence of an increase in the composite of major and clinically relevant non-major bleeding in the LMWH arm. LMWH did not improve overall survival in the patients with lung cancer in this trial. A significant reduction in VTE is associated with an increase in clinically relevant non-major bleeding. Strategies to target those at greatest risk of VTE are warranted. These results were of great importance to health professionals and researchers not only in the UK but around the world as heparin clearly does not give a clinically significant benefit to newly diagnosed lung cancer patients. The publication has been received 44 citations prior to 24/04/2018. The two follow up publications concluded that although no survival difference was seen from the addition of dalteparin, there was a significant reduction in the incidence of radiologically confirmed venous thromboembolism (VTE) in both tumour types for those patients randomised to dalteparin. However, this was at the expense of an increase in clinically relevant non-major haemorrhage, thereby making routine thromboprophylaxis difficult to justify. Interestingly, of 910 lung cancer patients receiving chemotherapy and no dalteparin, 108 (11.9%) developed VTE, although the Khorana score had a limited utility in identifying those at greatest risk. Actual yielded benefits include: 1) Based on the results of the FRAGMATIC study the revised International Society on Thrombosis and Haemostasis (ISTH) Guidelines now recommend consideration of LMWH for VTE prophylaxis in ambulant lung cancer patients receiving chemotherapy. 2) Data has been published highlighting the poor utility of the Korhana score in lung cancer, based on FRAGMATIC data. 3) The limitation of the FRAGMATIC study, i.e the majority of patients being advanced cancers, led to the implementation of other research (TI|LT study) which has recently been reported.

Expected Benefits:

The FRAGAMATIC study was the largest randomised trial in lung cancer patients that addressed the important clinical uncertainty about how patients should be managed. The data has been used in three publications to date as described in the "Outputs" section. The results of these publications have been of great importance to health professionals and researchers and have contributed to an increased knowledge base of the management of patients with lung cancer in the UK and worldwide.

Future analysis will provide further information on the health economics and cost effectiveness of using fragmin in this patient set and may, therefore, influence national and international policy further.

The benefits of the McMaster research include scientific publication of analysis and use in clinical practice guidelines.

Processing:

Under this Agreement, the data may be securely stored and processed by Cardiff University. Velindre University NHS Trust will not store or process the data. No new data will be provided by NHS Digital under this Agreement.

The WCTU originally provided the following fields to NHS Digital under the original Agreement:
- Trial number (specific to FRAGMATIC),
- Surname,
- Forename,
- Initials,
- NHS number,
- Date of Birth, and
- Postcode.

NHS Digital then flagged this cohort and linked this information to Mortality data and returned outputs to Cardiff University. NHS Digital supplied death registration data (including registration district, sub district, number and entry number, date cause (text) and place of death (text), ICD coding, date and place of birth, occupation and address) and patient status (if alive and registered with a GP). This data included sensitive, patient identifying data. This information was used to assess survival in the two arms of the randomised clinical trial summarised above.

All FRAGMATIC data that was not sourced from NHS Digital or generated directly through non-routine trial-specific procedures was collected from routine and trial-specific medical notes at participating local sites via entry by staff onto trial specific paper case report forms (CRF). The only exclusion to this was data directly sourced from trial participants via questionnaires completed by the participant (e.g. EQ5D, HADs, Dyspnoea). All CRFs and questionnaires were submitted by sites to the WCTU (as it was known then) for entry onto a WCTU trial-specific database following standard clinical trial practice and under the remit of the FRAGMATIC immediate care facility, FRAGMATIC REC and MHRA approval, trial-specific contract between the participating site and Sponsor, and delegation of duties contract between the Sponsor and WCTU.

The data is stored in a safe data haven at Cardiff University within which all analyses continues to be undertaken. The dates of death of specific identified patients in the clinical trial who agreed that such information could be accessed were required.

The study findings have been published in four publications to date as specified in the "Outputs" section below.

Velindre University NHS Trust and Cardiff University have shared a subset of data with McMaster University in Canada for use in a very large individual patient meta-analysis of trials on the use of heparins in cancer patients. This contains a limited number of variables derived from the data supplied by NHS Digital. These include an indicator of whether individuals are alive or deceased and, if deceased, an indicator of whether the cause specific mortality was listed as lung cancer or ‘Other’. Date of death was not supplied but was used by Cardiff University to derive the number of days from date of recruitment to death. The date of recruitment was not supplied.

The following fields have been shared with McMaster University:
- Age at randomisation (years)
- Gender (Male / Female)
- Time in study - Time to death (Derived from Date of Death)
- Event (Fact of Death)
- Cause Specific Mortality (Lung Cancer / Other)
- Type of VTE event (Derived from Date of Death)

No personal data supplied by NHS Digital has been shared in any form. No variables which could be used to identify a data subject have been shared.

NHS Digital has reviewed the specification and description of the data that was shared and is satisfied that the data is sufficiently derived in accordance with its definition of derived data in the Data Sharing Framework Contract. NHS Digital is content for McMaster University to have access to this derived data subject to the following conditions which ensure the data cannot be identified as originating from the data supplied under this Agreement:
i. McMaster University is permitted to process the data for the purpose of its meta-analysis only;
ii. McMaster University is not permitted to combine the data with other datasets which could potentially increase the risk of reidentification for individuals in the dataset;
iii. McMaster University must not attempt to re-identify individuals in the dataset;
iv. McMaster University must not onwardly share the dataset;
v. McMaster University must not publish the data.

Under the terms of this Agreement, Velindre University NHS Trust and Cardiff University are jointly responsible for ensuring compliance with the above conditions and for confirming destruction of the data by McMaster University once the data is no longer required for the purpose for which it was shared.

A protocol paper has been published for this study by McMaster University. The results of the McMaster study are yet to be published. The primary results of the McMaster study have been accepted for publication and are pending publication. Analysis of secondary objectives is already complete, and ready for submission to Lancet Oncology. Further information about publication outputs from McMaster that the FRAGMATIC study has contributed to are provided in the outputs section.

STEADFAST Modelling the associations between wider health and social characteristics and diabetes-related health - Substantive data flows to ONS — DARS-NIC-669962-W1F6D

Type of data: information not disclosed for TRE projects

Opt outs honoured: Anonymised - ICO Code Compliant, Yes (Does not include the flow of confidential data)

Legal basis: Health and Social Care Act 2012 – s261(2)(a)

Purposes: No (Academic)

Sensitive: Sensitive

When:DSA runs 2022-12-23 — 2023-12-22 2023.04 — 2023.04.

Access method: One-Off

Data-controller type: CARDIFF UNIVERSITY

Sublicensing allowed: No

Datasets:

  1. National Diabetes Audit

Objectives:

On 1 February 2023, NHS Digital merged with NHS England. NHS England has assumed responsibility for all activities previously undertaken by NHS Digital. The merger was completed by a statute change. Any reference made to NHS Digital within this Data Sharing Agreement is in reference to the merged organisation known as NHS England.

Cardiff University requires NHS Digital data for the research study: “STEADFAST - The personal cost of health conditions in childhood”.

Cardiff University’s overall aim is to quantify the links between educational outcomes and diabetes-related health outcomes, for example, how educational settings influence blood glucose levels and the time to onset of diabetes-related complications. This linkage of diabetes and education data for England and Wales arose from work by Cardiff University modelling whether rates of hospital admissions for young people living with diabetes were higher than for their peers without diabetes.

The purpose of this study is to provide a greater understanding of the interrelationship between diabetes-related health and education. The substantive motivation for this study is grounded in the evidence that most of the costs arising from diabetes come not from the day-to-day care and medications, but the complications arising from elevated blood glucose over the life course. There is currently limited evidence on the causes of less optimal diabetes management and the potential mechanisms for interventions to improve this. Thus, this study hopes to broaden the evidence base beyond the purely clinical factors to investigate the wider health and social determinants of diabetes-related health, using linked administrative data and focussing on education.

Children of school and university age with diabetes are most frequently living with type 1 diabetes, accounting for approximately 98% of cases, so the research initially focuses on this population. Education and health outcomes for children with type 2 diabetes (and other rarer forms of diabetes) are equally important however, and the analysis is replicated for each of these groups.

Under a linked Data Sharing Agreement (DSA), DARS-NIC-158283-T2Q2D versions 0-2, diabetes health data covering England and Wales was shared for people born from the 1983-1984 academic year to the 2009-2010 academic year. Data linkage of the diabetes health data with school and university records was completed for a subset of 2080 people who studied at a Welsh educational institution (school, university, college). Diabetes health data included those linked from the national paediatric diabetes audit data (controlled by the Healthcare Quality Improvement Partnership (HQIP)) for individuals born 1992-1993 to 2009-2010, and the adult national diabetes audit data (from NHS Digital) for individuals born 1983-1984 to 2001-2002.

This Agreement is a new request to obtain the equivalent diabetes health and education data for people living in England and Wales. This Agreement should be read in parallel with DARS-NIC-674735-Z0H6K-v0 and DARS-NIC-669808-V6T0M-v0. Each DSA covers a different element of the required data flows for this study. DARS-NIC-158283-T2Q2D-v2 is being retained to cover the historic data flows described above, and may be amended in future to request further diabetes audit years to be linked with Welsh educational records.

The substantive dataset requested from NHS Digital is the National Diabetes Audit (NDA) which provides information regarding diabetes-related health for all people with diabetes. NDA data for linkage in the Secure Anonymised Information Linkage (SAIL) databank was disseminated and processed under DARS-NIC-158283-T2Q2D versions 0-2. Data was disseminated to SAIL databank at the University of Swansea with a pseudonymised identifier only.

NDA data for people living in England and Wales is requested under this Agreement. The NDA data that is required relate to characteristics of the diabetes diagnosis (diabetes type, age at diagnosis) and measures of diabetes-related health (e.g. blood glucose levels, levels of protein in urine), and care processes (e.g. retinopathy screening for damaged blood vessels in the eye; foot exams to assess nerve or blood vessel damage). Diabetes-related health is primarily measured using blood glucose levels, directly or by using ‘HbA1c’, a proxy for blood glucose management control measured regularly as part of diabetes clinical management and recorded in the diabetes audit data. This data will be disseminated to the Office for National Statistics (ONS) – Secure Research Service (SRS) with a pseudonymised identifier only.

Cardiff University are also requesting that NHS Digital provide identifiable information of relevant individuals in the NDA to the organisations who hold their educational records.

Under the linked agreement, DARS-NIC-158283-T2Q2D versions 0-2, identifiable information for people with diabetes living in England and Wales was disseminated to Digital Health and Care Wales (DHCW), in order that the relevant primary and secondary education (provided to DHCW from the Welsh Government), and higher education (provided to DHCW from the Higher Education Statistics Agency (HESA)) records could be identified. The identifiers were provided alongside the same pseudonymised identifier provided to SAIL databank with the NDA data. The education data was made available for analysis by Cardiff University in a de-identified format in the SAIL databank, linked to the applicable diabetes health record via the pseudonymised identifier.

Under linked agreement DARS-NIC-674735-Z0H6K-v0, NHS Digital are requested to provide identifiable information of relevant individuals in the NDA who were living in England and Wales to the Department for Education (DfE) in order that their primary, secondary and higher education records can be identified. The identifiers will be provided alongside the same pseudonymised identifier provided to the ONS-SRS with the NDA data. The education data will be made available for analysis by Cardiff University in a de-identified format in the ONS-SRS, linked to the applicable diabetes health record via the pseudonymised identifier.

Under linked agreement DARS-NIC-669808-V6T0M-v0, NHS Digital are requested to supplement the identifiable information of individuals in the National Paediatric Diabetes Audit (NPDA), provided by the Royal College of Paediatric and Child Health (RCPCH), with their names. This would be done using NHS Digital’s ‘Demographics’ dataset. Names can then be used by the DfE alongside other identifiers supplied by the RCPCH to retrieve the relevant educational records of individuals in the NPDA. These would be made available for analysis by Cardiff University in a de-identified format, linked to the applicable diabetes health record.

The NDA and identifiable data from NHS Digital will, when combined with the associated education data, allow the study to model how characteristics of a person’s diabetes impact their education and, simultaneously, how their education affects their diabetes-related health.

Educational outcomes are recorded from the time a child enters school until they leave university, including measures of attendance, attainment, and broader characteristics of the educational experience such as special educational needs and school exclusions.

The mechanisms driving the relationships between education and health can be broken down into three pathways. Firstly, diabetes-related health may affect educational outcomes, for example, through biological mechanisms, including the effects of excess glucose on the brain structure and social mechanisms such as adjusting management routines to fit in with a university lifestyle. Secondly, education may affect diabetes-related health. For example, continuing education beyond compulsory schooling might provide structure and support that facilitate better management. Thirdly, individual characteristics (observed and unobserved) may directly affect both education and diabetes-related health, such as motivation and intelligence.

To help unpick which of these are happening in the data, Cardiff University uses repeated measures of educational outcomes and health outcomes for an individual to tease out the ordering of events. For example, if a person with less optimal blood glucose levels has high rates of school absence, it would be possible to look back and see if that person had high rates of absence before they were diagnosed with diabetes. In practice, the statistician will look at many thousands of individuals at once and consider many such differences simultaneously, but the principle is the same.

The data request covers people born from 01/09/1983–31/08/2002 with diabetes, living in England and Wales, who are included in the diabetes audits (NDA and/ or NPDA) from 2003 onwards. This is anticipated to include 30,000 patient records, with linkage anticipated for 28,000 individuals. Individuals born prior to this have been excluded since their full education data, including data on their first year of university, is not available. Individuals born after 2002 will be recorded in the NPDA, which is not provided by NHS Digital. As the study is primarily interested in the effect of diabetes during education, cases are restricted to those diagnosed with any form of diabetes prior to age 24 or younger. Whilst university cohorts are typically 18-21, due to the high prevalence of delays in starting time at university (gap year, changing university, changing course etc.) and the varying length of courses (sandwich degrees, placements, conversion courses), the request will include all ages up to the government definition of the end of youth education, i.e. up to age 24.

The predicted ages of school students include early years (nursery and reception) aged 3-5, compulsory schooling aged 5-16, and key stage 5 ages 16-18. Some students start school early or leave later than normal, however there will be no request for additional cohorts for this contingency and will instead make this an amendment if it transpires to be an issue.

The DfE do not hold data for children educated at private schools or educated home, thus for these cases (~7%) the diabetes audit data would not be assigned a linkage ID by DfE. Unlinked cases would not form part of the core analysis, however Cardiff University would carefully examine why different cases did not link. For example, there may be noticeable patterns in children who did not link for technical linkage reasons, such as lower matching rates for children from certain minority ethnic groups where the matching algorithm may not be as efficient for those names. Cardiff University would also consider if the diabetes health outcomes are different for those children who do not link, since that may motivate a change in monitoring, for example for children who are educated in settings other than at school (home educated).

There should be no NDA cases in the requested extract that will not meet the conditions of the study. It would be expected all cases to have attended school, and those that do not attend university are still important as controls for comparison with those that did. There will be no request for controls from NHS Digital.

Although linkage with Welsh educational data was undertaken under DARS-NIC-158283-T2Q2D versions 0-2, data is requested for individuals resident in both England and Wales for the current request because individuals can be located across borders for their medical care or educational settings, and individuals move across borders over time. The study will combine estimates from data linked to Welsh educational records in the SAIL databank, and data linked to English educational records in the ONS-SRS, to create a single estimate for England and Wales.

Cardiff University request individual-level data for modelling changes over time for individuals, e.g. how individual differences in education experiences affect diabetes management and vice versa. Cardiff University requests identifiers only for the purpose of data linkage; however, the analysis will be carried out on de-identified data.
Cardiff University request data from 2003/04 to 2017/18 to best model trajectories of HbA1c and the time to the first onset of early complications.
Cardiff University requests data for England and Wales to ensure maximum power and generalisability of the results.

Cardiff University confirm there are no alternative, less intrusive ways of achieving the purpose. A high degree of data linkage is required for this study in order to follow the entire life course trajectories of both diabetes related health and education through the combined linkage of both the paediatric and adult diabetes audits, and education data from primary school to university. Patient identifiers are sent to different organisations than the substantive clinical data to minimise the risk of disclosure.

The request has been restricted to only the essential clinical measures and associated metadata such as date of measure, location of measure (clinic).

Cardiff University are the research Sponsor and sole data controller for this DSA, who also process the data.

The personal data for this study is processed under UK GDPR Article 6(1)(e) - Public Task - for academic medical research carried out as a task in the public interest. The processing is necessary for Cardiff University (as a public authority for the purposes of data protection legislation) to perform a task in the public interest. The task has a clear basis in law.

The special category personal data for this study is processed under UK GDPR Article 9(2)(j) for 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 right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject. The public interest lies in the improved evidence on the links between educational measures (school settings, absence, attainment) and health measures (HbA1c levels) for young people with diabetes. The study aims to improve care for all young people with diabetes in the education system by informing clinicians and commissioners of variation and outcomes and complications to support work to improve and standardise treatment selection choices. For these reasons, the processing also meets the conditions of Schedule 1 Part 1 paragraph 4 of the Data Protection Act 2018.

Cardiff University has approval from the Health Research Authority for data providers to set aside the common law duty of confidentiality with respect to the sharing of person identifiers in health datasets, known as a ‘Section 251 exemption’.

SAIL and DHCW are data processors for the Welsh subsets of data as described above. The DfE and ONS are data processors for the English subsets of data as described above.

The Welsh Government and HESA provide education data which is linked with NHS Digital data.

The RCPCH provide NPDA identifiable data to NHS Digital, and NPDA substantive data to the ONS-SRS. The Healthcare Quality Improvement Partnership (HQIP) are the data controller of the NPDA.

The Welsh Government, HESA, RCPCH and HQIP do not process NHS Digital data.

None of the organisations listed above, except Cardiff University, determine the purposes or the means of the processing of NHS Digital data. They are not therefore considered joint data controllers of NHS Digital data under this DSA.

The Medical Research Council (MRC) and UK Research & Innovation (UKRI) have provided funding for this study. The funders do not determine the purposes or the means of the processing of NHS Digital data and do not process the data. They are not therefore considered joint data controllers or data processors of NHS Digital data under this DSA.

Yielded Benefits:

The study has been lucky in terms of the timings of the opportunities to contribute to the policy debate. Cardiff University were able to provide evidence, based on analysis of the Welsh data, to the new Additional Learning Needs Bill for Wales and the Special Educational Needs and Disabilities Green Paper in England, though these are yet to report their findings so have not yet resulted in any change in practice. To support public involvement and engagement, the study team has given presentations of the results obtained so far to practitioners (e.g., the Royal College of Paediatrics and Child Health annual conference, the Brecon Annual Meeting of type 1 diabetes practitioners in Wales), patients (e.g., the Yorkshire and Humber Young People’s Diabetes Network Annual Meeting), and researchers (The Type 1 Diabetes Consortium at the Diabetes UK conference). The findings have been well received by clinicians, who report that they feed back some of the findings, particularly the ‘good news story’ that overall, children with diabetes attain in school as well as their peers without diabetes. Practitioners have also reported that the quantification of expected absence (average and range) has also been useful to provide a benchmark for children with diabetes.

Expected Benefits:

Although laws exist in England and Wales to support people with health conditions in schools and universities, the legislation (and associated guidance and resulting practice) does not always meet the child's needs. In schools, there is a gap between support for special needs (which stems from the Children and Families Act 2014 and traditionally focused on support for ‘learning conditions’, e.g. dyslexia) and support for children with health conditions (which stems from the Equalities Act 2010 and traditionally focuses on support through reasonable adjustments). In universities, students lack clarity on how financial support for health conditions works, particularly for conditions like diabetes, where there is wide variation in the required amount and type of support. Students will also be adjusting to having less clinical support than they are accustomed to, having only recently transitioned from paediatric to adult services before moving to a university campus far from their ‘home’ clinical team.

The high-level goal of this study is to ensure that children living with diabetes (and their families) can (i) manage their condition to stay healthy and (ii) fulfil their potential in education. By combining health data (including HbA1c) and education data (including educational outcomes and support), the project hopes to generate robust evidence of variation in outcomes, which may motivate changes in legislation and practice. Although the primary route for an impact on health is through policy change (and the associated improvement in guidance and implementation); the project will continue to inform practitioners of the range of outcomes and thus show the scope and potential benefit from improved support, both in terms of educational outcomes and diabetes-related health outcomes, and the link between the two.

Study outcomes for children in schools: This will focus on how rates of absence and attainment trends are associated with diabetes management. The longitudinal (annual measures) nature of the data means that researchers can compare outcomes over time; for example, annual measures of school absence can help show the changes in missed schooling before vs after a diagnosis of diabetes, in addition to comparing absence rates with similar children who were not diagnosed with diabetes. Similarly, regular measurements of HbA1c allow researchers to identify changes in blood glucose management as a child moves from primary school to secondary school. If differences exist, clinical teams may choose to give additional support based on non-clinical (educational) challenges that may influence diabetes management.

Study outcomes for children leaving school: Students with diabetes who are leaving compulsory schooling will transition to A-levels, vocational courses, or the labour force, and the study will model how these choices subsequently impact the trajectories of HbA1c. For example, the study may find that those who choose not to continue formal education beyond age 16 have less structure, affecting their diabetes management. If differences exist, clinical teams may choose to provide extra support or delay the transition from paediatric to adult diabetes care until after the transition to employment is complete.

Study outcomes for higher education students: The work focusing on universities will look for changes in trajectories of diabetes management, for example, comparing those that live at home with those who move away to attend university, identifying whether there is a significant change in management and the timings of that change. Differences may be seen in HbA1c during the first term as students try to reconcile good management routines with university social life, or perhaps HbA1c may not alter until the stress of the final year of university with impending high-stakes exams. If differences exist, clinical teams may choose to be more proactive in supporting the transition between home and university care and arrange follow-up appointments, both at home and at university, scheduled at times that fit around academic term dates.

Study outcomes for young adults: The research on young adults focuses on how childhood HbA1c levels and educational outcomes determine early adult health outcomes. This study focuses on the ‘double whammy’ effects of a history of sub-optimal HbA1c levels and lower educational outcomes. Relevant determinants might include individual factors (e.g. special educational needs or behavioural issues), school factors (e.g. lack of appropriate support), or family-level factors (e.g. socioeconomic status). If differences do exist, clinical teams may choose to flag individuals earlier who are struggling with diabetes management and education as being particularly vulnerable and receive additional support in the adult care setting, with potential health and social benefits to the individuals (who are most likely to develop early diabetes-related complications).

In partnership with Diabetes UK the project held a series of public workshops which reported the expected measurable benefits from the perspective of patient and public involvement participants (https://www.adruk.org/fileadmin/uploads/adruk/Documents/Diabetes_Education_Public_Workshop_Report_Aug_2021_01.pdf), and included: more understanding by others of what they are going through, more support with managing diabetes in schools, and earlier intervention at times of high stress. There was a focus on diabetes-education interactions during transition from paediatric to adult diabetes care – this comes amongst many other stressful events that teenagers go through, not least of which were the GCSE exams.

Participants focused on the benefit of bridging the gap in educational experiences between children with type 1 diabetes versus those without – “ensuring an equal playing field”. Participants reported a lack of knowledge around the impact of type 1 diabetes on education, which they felt was likely to be large. They felt that a study of this kind could influence policy, and ultimately the education setting, which in turn could result in a better experience for young people. A related point was the hope for increased general awareness and understanding of how type 1 diabetes might affect individuals beyond direct medical outcomes.

Outputs:

The expected outputs will be academic journal papers modelling the associations between education outcomes and trajectories of diabetes-related health.
The nature of data-based outputs would be descriptive statistics and regression coefficients, all of which would be aggregate data with small numbers (less than 10) suppressed; this is a requirement of taking any results out of the ONS-SRS and is rigorously checked by ONS analysts.

The original aim was to finish the draft of the outputs using data in SAIL disseminated under linked agreement DARS-NIC-158283-T2Q2D versions 0-2 between 21/01/2019 to 21/08/2020, though the first of these papers was only accepted for publication on 01/08/2022 by the journal ‘Diabetes Care’. It is likely that future analysis will combine estimates from data deposited in both SAIL databank and ONS-SRS, and is expected to be published before 2025.

Findings from the first paper have been presented at an invited address at the Royal College of Paediatric and Child Health annual conference and a presentation to clinicians at the Brecon Group Annual meeting.

The first research paper is based upon how measures of diabetes-related health influence educational trajectories, and shows that (i) children with diabetes have higher absence rates, but similar attainment and progression to university as their peers and (ii) HbA1c tracks with educational outcomes. Cardiff University provide tentative evidence that this link is likely to be the result of external factors affecting both education and health outcomes. Due to issues with the NPDA data, Cardiff University have not yet been able to undertake any analysis pertaining to how education settings affect HbA1c trajectories, or how external factors affect both diabetes-related health trajectories and educational trajectories.

Any further study findings will be shared with clinicians through academic publications (e.g., Diabetes Care) and presentations at national forums (e.g., Diabetes UK Conference & Brecon Scientific Meeting). The findings will be shared with the public more broadly using a plain English version of the study; this will hopefully be published in the diabetes audit report and presented by the project lead at patient forums. Further engagement work with the results will be coordinated with Cardiff University’s coproduction group of young people.

An agreement is in place with Royal College of Paediatric and Child Health to include a patient level report on associations between diabetes related health and education for their annual report. Updates and summaries will be published on the project website (https://www.cardiff.ac.uk/research/explore/research-units/childhood-health-and-education?).

A dissemination plan has been produced in partnership with Diabetes UK. This includes two papers supported by plain English reports produced by the Diabetes UK policy team. The research team and Diabetes UK have begun to meet with policymakers from England and Wales, but it is too early to say how these connections will be exploited to communicate results. All outputs will belong to Cardiff University but will be freely available to Diabetes UK to use as agreed under Cardiff University’s heads of terms. Information about how the study has already consulted young people are provided on the webpage for the first 2021 patient and public engagement workshop: https://www.adruk.org/our-work/the-personal-cost-of-health-conditions-in-childhood-engaging-the-public/public – ADR UK and the workshop report: https://www.adruk.org/fileadmin/uploads/adruk/Documents/Diabetes_Education_Public_Workshop_Report_Aug_2021_01.pdf

In the second workshop, the young people produced a video explaining how researchers use data relating to young people with diabetes. Although the video and report produced are not yet released, there is a holding version on YouTube: https://www.youtube.com/watch?v=oimfnSoENxo.

Several of the representatives from Cardiff University’s patient and public involvement sessions have become the coproduction group, helping steer the project and the next iteration of the public engagement.

Processing:

This study uses a split file process to transfer individual-level data. The objective is to ensure that, as far as possible, every person’s health data is processed separately from the identifiers that would link that health data to a person.

Each clinical data provider assigns a study-specific pseudonymised identifier (study ID) to each participant. They then split the whole dataset into an identifiers dataset (containing variables such as NHS number, name, date of birth, postcode, and gender) and a substantive dataset (containing de-identified clinical/ education data).
The substantive data is transferred directly to the repository.

The identifiers file is shared with a trusted third party (details below) who uses the identifiers to match individuals’ ‘study ID’ (pseudonymised identifiers) to their ‘linkage ID’, and then deletes all the real-world identifiers such as names and dates of birth, before transferring the ‘study ID’ and ‘linkage ID’ into the repository where the ‘study ID’ enables re-joining of the ‘linkage ID’ to the substantive data, and the ‘linkage ID’ enables linkage to the other datasets which have been processed in the same way.

Under linked agreement DARS-NIC-158283-T2Q2D versions 0-2, the data flows were as follows:

NHS Digital identified people born from 01/09/1983 to 31/08/2002 (ie the 1983/84 to 2001/02 academic years) who appeared in the National Diabetes Audit (NDA) dataset and assigned them a unique study ID.

From this cohort, NHS Digital provided a one-off drop of data covering 2003/04 to 2017/18 audit years, containing the substantive diabetes-related data, accompanied by study ID only, to the Secure Anonymised Information Linkage (SAIL) databank at the University of Swansea. This file contained no identifying information other than the study ID.

From this cohort, NHS Digital also provided a one-off drop of data covering 2003/04 to 2017/18 ‘audit years,’ containing study ID, patient name, NHS number, date of birth, gender, and postcode only to Digital Health and Care Wales (DHCW).

DHCW created the ‘linkage ID’, a hashed version of the NHS Number, referred to by DHCW and SAIL as the Anonymised Linkage Field (ALF), then destroyed all the identifying information (name, NHS number, date of birth, gender, postcode). After this process, the identifiers file contains only the study ID and linkage ID, along with less disclosive versions of the demographic data (gender, week of birth, lower super output area). This file was onwardly flowed to SAIL.

Within SAIL, the substantive NDA file was re-joined to the identifiers linkage field file using the study ID so that each case of the substantive data got a linkage ID.
The same process as outlined above for NDA data also happened for the other de-identified datasets that were placed in SAIL, including the National Paediatric Diabetes Audit (NPDA), Higher Educational Statistics Agency dataset (HESA), and Welsh Dataset Education Records (WED, also known as the National Pupil Database or NPD). The NPDA, HESA, and WED data were linked to the NHS Digital NDA data using the linkage ID.

Data accessible to Cardiff University via SAIL therefore are:
- NDA data covering England and Wales
- NPDA data covering England and Wales
- HESA (higher education) data covering England and Wales
- Primary and secondary education data for Wales only (WED)

Under linked agreement DARS-NIC-674735-Z0H6K-v0.0, NHS Digital identify individuals based in England and Wales and born between the 01/09/1983 to 31/08/2022 (ie the 1983/84 to 2001/02 academic years), in the NDA dataset under DARS-NIC-669962-W1F6D and assign them a unique study ID.

Using the NHS numbers held in the NDA, NHS Digital retrieves the names, date of birth, latest postcode, and gender from the Personal Demographics Service (NHS number was included for DHCW, but is not used here as Department for Education (DfE) are unable to process it).

These identifiers are sent to the DfE alongside the unique study ID. This will be a one-off deposit of data, though permission for annual refreshes of data will be sought in the future. DfE, acting as the trusted third party, will use the identifiers to match individuals’ ‘study ID’ to the ‘linkage ID’ (referred to by DfE and the Office for National Statistics (ONS) as the Pupil Matching Reference (PMR)), then destroy all the identifying information (name, date of birth, gender, postcode). After this process, the identifiers file contains only the study ID and linkage ID. (Unlike for the SAIL process, less disclosive versions of the demographic data [gender, week of birth, lower super output area], are not retained in the identifiers file). This modified identifiers file, containing only study ID and linkage ID is onwardly flowed to ONS-SRS repository where the ‘study ID’ enables re-joining of the ‘linkage ID’ to the substantive data, and the ‘linkage ID’ enables linkage to the other datasets which have been processed in the same way.

In parallel, under this Agreement (DARS-NIC-669962-W1F6D-v0), substantive de-identified NDA health data for the 2003/4 to 2017/18 audit years for the above-described cohort will be sent to the ONS Secure Research Service (SRS), with the unique study ID only. This will be a one-off deposit of data, though permission for annual data refreshes will be sought in the future.

Under linked agreement DARS-NIC-669808-V6T0M-v0, NHS number, date of birth, postcode and gender will be provided to NHS Digital by the Royal College of Paediatric and Child Health (RCPCH) for individuals born in the academic birth cohorts from 1983/4 to 2001/2 who are in the NPDA. NHS Digital will retrieve the names of these individuals and send identifiable Demographics data (names, date of birth, latest postcode, and gender) to the DfE alongside a unique pseudonymised identifier provided by the RCPCH. This will be a one-off deposit of data, though permission for annual data refreshes will be sought in the future.

The project includes four further data flows which are not covered by the above-listed DSAs:

1. RCPCH send substantive deidentified NPDA data directly to the ONS-SRS, containing the same unique pseudonymised identifier as provided to NHS Digital under DARS-NIC-669808-V6T0M-v0
2. DfE send substantive deidentified education records (compulsory education) directly to the ONS-SRS, alongside the linkage ID
3. HESA send substantive deidentified education records (higher education) directly to the ONS-SRS, alongside the linkage ID.
4. After the DfE ‘hashing service’ use the identifiers supplied by NHS Digital to identify the correct individual against their records and retrieve their PMR, the original study ID supplied by NHS Digital, alongside the PMR retrieved by the DfE, are onwardly flowed into the ONS-SRS.

DfE will destroy the real-world identifiers as soon as the PMR field is retrieved and the necessary pseudonymised identifiers provided to the ONS-SRS. They will provide NHS Digital with a data destruction certificate.

ONS-SRS staff will work with the Cardiff University project lead to link the data records within the ONS-SRS environment. Substantive NDA and NPDA data will be linked with a PMR via the original study IDs, where the DfE have been able to retrieve a PMR.

NDA, NPDA, and education data will then be linked using the PMRs.

ONS-SRS staff will support Cardiff University with the linkage evaluation for the extract to be used for research, determining what percentage of clinical records have been successfully linked together (NPDA to NDA) and linked to associated education records (NPDA to NPD, NPDA to HESA, NDA to NPD, NDA to HESA).

Unlinked clinical data will remain in the ONS-SRS environment but will not be further processed for the purposes of these Data Sharing Agreements (DARS-NIC-674735-Z0H6K-v0.0, DARS-NIC-669808-V6T0M-v0, and DARS-NIC-669962-W1F6D-v0). Other linkage mechanisms may be explored in future to improve the success of the linkage, and data on individuals not in school or university may give rise to important information for diabetes management.

The linked deidentified datasets will be made available for multilevel modelling analysis of the associations between education and health by Cardiff University on the ONS-SRS. The DfE will provide educational data for controls who do not have diabetes, alongside the educational data for those with diabetes, to enable Cardiff University to model the differences in outcomes.

The combining of health data with education data increases the likelihood that a person may appear as unique in the dataset; however, given that there are 30,000+ cases of young people with diabetes in England and Wales, and the high-level nature of the variables, it is unlikely that this dramatically increases the risk of reidentification. There will be no requirement or attempt to reidentify individuals for the purposes of the study or any other reason. The primary protection against reidentification is the creation of the linkage ID along with the destruction of the real-world identifiers prior to the data being put in the repository, meaning the data accessed by Cardiff University is deidentified. Secondly, ONS-SRS and Cardiff University will check data at the outset for any identifiability risks in the raw data or linked data before using it for analysis. This checking is routinely conducted by experienced analysts at ONS who use a mature process to identify small numbers, unique/rare cases, or other identifiability risks. Thirdly, researchers from Cardiff University who are accessing the data are trained not to re-identify data.

NHS Digital data is not being linked to any publicly available data.

Data processing is only conducted by substantive employees of the data processors who have specific teams and staff trained and explicitly employed for this purpose. Data will be accessed for analysis through secure remote gateways into the ONS-SRS. No data will be accessed outside the UK. Person identifiers (such as name and date of birth) will not be shared with third parties beyond those required to create the pseudonymised linked datasets (i.e. only DfE).

Technical infrastructure:
The data, including data originating from NHS Digital, are held and processed in two separate environments at ONS.

Environment 1: Data security for storage and linkage of the data will be provided with an assured ONS data analysis environment (DAP) that includes the following elements of security control:
- Need To Know applied through user account access and management
- Controlled ingest and export of data into/out from the environment
- Controlled account access using unique credentials based on job role
- Logged and monitored access of user activity within the environment
- Secure build configuration for infrastructure, including cloud services
- Vulnerability tested infrastructure with appropriate remediation and patching
- Compliance checks against security enforcing controls
- Architectural review against standards and best practice
- Staff security cleared to the appropriate level based on their supervised and/or unsupervised access to sensitive data in accordance with ONS clearance policies and data access processes
- Education and awareness of environment users covering security policies and secure working practices
- Operational support processes to securely manage the environment
- Risk assessment to identify security risks and mitigation actions to reduce this risk. ONS employs rigorous disclosure controls and access restrictions ensuring that physical, technical, procedural and personnel security is kept to the highest and most up to date standards. Following policy specified by the ONS Chief Security Officer, users will be granted supervised or unsupervised access following a clearance application. The ONS Information Asset Owner (IAO) grants access and a list of all authorised users is available on request.

Environment 2 is the Secure Research Service (SRS). The SRS gives accredited researchers secure access to pseudonymised, unpublished data in order to work on research projects for the public good. The SRS has been accredited as a Digital Economy Act (DEA) 2017 processor by the UK Statistics Authority for the preparation and provision of data for research purposes. The SRS runs on cloud-based infrastructure provided by UKCloud. UKCloud is included within the data processor section of this Agreement. Security assurance documentation for UKCloud has been supplied to NHS Digital and approved by the NHS Digital Security Team. All implementation and maintenance services related to the SRS’s use of UKCloud infrastructure are provided by Equiniti. Equiniti is included within the data processor section of this Agreement.

Public engagement focused on both the research questions and the data processing has been carried out throughout the project, going back to 2014 before the study was funded. The largest public involvement work has been the sessions with young people with type 1 diabetes in 2020, in partnership with Diabetes UK and the MRC Regulatory Support Centre. This work is summarised in a website (https://www.adruk.org/news-publications/news-blogs/public-views-on-the-use-of-personal-identifiers-for-linking-diabetes-and-education-data-for-research-439/) with links to the full report towards the bottom of the page. Further public engagement was carried out in 2022 across 19 focus groups, the link to the work is https://dareuk.org.uk/sprint-exemplar-project-steadfast, and the report summarising this work should be released in early 2023.

Modelling the associations between wider health and social characteristics and diabetes related health — DARS-NIC-158283-T2Q2D

Type of data: information not disclosed for TRE projects

Opt outs honoured: Yes - patient objections upheld, Identifiable, Yes (Section 251 NHS Act 2006)

Legal basis: Health and Social Care Act 2012 – s261(7), Health and Social Care Act 2012 – s261(7)

Purposes: No (Academic)

Sensitive: Sensitive, and Non-Sensitive

When:DSA runs 2019-03-01 — 2022-02-28 2020.04 — 2020.06.

Access method: One-Off, Ongoing

Data-controller type: CARDIFF UNIVERSITY

Sublicensing allowed: No

Datasets:

  1. National Diabetes Audit

Objectives:

Cardiff University requires National Diabetes Audit Data for use in a research project, ‘Investigating interrelationship between diabetes and children’s educational achievement’, which aims to model the association of the relationship between wider health and social characteristics and diabetes related health and better understanding the effects of educational trajectories, experiences and outcomes on diabetes management. The background to this study is the evidence that most of the costs arising from Type 1 diabetes come not from the day to day care and medications but from the complications arising from poor management over the life course, yet the causes of poor management are poorly understood. An aim of this study is to broaden the search beyond the purely clinical setting to investigate the wider health and social determinants of diabetes related health, using linked administrative data and focussing on education.

The project aims to quantify the links between diabetes related health trajectories and educational trajectories. Diabetes related health is measured using HbA1c, a proxy for blood glucose management control which is measured regularly as part of diabetes clinical management and recorded in the diabetes audit data. Educational outcomes are recorded from the time a child enters school until they leave university, including measures of attendance, attainment, and broader characteristics of the educational experience such as special educational needs and school exclusions.

The causality in this relationship is complex. It is hoped that three potential causal relationships will be identified. Firstly, diabetes related health may affect educational outcomes, for example through biological mechanisms including the effects of excess glucose on the brain and social mechanisms such as adjusting management routines in order to fit in with a university lifestyle. Secondly education may affect diabetes related health, for example staying on in education beyond compulsory schooling might provide structure and support that facilitate better management. Thirdly there may be individual specific characteristics that directly affect both education and diabetes related health, for example motivation and intelligence.

The processing of the data under this Data Sharing Agreement involves the following organisations:
- Cardiff University is the data controller and will perform analysis of the linked data and publish the findings;
- North NHS Wales Informatics Service (NWIS) is a data processor for Cardiff University and will receive, link and pseudonymise the data through the creation of a hashed linkage field before passing it on to SAIL;
- The Secure Anonymised Information Linkage Centre (SAIL) at the University of Swansea is providing secure facilities to store the datasets as linked datasets using the hashed identifier provided by NWIS so that it can be remotely accessed by an individual from Cardiff University for the purpose of analysis.

The Welsh Government (as the data controller for the schools and college data) and HESA (Higher Education Statistics Agency) will supply educational data to NWIS which will be linked with the data from NHS Digital. Neither organisation will process data supplied under this Agreement.

Cardiff University is the sole data controller for the data under this Agreement and for data from other sources to which the data will be linked for the purpose of this study. The data will be processed by Cardiff University alongside NHS Wales Informatics Services (NWIS) and Swansea University Secure Anonymised Information Linkage (SAIL) Databank. The Medical Research Council (MRC) are funding the work to undertake research such as this but cannot access data nor has a role in analysis or interpretation.

This work is research in the public interest as it aims to improve care for all patients considering undergoing this type of process – informing clinicians and commissioners of variation and outcomes and complications to support work to improve and standardise treatment selection choices.

Yielded Benefits:

The data is still being analysed and it is therefore too early to report any benefits.

Expected Benefits:

The outcomes of the study will provide clinicians with a robust quantification evidence for the links between education factors and diabetes-related health (defined using HbA1c measures from the diabetes audit data) for both children, young people and adults. The findings will be shared with clinicians both through academic publications such as BMJ, Lancet Diabetes Endocrinology, Diabetes Care and through presentations at national forums, Diabetes UK & Brecon Scientific Meeting. The team have already presented the planned work to a national meeting of diabetes clinicians and the response was that such evidence would be useful to help them include educational factors in the way they differentiate practice.

The work on children in schools will focus on how rates of absence and attainment trends are associated with diabetes management. Whilst it is expected those children, who have always had poor attendance or attainment (both before and after diagnosis) to have worse diabetes management, the study will additionally differentiate between those who experience larger changes in attendance or attainment following diagnosis. The applicant will also consider if the characteristics of schools are associated with management, and whether there are changes during the transitions between primary and secondary schooling. Due to changes in legislation and the improvement in communication between diabetes nurses and school staff, there are robust links that can be exploited to seek feedback from teachers where issues over attendance or attainment that can be used in clinical care.

The work around the period of transitions from compulsory schooling to A-levels, vocational courses, or the labour force will help inform how these choices subsequently impact the trajectories of HbA1c. For example, the study may find that those who choose not to continue formal education beyond the age of 16 have less structure and see a worsening of diabetes management. Clinicians could provide extra support or delay transition from paediatric to adult diabetes care until after the transition to employment is complete.

The work focusing on universities will look for changes in trajectories of diabetes management, for example, comparing those that live at home with those who move away to attend university, if there is a significant change in management and the timings of that change, perhaps in the first term as students try to reconcile good management routine with university social life or perhaps it may not alter until the stress of the final year with impending high stakes exams. Clinicians maybe more proactive in supporting the transition between home and university care and arrange follow-up appointments both at home and university at times that fit with any sensitive periods and academic terms.

The work on adult outcomes identifies characteristics of the educational history that predict poor adult control and propensity to earlier complications of diabetes. The team expect those with more years of schooling to have better control and seek to identify whether the wider characteristics of the school experience are associated with worse control later in the life course, these might include school factors such as special educational needs or behavioural problems, or factors that interact with education such as socioeconomic status. The benefits to practice are on better identification of the important non-clinical risk factors for poor adult diabetes management to support interventions to prevent costly complications.

Outputs:

The expected outputs will be academic journal papers modelling the associations between education outcomes and trajectories of diabetes related health. The nature of data-based outputs would be descriptive statistics and regression coefficients, all of which would be aggregate data with small numbers suppressed, this is a requirement of taking any results out of the SAIL databank and is rigorously checked with any outputs from analysis in the secure gateway having to be screened by a SAIL analyst.

The aim is to finish the draft of the outputs before 21/01/2019 21/08/2020. The findings will be shared with clinicians both through academic publications such as BMJ, Lancet Diabetes Endocrinology, Diabetes Care and through presentations at national forums such as the Diabetes UK & Brecon Scientific Meeting.

There will be a plain English version available to the audit for sharing more broadly and the applicant will present results at patient forums in Cardiff and Swansea.

Processing:

The request involves the disclosure of confidential patient information from the National Diabetes Audit (Adults – England), National Diabetes Audit (Adults – Wales) (both held by NHS Digital) and the National Paediatrics Diabetes Audit (held by the Royal College of Paediatrics and Child Health) to NHS Wales Informatics Services (NWIS).

In practice, NHS Digital will produce an extract of relevant variables and cohorts (based on education cohort availability) for patients born after 31/08/1984 from the NDA data. A new ‘system ID’ (AKA project specific ID) will be generated for each case in this extract and used exclusively for re-joining the two ‘split files’ , the split file process is used to avoid flowing identifiers alongside substantive data.

The NDA extract will be split into two halves i.e. the split files:

The first split file is the ‘identifiers file’ and will contain the system ID plus the identifying information, in the case of the NDA extract this will be the patient name, NHS number, date of birth, gender and postcode. This file will have no substantive diabetes data.

The second split file is the ‘substantive file’, and will contain the system ID, plus the substantive diabetes related data. This file will have no identifying information other than the system ID.

The file containing identifying information will be sent by NHS Digital to NWIS. NWIS will create a hashed version of NHS Number, referred to as the Anonymised Linkage Field (ALF), then destroy all the identifying information (NHS number, date of birth to nearest day, postcode). After this process, the identifiers linkage field file will contain only the system ID and linkage field (ALF), along with less disclosive versions of the demographic data (gender, week of birth, lower super output area). This version will then be flowed to SAIL.

The substantive file is sent directly to SAIL.

Once the substantive file and linkage field file are within SAIL, the substantive file is rejoined to the linkage field file using the system ID, this is so that each case of the substantive data gets a linkage field (ALF). It is the linkage field (ALF) that is used to link data within SAIL.

The same process as outlined above for NDA data, also happens for the other datasets that are placed in SAIL. Datasets can be linked in SAIL using the linkage field (ALF). No identifiable information is ever put into SAIL. Other datasets that will be brought into SAIL and linked to NDA data are National Paediatric Diabetes Audit, Higher Educational Statistics Agency dataset (HESA), and Welsh Dataset Education Records (WED, AKA National Pupil Database NPD).

Alternatives to these data flows are not viable to achieve the amount of linkage required for this study. Single datasets could be flowed, for example HESA data could be shared with NHS Digital to look at the effect of higher education on adult diabetes trajectories, however single flows would only enable exploration of one substantive area, whereas the objective of the study is to follow the entire life course trajectories of both diabetes related health and education through the combined linkage of both the paediatric and adult audits and education data from primary school to university.

Data is flowed using the split file technique to minimise the risk of disclosure, which means there are two streams of data, one for the identifiers and a second for the clinical data. The level of data is individual for all flows of data leading to the research dataset used for analysis in SAIL. All outputs arising from the research in the databank for publication are small number suppressed and aggregated.

Only a single researcher from Cardiff University will be accessing the linked dataset in SAIL. There will not be anyone who is not a direct employee of the organisation requesting data (e.g. users working on honorary contracts, charity staff helping with analysis, etc.) who will be accessing data, no data will be accessed outside the UK. Data will not be shared with third parties beyond those required for the creation of the pseudonymised linked dataset used for analysis in the SAIL databank (i.e. only North Wales Informatics & SAIL). The data will be used for multilevel modelling of HbA1c and education.

There should be no NDA cases in the requested extract that will not meet the conditions of the study. It would be expected all cases to have attended school, and those that do not attend university are still important as controls for comparison with those that did.

Data minimisation has been discussed at length with NHS Digital’s National Diabetes Audit Team and the request has been restricted to only the essential clinical measures and associated metadata such as date of measure, location of measure (clinic). The team have also restricted the participants to only those birth cohorts that would also be included in the linked administrative education data.

The NDA data extract has been designed to only include data of individuals that meet the conditions of the study. This excludes:
• Individuals born prior to 1979 since their education data is not currently available;
• Individuals born from 1979 to 1983 since the data on their first year of university is not available;
• Individuals born after 2002. Data will be collected about individuals born between 2003 and 2013 but not from NHS Digital. Those individuals should still be in paediatric care and recorded in the NPDA.

NDA cases born 1984 – 1996 have higher education data for the most common ages at university, in addition the later of these cohorts have some schools data, including key stage 4 and key stage 3 assessments, these cohorts will form the core of the analysis, all of these cohorts are requested for analysis.

NDA cases born 1997 – 2002 have sufficient university or schools data to make inference on the associations between diabetes related health and education, though unlike the core cohorts there will not be a record of whether they completed university.

For these cohorts the study is primarily interested in the effect of diabetes during education and thus restricting cases to those diagnosed with any form of diabetes prior to age 24 or younger.

The predicted ages of school students includes early years (nursery and reception) aged 3-5, compulsory schooling aged 5-16, and key stage 5 ages 16-18. Some students start school early or leave later than normal, however there will be no request for additional cohorts for this contingency and will instead make this an amendment if it transpires to be an issue. University cohorts are typically 18-21, however due to the high prevalence of delays in starting time at University (gap year, changing university, changing course etc.) and the varying length of courses (sandwich degrees, placements, conversion courses) and will include all ages up to government definition of the end of youth education, i.e. up to age 24.

The following items of confidential patient information are required as part of the NDA extract to be supplied by NHS Digital for the purposes of creating the linkage field:

> NHS number: Used to create Anonymised Linkage Field (ALF).
> Date of birth: Linkage validation (also translated to week of birth for analysis).
> Gender: Linkage validation (also used for analysis).
> Postcode. Linkage validation (also translated to LSOA for analysis).

The researcher has Section 251 approval for the NDA Identifying file to be linked to the patients' name to facilitate linkage with the wider data sources in the study. NDA does not hold the patient name so NHS Digital will trace the patient name once the NDA data has been prepared.

Wider clinical information will be provided from the diabetes audits for inclusion in the analysis dataset.

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 Data will only be used for the purposes described in this Agreement.

MR1029:Patient Tracking — DARS-NIC-316558-W0T8G

Type of data: information not disclosed for TRE projects

Opt outs honoured: No - consent provided by participants of research study, Identifiable, Yes, No (, )

Legal basis: Informed Patient consent to permit the receipt, processing and release of data by the HSCIC, Informed Patient consent to permit the receipt, processing and release of data by NHS Digital,

Purposes: No (Academic)

Sensitive: Sensitive, and Non Sensitive, and Non-Sensitive

When:DSA runs 2010-01-15 — 2024-01-14 2017.03 — 2019.03.

Access method: Ongoing, One-Off

Data-controller type: CARDIFF UNIVERSITY SCHOOL OF MEDICINE

Sublicensing allowed: No

Datasets:

  1. MRIS - Cause of Death Report
  2. MRIS - Cohort Event Notification Report
  3. MRIS - Scottish NHS / Registration
  4. MRIS - Flagging Current Status Report
  5. MRIS - Members and Postings Report
  6. MRIS - Personal Demographics Service

Objectives:

Treatment choice and outcome substantially decided by age.
Patients <60 years have a significantly improved survival.
Of patients >60 years treated intensively over the last 15 years there is little evidence of improvements in survival, CR rate not improved beyond 60% and relapse in 85% of patients.
Large groups of patients >60 years are not considered fit for intensive treatment - median survival 4 months.
Median age of disease is 65 years
As the general population lives longer, the number of patients in this age group will increase.
Urgent need to find new treatments for patients >60 years who are not catered for in most trials.

Project 9 — DARS-NIC-79434-P1T7D

Type of data: information not disclosed for TRE projects

Opt outs honoured: Y ()

Legal basis: Section 251 approval is in place for the flow of identifiable data

Purposes: ()

Sensitive: Non Sensitive

When:2017.12 — 2018.02.

Access method: Ongoing

Data-controller type:

Sublicensing allowed:

Datasets:

  1. MRIS - List Cleaning Report

Objectives:

This application seeks to provide a list clean of a subset of participants from the Building Blocks study who have been lost to follow up to ascertain whether they are still eligible to be included in the study.

The request for HES/ONS data has already been submitted and approved under a separate application and agreement (NIC-333498-D1K7G) this application is looking for updated identifiers and any death notifications for approx 110 participants for who the study cannot match with when sending them to the National Pupil Database.

The cardiff study team will use the updated identifiers to send to The National Pupil Database (Dept. for Education) (NPD) for matching to their records to extract routine education and social care data. Where the study update details using the tracing service they will be using these details to then match to NPD.

Details of the wider study are below;

The Family Nurse Partnership (FNP) has been developed and licenced by the University of Colorado. It is a voluntary, preventive programme for vulnerable young first time mothers. It offers intensive and structured home visiting, delivered by specially trained nurses, from early pregnancy until age two. Its three aims are: to improve pregnancy outcomes, improve child health and development and improve parents’ economic self-sufficiency.

A strong and rigorous US evidence base, developed over 30 years, has shown FNP benefits the most needy young families in the short, medium and long term across a wide range of outcomes helping improve social mobility and break the cycle of inter-generational disadvantage and poverty. Proven benefits include:

• improvements in antenatal health
• reductions in children’s injuries, neglect and abuse
• improved parenting practices and behaviour
• fewer subsequent pregnancies and greater intervals between births
• improved early language development, school readiness and academic achievement
• increased maternal employment and reduced welfare use
• increases in fathers’ involvement 

The University of Colorado (UC) has licensed the FNP to the Department of Health (DH). The model of international replication for FNP specified by UC follows four stages:

(i) adaptation to local context;
(ii) pilot testing of feasibility and acceptability;
(iii) randomised controlled trial, and
(iv) replication and expansion.

The programme was adapted for implementation and introduced in England in 2007. Due to the greatly differing nature of publicly funded health and social care service provision and socio-cultural context between England and the US, the relative benefits of the programme need to be replicated in England and costs determined before wide-spread implementation can be recommended.

Department of Health (DH) is committed to strengthening the evidence base for FNP in an English context. To that end, DH commissioned the ‘Building Blocks’ randomised controlled trial (RCT) from Cardiff University to provide independent evidence on the effectiveness of the FNP programme in improving short term outcomes for young parents and their babies. The trial began in 2009 and the findings which cover the period from pregnancy to the child’s second birthday' were published in October 2015. The FNP described these as “important early findings and add to the evidence we have from the US, Netherlands and other early evaluation in England to help improve FNP in England.”

The National Institute of Health Research has now funded Cardiff University to undertake a follow up study to examine child outcomes to age six. This will build on the original study examining the longer term impact of FNP intervention. The study objectives are:
1. To determine the effectiveness of the FNP programme in reducing objectively measured long-term maltreatment outcomes when compared to usually provided health and social care alone. Using a multi-method multisource approach to maltreatment research main outcomes will be: Child in need status, child protection registration, referral to social care (overall; child protection; Child in Need)
2. To determine the long-term effectiveness of the FNP programme in reducing maltreatment when assessed using associated measures of injuries and ingestions, hospital DNA rates and immunisation rates.
3. To determine the long-term impact of the FNP programme upon intermediate programme outcomes, most notably subsequent pregnancies.
4. To explore the impact of theoretical moderators of programme effect, including domestic abuse and baseline client characteristics
5. To determine the costs and consequences of the FNP programme over the full period of available follow-up.

Cardiff University will follow up the mothers and children who took part in the first Building Blocks trial (BB:0-2) by obtaining health and mortality data from NHS Digital and data from Department for Education (DfE) which will be linked with the original trial data.

The original Building Blocks trial (BB:0-2) provided evidence for the short-term effectiveness of the programme (up to 2 years after birth).

The Building Blocks: 2-6 (BB:2-6) study will provide evidence for the long-term effectiveness and costs of one of the most promising early intervention programmes for reducing risk of child maltreatment in a targeted vulnerable population. Specifically, data requested from NHS Digital will provide the basis for key study outcomes which are indicators of maltreatment.

The study will provide evidence to inform policy about whether to continue implementing a programme. The proposal presents a unique opportunity to extend learning from the trial by using existing trial outcome data in combination with newly arising routinely recorded data.

Expected Benefits:

The main benefits will be to ensure the wider study succeeds be ensuring that the only eligible participants are included. The benefit will also be that any participant who is no longer eligible will not be included in any future communication thus preventing any distress or harm being caused to family members.

The outputs for the wider study as recommended for approval under NIC-333498 are detailed below.

This study will provide evidence for the long-term effectiveness and costs of one of the most promising early intervention programmes in a targeted vulnerable population. It will inform policy about whether to continue implementing a programme for which there is no existing UK evidence for effectiveness. The recognised potential programme benefits – in particular for child maltreatment have largely been evidenced in the longer term. This project presents a unique opportunity to extend learning from the trial by using existing outcome data in combination with newly arising data.

The original study (BB:0-2) has already been viewed by the relevant DH Policy team. Cardiff University expect the trial results will inform post-election decision-making regarding the implementation of this programme. The policy team within the Department of Health are fully aware of the follow-on study. The academic research team and DH policy team have a rolling joint dissemination meeting (to manage trial results dissemination) and Cardiff University will continue this relationship into the work of the follow on study, whilst being mindful of maintaining independence of the research team. The Department of Health hold the licence for the program in the UK. Its continued availability will depend upon evidence produced by the trial, and Cardiff University expect by the follow-on study.

This project is looking at the long term effects of a home-visiting intervention commissioned by the Department of Health. The study aims to determine the long-term effectiveness of this intervention in reducing objective and associated measures of maltreatment.

There was no pre-existing evidence for programme effectiveness in the UK (England). The completed trial provided definitive evidence of short-term impact. The current work will provide new evidence about longer-term impact on maltreatment. Existing evidence for programme effectiveness related to maltreatment exists in the US context only. US evidence (specifically for maltreatment) includes Olds et al JAMA 1997 278 (8); 637-643.

Both the original trial (Funder: DH PRP) and the follow-on study (Funder: NIHR-Public Health Research Programme NIHR PHR) are independent evaluations of the intervention. It is important for the research team to retain this independence and aim to produce high quality evidence to inform practice and policy. The Department of Health have indicated the importance of the work by funding both studies to a combined value of £6M. The evidence base for policy should comprise all relevant research and not the results of a single trial cohort, although Cardiff University’s trial cohort will produce the most directly relevant evidence.

The Department of Health funded the research team to run a large stakeholder event in January (2016) to which practitioners, policy leads and lay representatives from across the UK attended.

Following discussion with the Department of Health (23 March 2015), Cardiff University can also state that the intervention under investigation is currently embedded across 135 local authorities and that the results of the trial will influence not only policy but commissioning decisions in local authorities who now have responsibility for commissioning public health services for children aged 0-5. Cardiff University expect that the results accruing from the current data request will have similar reach and engagement from commissioners.

Outputs:

The output for this application will be an updated cohort list.

The outputs for the wider study as recommended for approval under NIC-333498 are detailed below.

Results of the final analyses (following the second data extraction) will be reported to the Department of Health (NIHR-Public Health Research, and the DH Policy Research Programme), and to the FNP National Unit (FNPNU). The FNPNU is responsible for national delivery of FNP and is commissioned by the Department of Health and Public Health England who hold the license in England and have the lead role for its strategic policy direction. All local authorities in England will be notified of the results, as (since October 2015) they have responsibility for commissioning public health services for children aged 0-5. Participants will receive a summary of the results and all reports and publications will be made available in full in the public domain on the Cardiff University website. The research team have convened and met twice with a stakeholder group, including relevant policy leads from each country in the UK delivering FNP (England, Scotland, Northern Ireland). Cardiff University will stage a similar event to present and discuss the implications for practice and policy of the results of this longer-term follow up of participants. The reports are planned for summer 2018.

In addition to the policy and public outputs, there will also be academic outputs which are outlined below. The purpose of these academic outputs are to report the methods used in order to answer the research question as well as the results of the study. These will be presented both in writing and at conferences for the purpose of sharing knowledge to aid other researchers using these methods, these data, and these topic areas. Publishing in scientific journals will involve rigorous independent scientific peer review. This provides additional reassurance to the funder, the public and other researchers that the methods and results presented are of high quality, credible and scientifically robust.

The study plans the following academic publications:

1. A ‘protocol’ paper has been submitted for publication subject to acceptance by BMJ Open. This describes the aims, objectives and research design of the study. This exposes our approach to the scrutiny of other academics, raises awareness of the work taking place and provides an indication about when to expect the results.

2. A paper describing the piloting process of the study and describing data quality, the success of data matching at multiple information centres and the linkage conducted at SAIL. Academics will learn from the methodology of this work and use this to inform their own research. This is planned for second half of 2017 and the aim is to publish in BMC Paediatrics.

3. A paper on main results of the study to be published in the Lancet is planned for mid-2018. This is a high impact international journal which will reach academics across the UK and in other countries with the results of the long-term outcomes of FNP.

The individuals who will be cleaning and analysing the de-identified data include an individual who is also studying for a PhD. This individual will be discussing in their PhD thesis how a variety of data sources can be combined to build a clear picture of confirmed maltreatment, markers of maltreatment, and predictors of maltreatment for women and children recruited to the original trial and this long-term follow up. This differs from the main results of the study which focuses on confirmed cases of maltreatment only. Results presented for this PhD will be the availability and validity of linking fields that allow different data sources to be linked for measuring maltreatment.

Processing:

Cardiff will send in the name, date of birth (or estimated delivery date), NHS Number, contact address and gender. for approx 110 participants.

NHS Digital will trace the individuals and provide an updated list to include the following data items: name, date of birth, NHS Number, most recent contact address and gender.

The list will be flagged with any individuals who have died in the interim.

The list of updated identifiers will be returned to the applicants at Cardiff University.

The applicants will remove those patients who no longer meet the eligibility criteria from the list i.e. those who have died. Remaining patients will be followed up as per the agreed protocol.

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 ie: employees, agents and contractors of the Data Recipient who may have access to that data).

The LUCI Study: The long-term follow-up of urinary tract infection (UTI) in childhood — DARS-NIC-74625-S1Q8X

Type of data: information not disclosed for TRE projects

Opt outs honoured: N, Anonymised - ICO Code Compliant, No (, Mixture of confidential data flow(s) with support under section 251 NHS Act 2006 and non-confidential data flow(s), )

Legal basis: Section 251 approval is in place for the flow of identifiable data, Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii); National Health Service Act 2006 - s251 - 'Control of patient information'., Health and Social Care Act 2012 – s261(2)(b)(ii); National Health Service Act 2006 - s251 - 'Control of patient information'.

Purposes: No (Academic)

Sensitive: Non Sensitive, and Non-Sensitive

When:DSA runs 2017-07-01 — 2020-06-30 2017.09 — 2017.11.

Access method: One-Off

Data-controller type: CARDIFF UNIVERSITY

Sublicensing allowed: No

Datasets:

  1. Hospital Episode Statistics Admitted Patient Care
  2. Hospital Episode Statistics Outpatients
  3. Hospital Episode Statistics Accident and Emergency
  4. Hospital Episode Statistics Accident and Emergency (HES A and E)
  5. Hospital Episode Statistics Admitted Patient Care (HES APC)
  6. Hospital Episode Statistics Outpatients (HES OP)

Objectives:

The LUCI Study (The long-term follow-up of urinary tract infection (UTI) in childhood) is funded by Welsh Government through Health and Care Research Wales. It aims to determine the short, medium and longer-term outcomes of urine infections (UTI) in childhood and to determine whether there is a difference between UTI that is identified through routine practice and UTI identified through systematic urine sampling (where all ill children have their urine sampled).

Current guidelines advise the prompt diagnosis of treatment of UTI in children because childhood UTI can lead to scarring of the kidneys which is believed to lead to long term complications such as high blood pressure, chronic kidney disease and kidney failure in some cases. However, UTI is difficult to diagnose in children as the symptoms are non-specific and similar to those found in many common childhood illnesses. Cardiff University have shown that as many as 80% of UTIs are missed in primary care. A urine sample is required to diagnose UTI but urine is infrequently sampled from ill children in primary care (in less than 2% of consultations). If urine is not sampled, a UTI cannot be diagnosed.

The evidence linking childhood UTI with long-term complications is weak and has been questioned. There is an urgent need to clarify the association between childhood UTI, renal scarring and long term complications as the correct approach to urine sampling and diagnosis of UTI in children hinges on this association. The studies, on which much of the current practice is based, are generally conducted in secondary care, and children have usually only had urine sampled when the clinician suspects a UTI rather than urine systematically sampled in all ill children. Therefore, the findings from these studies are not necessarily applicable to the population of acutely ill children in primary care.

Not all children with UTI will have had their urine tested and therefore not all will have been diagnosed. In addition, not all of the children diagnosed will have been seen in secondary care and not all will have had a scan to look for renal scarring. Cardiff University have previously conducted two large studies of acutely ill children less than five years old in primary care (EURICA study n=1003 Wales only [Funder: Welsh Office for Research and Development (WORD), now known as Health & Care Research Wales]; DUTY study n=7163 England and Wales[Co-led by Cardiff and Bristol Universities; Funder: National Institute for Health Research Health Technology Assessment – NIHR HTA) and collected urine samples systematically from all ill children. The DUTY & EURICA studies found UTI was present in 5.9% and 5.6% respectively in NHS laboratories in the two studies. Cardiff University followed up EURICA children for six months and a sample of DUTY children for three months.

Systematically sampled (DUTY and EURICA children): Cardiff University will use HES/NHS Digital data (for participants living in England) and Secure Anonymised Information Linkage (SAIL) data (for participants living in Wales). Routinely sampled (NOT a participant in DUTY or EURICA): Cardiff University will use data from SAIL ONLY.

In the current study, it is proposed to use routinely collected health data (GP, hospital, microbiology) to follow DUTY and EURICA children up for 5 years to determine short (up to 1 year) and medium (1-5 years) term outcomes for UTI identified through systematic urine sampling. The outcomes of interest will include further UTI episodes, difficulties with passing urine (dysfunctional voiding syndromes), hospital admissions, any scans on the kidneys or bladder, kidney scarring, high blood pressure and kidney failure.

Cardiff University want to compare the presenting features, initial management and short and medium-term outcomes of UTI identified through systematic urine sampling (using DUTY and EURICA children) with UTI identified through standard (clinician-led) urine sampling. Cardiff University will use routinely collected data to identify a group of children who had a UTI identified in childhood (before the age of five) who were not part of the DUTY or EURICA studies and therefore identified through standard sampling rather than systematic sampling. Cardiff University will compare this cohort of children with the DUTY and EURICA cohorts. Cardiff University will also determine longer term outcomes (more than five years) for those with a UTI identified through standard practice aged less than five years old where the data is available.

This study maximises the benefits of the previously funded DUTY and EURICA cohorts, representing over 8000 acutely ill children recruited from UK primary care. Significant resources were invested by funders, patients and staff to develop these cohorts. Access to HES data will allow the University to determine longer-term outcomes for these children and to determine risks of adverse outcomes. It will also pave the way for even longer-term follow-up of cohorts of children with UTI (diagnosed both systematically and non-systematically) which has been identified as a high research priority by NICE.

A group of 4 lay members have contributed to the development of the participant letter and the content and design of the website. Prior to the letters being sent out, one of the lay members talked through with the study team possible questions they might get from participants. The LUCI Study continue to involve this group in the research by updating them on progress.

The University of Bristol and the University of Oxford will not act as data processors or data controllers in respect of this application.

Expected Benefits:

This study will clarify the association between childhood UTI, renal scarring and long-term complications and determine the best approach for managing ill children and UTI.

Urinary tract infection (UTI) is the cause of 6% of all acute illness episodes in children presenting in primary care [Refs 8-9]. It is the most common cause of serious bacterial illness in children, and an important cause of hospital admission. [Refs 1-2] The results from this study will improve the diagnosis and management of UTI in primary care which in turn will reduce morbidity and hospital admissions.

Childhood UTIs have also been associated with serious long-term complications including renal scarring, hypertension and renal failure. [Refs 3-7]. Current guidelines recommend the prompt diagnosis of childhood UTI in order to try to prevent these complications. [Ref 3] However, urine sampling from children is difficult, particularly in primary care settings. This study aims to determine whether increased urine sampling strategies benefit children with UTI and reduce complications.

Although the association between childhood UTI and long-term complications are generally accepted and drive much of the current advice concerning UTI in children, the evidence for the strength of the association is weak. The National Institute for Health and Care Excellence (NICE) guideline concludes that ‘there are no appropriate studies that accurately estimate the risks of long-term complications as a result of childhood UTI’, highlighting the importance of cohort studies, such as this one.[ Ref 3]. Our study will clarify the association between childhood UTI, renal scarring and long-term complications and help to determine the best approach for managing ill children and UTI.

If childhood UTI is confirmed as a significant contributor to chronic conditions such as hypertension, chronic kidney disease and renal failure, there will be an opportunity to reduce some of the burden of these conditions by the early diagnosis and treatment of childhood UTI, improving health outcomes for both children and adults, reducing hospital admissions and reducing NHS costs.

This study will also determine whether a broader urine sampling strategy is warranted for acutely ill children presenting in primary care. Despite current guidelines [ref 3] which recommend urine sampling in many acutely ill children, the Cardiff and Bristol co-led DUTY study found that urine is sampled in less than 2% of all consultations with acutely ill children [ref 8]. In the EURICA study, Cardiff University found that as many as 80% of UTI in children is missed using current (non- systematic) urine sampling strategies [ref 9]

It is not known whether UTI case finding with systematic urine sampling in acutely ill children is necessary to reduce the risk of serious long-term complications. If GPs continue with current practice, we know the majority of childhood UTIs in primary care will be missed but the consequences of this are not known as it unclear whether or not these missed UTIs represent increased risk of serious complications. UTI diagnosed through systematic urine sampling (obtaining urine samples from all presenting ill children) is likely to be different to UTI diagnosed using standard urine sampling on which much of the current evidence is based.

The study will be able to compare outcomes for children with UTI identified through systematic sampling as part of the DUTY or EURICA study and children with UTI identified through routine sampling.

It is essential to clarify the evidence base in order to be able to weigh up the cost of systematic urine sampling (which would be significant due to the large numbers involved) against the cost of morbidity associated with missed UTIs and long term complications which could also be extremely costly to the NHS. This study will determine whether there is a need for systematic urine sampling in primary care and in turn change guidelines such as NICE to reflect the evidence.

Results will be communicated via peer-reviewed publications, conference presentations, and through Cardiff University websites. The National Society for Academic Care (SAPC) and the International General Practice Research on Infections (GRIN) Conferences will be targeted. These conferences are attended by practicing clinicians as well as academics from the UK and Europe. Peer-reviewed publications concerning the primary care management of common conditions are reviewed by GP update course organisers, for example the findings from the EURICA study were presented in the NB Medical Update Course in 2016; enabling dissemination of findings to clinicians directly involved in the management of ill children and UTI in primary care.

Findings from this study will inform clinical guidelines. The NICE childhood UTI guideline is currently being updated and findings from the Cardiff and Bristol co-led DUTY study are being included in this review. Further information was requested from the DUTY study team by the NICE guidelines Update Team. The NICE guidelines are regularly updated and we will inform them of the results from this study when they are available.

Cardiff University will also be able to describe variation in urine sampling and UTI diagnosis across Wales which will inform interventions to increase urine sampling if the findings from this study suggest that is warranted.

Health and Care Research Wales are funding this study. This study was awarded the grant on the strength of the application in terms of the clinical need and impact of the research, fit with Welsh Government priority areas and on the scientific merit of the study.

The early diagnosis and treatment of childhood UTI is key to preventing future health problems and is consistent with the Welsh Government's commitment of giving every child a healthy start. The Early Years and Childcare plan (Building a Brighter Future), highlights the importance of improving the quality of care, outcomes and health for every child in Wales. The project concerns both the reduction of acute morbidity in children and the prevention of chronic conditions. The importance of prevention is a strong theme throughout Welsh Government policy. The Chief Medical Officer's 2008 and 2013/14 reports emphasize 'preventing the preventable', and prevention was highlighted as one of six key areas for action in 'Our Healthy Future' and further strengthened in 'Together for Health'. In Together for Health a commitment was made to improve the health of everyone in Wales, particularly children. Improving the diagnosis of childhood UTI and determining appropriate urine sampling strategies in primary will help to reduce acute morbidity and suffering in young children and may reduce hospital admissions. Chronic conditions are highlighted as one of the most serious challenges for the NHS. Prevention is also a key priority in NHS England’s ‘Five Year Forward View (2014).

Determining the most appropriate urine sampling strategy is in line with the policy of developing 'systems for assuring high quality care', identified in Together for Health and consistent with 'Prudent healthcare'. An investment in better determining causative factors and developing strategies which may be able to prevent the development of these chronic conditions will result in better health, better quality of life and ultimately better value for money.

This study also maximises the benefits of the previously funded DUTY & EURICA cohorts which together represent the largest cohort of acutely ill children recruited from UK primary care (n=8166). Significant resources were invested by funders, patients and staff to develop these cohorts. There is rich clinical and economic data and uniquely, UTI status from systematically sampled urine. Long-term follow-up studies of children with UTI are urgently needed to determine risk of complications and to inform clinicians on appropriate sampling and management strategies. This study provides this opportunity without the need to recruit a new cohort of children.

Outputs:

This study aims to clarify the association between childhood UTI, renal scarring and long-term complications and to determine whether systematic urine sampling improves outcomes for children with UTI.

In addition to the funders of the study, Health and Care Research Wales, which is part of the Welsh Government, the findings from this study will be of interest to clinicians, parents of children and groups reviewing and developing clinical guidelines (such as NICE) across the UK and Europe with similar patient populations and clinical practices. This study was awarded the grant on the strength of the application in terms of the clinical need and impact of the research, fit with Welsh Government priority areas and on the scientific merit of the study.

The results from this study will be reported to the funder – Health and Care Research Wales at the end of the study (September 2018) who will review the report with a view to making recommendations on the management of UTI infections in childhood, based on the findings in the report, prior to making it publicly available. The funders will not suppress the findings from the report before it is published - the requirement for review is to ensure their own Public Relations (PR) team is prepared for the publication of the report to the public domain. The early diagnosis and treatment of childhood UTI is key to preventing future health problems and is consistent with the Welsh Government's commitment of giving every child a healthy start.

A lay summary of the results will be made available on the University project website, the participant website they are sent via opt-out and both will link to any publications and news items. All DUTY and EURICA participants were contacted to inform them of the study and given them the option of opting out of the study. They were given the link to the participant website and informed that the website will be updated throughout the course of the study. Results will therefore be made available in lay terms on the website.

In addition to the policy (via report for funder) and public outputs (via website), there will also be academic outputs which are outlined below. The purpose of these academic outputs is to report the methods used in order to answer the research questions and the results and implications of the study findings. These will be presented both in writing and at conferences for the purpose of sharing knowledge, to aid other researchers using these methods, these data and these topic areas. Publishing in scientific journals will involve rigorous independent scientific peer review. This provides additional reassurance to the funder, the public and other researchers that the methods and results presented are of high quality.

The study plans the following academic publications:
A ‘protocol paper’ to be submitted for publication to BMJ open or BMC Family Practice. This is planned for Summer 2018. Proposed Title - The long-term follow-up of urinary tract infection (UTI) in childhood: A Protocol Paper. This will describe the aims, objectives and research design of the study. This exposes the Universities approach to the scrutiny of other academics, raises awareness of the work taking place and provides an indication about when to expect the results.

A paper on the main results of the study will be submitted for publication to a UK clinical journal such as the BMJ or BJGP. This is planned for early 2019. Proposed Title - The long-term follow-up of urinary tract infection (UTI) in childhood. Cardiff University will target a high impact journal which will reach academics and clinicians across the UK and in other countries with the results of the long-term outcomes of childhood UTI.

All access for analysis is to de-identified data; no identifiable data will be accessed. Published results will only contain aggregated data with small numbers suppressed in line with the HES analysis guide, indeed all outputs from the SAIL data safe haven will abide by the small number policy.

Cardiff University plan to present results at conferences in 2018-2019. This is likely to include the Society for Academic Primary Care National Conference (July 2019) and the General Practice Research on Infections International Meeting (October 2018). A paper on the main results of the study will be submitted for publication to a UK clinical journal such as the BMJ or BJGP. This is planned for 2019.

Processing:

Cardiff University are the Data Controllers. A contract between Cardiff University and Swansea University has been set up to allow SAIL staff (Swansea University) to process the data from NHS Digital on the study team's behalf. SAIL databank will also act as a data provider of Welsh Health data. All data that the study team in Cardiff University will access is pseudonymised. SAIL are responsible for converting the study ID to an anonymised linking field. SAIL staff are all substantive employees of Swansea University. The team in Cardiff are all substantive employees of Cardiff University.

Additional detail about SAIL - lifted from the SAIL Databank Website (https://saildatabank.com/faq/):
“SAIL stands for Secure

Cardiff University will provide the following fields to NHS Digital:
Study ID
NHS Number
Date of Birth
Sex
Postcode

NHS Digital will link this information to HES Accident and Emergency, Admitted Patient Care and Outpatient data, strip out identifiers and return a pseudonymised output to the data processor, the Swansea University's Secure Anonymised Information Linkage (SAIL) Databank.

Cardiff University will also send SAIL a copy of the DUTY and EURICA data (their answers from questionnaires conducted during the original study) and will separately send participant identifiers to NWIS which will supply pseudonymised Welsh health data on its records to SAIL.

The only identifier provided to SAIL will be the Study ID. SAIL will assign an anonymous linking field (ALF) to each individual to replace the study ID. The study ID-ALF key will be encrypted and stored securely in SAIL. The individuals analysing the data will not have access to this key.

The key will be the same for each of the datasets so that all data can be linked up without using identifiers to do so. The key will be retained by SAIL so that any individual who expresses at a later date a wish to be removed from the study can then be removed from the dataset. This data linkage model is the same as the one used, and approved, in the Building Blocks: 2-6 study (NIC-333498). In the event that a participant wishes to opt-out after data has been sent to SAIL, the Cardiff Research team will notify SAIL to request that a participant be removed from the study database. SAIL will be notified of the original study ID and will remove the corresponding ALF and data relating to that individual.

The research team work at South East Wales Trials Unit, part of the Centre for Trials research at Cardiff University. The unit is fully registered with the United Kingdom Clinical Research Collaboration (UKCRC) fully registered clinical trials unit - South East Wales Trials Unit at Cardiff University. Definition of UKCRC - A Registration Process has been established for Clinical Trials Units responsible for coordinating multi-centre clinical studies. This is intended to help improve the quality and quantity of available expertise to carry out UK clinical trials. To gain UKCRC Registration, trials units must demonstrate a track record of experience in coordinating multi-centre trials, expert staff to develop studies, robust quality assurance systems and evidence of long term viability of capacity for trials coordination.

All data will be maintained in the safe data haven in SAIL within which all analyses will be undertaken. The research database will not be made available to other researchers (this will be a project specific resource). Access to the pseudonymised dataset will be via a secure remote portal. No data will leave the secure environment at SAIL. Approved, named data users (the research team - as listed in the application) can access the portal and defined data views remotely, subject to the appropriate access level being set and secure access keys being provided to them.

The contract between Cardiff University and Swansea University will confirm who will have access to the project database. Access is restricted to only those individuals listed in the application to NHS Digital, who are either staff at Cardiff University or Swansea University.

In addition, data cannot be downloaded from the portal, all exports of data are approved by SAIL who ensure raw data and results with small numbers (and therefore a risk of identification) are not exported. Only graphs, statistical analyses outputs and aggregated tables of data with small numbers suppressed in line with the HES analyses guide will be exported out of the secure portal.

Systematically sampled (DUTY and EURICA children): HES data (for participants living in England) and SAIL data (for participants living in Wales) will be used. The only data linked with HES data are the EURICA, DUTY questionnaire data and the SAIL equivalent of HES (i.e. inpatient, outpatient, accident and emergency). No other data will be linked to the HES data.

Routinely sampled (NOT a participant in DUTY or EURICA): Data from SAIL ONLY will be used. [SAIL data sources are: inpatient, outpatient, accident and emergency, GP, microbiology results of urine samples, congenital anomaly data and radiology data]. These data will not be linked to HES data (as these are a different cohort of individuals and therefore it would not be possible to link)

Summary of the main analysis of the linked pseudonymised dataset:

Children aged <5 years of age with an acute illness who had at least one microbiologically confirmed UTI identified through routine sampling will be compared with children who had at least one UTI confirmed through systematic sampling through participation in the EURICA/DUTY studies. Outcomes will include serious short-term (less than 1 year) and medium-term (1-5 years) outcomes, including hospital admission, renal imaging, renal scarring (primary outcome), vesicoureteric reflux (VUR) and renal failure outcomes for all children in these groups using the Inpatient and outpatient data from SAIL databank in Wales and HES hospital data in England. Cardiff University will also describe the two populations by children with and without confirmed UTI in each of these populations (routine and systematic sampling).

More detail of the analysis plan is outlined below:
• Describe the presenting features for children with and without microbiologically confirmed UTI in terms of age at consultation, gender and presenting symptom or working diagnosis from general practice records by group (routine and systematic sampling).
• Describe the initial management for children with and without microbiologically confirmed UTI in terms of antibiotic treatment, hospital admission and repeat urine sampling from general practice records by group (routine sampling).
• Describe the short term (up to 12 months) and medium term (one to five years) clinical outcomes (listed above) using general practice and hospital records for children with and without UTI (routine and systematic sampling).
• Routine versus Selective sampling analysis:
o Presentation factors, acute management, microbiology results will be examined to ascertain any marked imbalance between the two groups and identify any potential confounders of outcome.
o The primary comparative analysis of renal scarring will be examined using a multilevel logistic regression model to investigate differences between the groups (routine v systematic sampling) and where numbers allowed, variation in outcome will be accounted for at the level of the general practice.
o Associations between potential confounders of outcome will be examined at the univariate level and significant predictors adjusted for in the multivariable model.
o Multilevel modelling will allow for clustering of effect within a general practice and where this indicates little impact of clustering on effect, results from the single level model will be presented.
o Comparisons will be presented as unadjusted and adjusted risk differences and odds ratios, alongside 95% confidence intervals and p-values.
o Logistic multilevel modelling will also be used to analyse the associated secondary outcomes (e.g. proportion of children with a hospital admission).
o Counts data such as the number of GP consultations, urine samples sent will be analysed using Poisson or Negative Binomial multilevel regression modelling.

Following analysis, aggregated results / publishable information can be requested out of the secure environment for wider disclosure (subject to the data file being approved by data guardians at SAIL). Data guardians check for sensitive data, and small numbers that could risk disclosure before approving the file.

All results will be aggregated with small numbers suppressed in line with the HES analyses guide.

MR826 - AML15 - MRC working parties on leukaemia in adults & children acute myeloid leukaemia trial 15 — DARS-NIC-184980-J5B6C

Type of data: information not disclosed for TRE projects

Opt outs honoured: N, Identifiable (Consent (Reasonable Expectation), Mixture of confidential data flow(s) with consent and flow(s) with support under section 251 NHS Act 2006)

Legal basis: Informed Patient consent to permit the receipt, processing and release of data by the HSCIC, Health and Social Care Act 2012 – s261(2)(c)

Purposes: No (Academic)

Sensitive: Sensitive, and Non Sensitive

When:DSA runs 2019-12-01 — 2020-02-29 2017.03 — 2017.05.

Access method: Ongoing, One-Off

Data-controller type: CARDIFF UNIVERSITY

Sublicensing allowed: No

Datasets:

  1. MRIS - Cause of Death Report
  2. MRIS - Cohort Event Notification Report
  3. MRIS - Flagging Current Status Report
  4. MRIS - Members and Postings Report

Objectives:

Treatment choice and outcome substantially decided by age. Patients <60 years have a significantly improved survival, Of patients >60 years treated intensively over the last 15 years there is little evidence of improvements in survival, CR rate not improved beyond 60% and relapse is 85% of patients. Large groups of patients >60 years are not considered fit for intensive treatment - median survival 4 months. Median age of disease is 65 years.

As the general population lives longer, the number of patients in this age group will increase. There is an urgent need to find new treatments for patients >60 years who are not catered for in most trials.

HCTU will use the data to support the analysis and long term follow up of the AML15 clinical trial with an aim of producing academic papers.

Yielded Benefits:

Results from the AML15 trial were used to further refine treatments in this population. The design of the subsequent AML17 and AML19 trials, in the same population of patients (i.e. patients within the same age range and with same condition but different patients), has been based on emerging results from the AML15 trial. Patients with APL (a sub-set of AML) have benefited significantly from the emergence of a chemotherapy-free regimen. Further advances have been made in refining the identification of patients suitable (or not) for stem cell transplant.

Expected Benefits:

The NCRI AML trials sought to improve treatments for patients with acute myeloid leukaemia for over 20 years, supported by the NCRI AML Working Group. The measure of benefits, particularly in the younger patient population, is reliant on survival status being followed up for all patients, usually well past the end date of the clinical trial itself. It is important to identify any patients or sub-populations that have benefited from trial treatment and any late effects of the treatment given. The tracking of some patients via NHS Digital allows the trial data to include patients that may have been lost to follow up at participating trial centres.

In any future application, the applicant will be required to provide details of the expected benefits resulting from the study.

Outputs:

No new outputs will be produced under this Data Sharing Agreement.

In any future application, the applicant will be required to provide details of any future outputs planned.

The data from the AML trials was used for two main purposes:

1. to contribute to the design of further trials and
2. to potentially update and improve on standard chemotherapy regimens.

The AML trials group has a history of defining standard of care treatment via the national trials, which have been adopted by the NHS, and have delivered benefit to healthcare systems on that basis.

All critical data in relation to AML15 has been published - therefore, these data will not be provided to any further public fora or pharmaceutical collaborators, unless a late safety signal emerges. The follow-on trials, which were informed by the methodology and findings from AML15, have now been designed and are underway. These trials involve different study cohorts and testing of different drugs. Therefore, the data received from NHS Digital covered by this Agreement will not be used to update future trial designs.

All outputs were aggregated with small numbers suppressed in line with the HES Analysis Guide.

Processing:

Under this Agreement, the data may be securely stored but not otherwise processed. No new data will be provided by NHS Digital under this Agreement.

The study data, including data provided by NHS Digital under previous Agreements, are currently held by Cardiff University. Under this interim extension all devices containing data will be securely stored at Cardiff University's storage address specified in this Agreement.

The following provides background on the processing activities undertaken for the original study:

Identifiable data was shared with ONS to carry out the linkage between the study data and civil registration data. Participants records were ‘flagged’ with the Office for National Statistics (ONS). ONS notified the study team at Cardiff University of participants’ deaths (date and cause) and cancer events when they occurred. The ‘flagging for long-term follow up’ service transferred from ONS to the HSCIC in 2008. Data was last supplied in March 2017.

Project 12 — DARS-NIC-42321-K7G1C

Type of data: information not disclosed for TRE projects

Opt outs honoured: N ()

Legal basis: Health and Social Care Act 2012, Section 42(4) of the Statistics and Registration Service Act (2007) as amended by section 287 of the Health and Social Care Act (2012)

Purposes: ()

Sensitive: Sensitive, and Non Sensitive

When:2016.12 — 2017.02.

Access method: One-Off

Data-controller type:

Sublicensing allowed:

Datasets:

  1. Hospital Episode Statistics Admitted Patient Care
  2. Hospital Episode Statistics Outpatients
  3. Hospital Episode Statistics Accident and Emergency
  4. Hospital Episode Statistics Critical Care
  5. Office for National Statistics Mortality Data (linkable to HES)
  6. Mental Health and Learning Disabilities Data Set
  7. Mental Health Minimum Data Set
  8. Improving Access to Psychological Therapies Data Set

Objectives:

The national Child Health Outcome Review continues the work of the Confidential Enquiry into Child Health (CMACE). This new programme is commissioned by the Health Quality Improvement Partnership (HQIP) on behalf of NHS England, NHS Wales, the Northern Ireland Department of Health, Social Services and Public Safety (DHSSPS), the Scottish Government, the states of Guernsey, the States of Jersey and the Isle of Man government. NHS representatives from each of the four nations (England, Wales, Scotland and Northern Island) )sit on the CMACE's Independent Advisory Group. The project is being undertaken as a partnership between the National Confidential Enquiry into Patient Outcome and Death (NCEPOD), Cardiff and Swansea Universities with Swansea's SAIL (The Secure Anonymised Information Linkage) Databank being at the core of project. This work will add to the findings of the CMACE which found that the highest proportions of child deaths were among those with chronic disabilities and mental health problems in adolescence.

There are two workstreams to this programme and are being run separately -

1) This is run by NCEPOD which consists of a case note review, organisational survey and consultation with children and their parents regarding their care. NCEPOD are performing this and it does not impact the work of the second workstream, and vice versa.

2) This is being run by Swansea University and Cardiff Universities. It will link administrative health care data, which offers a source of data to provide a population based overview of adolescent mental health (AMH) and childhood neurodisability for children and young people (CYP). The data will be stored and accessed within SAIL, based at Swansea University along with the AMH team. The chronic neurodisability team are based at Cardiff University who will travel to Swansea to access the data.

The project arose from the findings of the report, “Clinical Outcome Review Programme: Overview of Child Deaths in Four UK Countries” commissioned by the Healthcare Quality Improvement Partnership (HQIP), which found 30-40% of children who died in England and Wales were affected by neurological or sensory conditions and 30-40% of 13-18 year olds were affected by mental health problems, learning disabilities and behavioural conditions.

The aim of this study is to explore:
1. Characteristics of hospital and primary care attendances including rates, reasons, length of stay and outcomes for the populations of interest and the general children and young people population
2. The interface between different care settings for example primary and secondary care and health and social care
3. Transition between child and adult services

These aims will be investigated according to condition, age and socio-demographic characteristics.

NHS Digital data is vital to this project in order to report on the rates at which children and young people with AMH and chronic neurodisabilities attend hospital in England, and the reasons why they are admitted. It will also provide information about the nature and outcomes of such attendances in comparison to children and young people without either of the conditions of interest. This data will feed into the wider project objective of exploring healthcare for children and young people with AMH and chronic disabilities across the four nations and will compliment the qualitative research carried out by NCEPOD.

Cerebral palsy (CP) has been selected as the index condition for chronic neurodisability and self-harm (SH), anxiety and depression (A & D), and eating disorders (ED) as the index conditions for AMH disorders.

The priority questions the study aims to answer for CP and AMH using NHS Digital data are as follows;

CP –
What is the annual incidence of hospital admissions per child with CP by disease severity?
What are the reasons for hospital admission (disease type /procedure) by disease severity?
How many inpatient, outpatient attendances/ED attendances are made annually by children with CP?
What is the pattern of primary care attendances per child by severity and age (subject to data availability from CPRD)?
How many children in each age group received MRI at diagnosis?
How many children receive regular hip X Ray surveillance?

AMH -
What is the annual incidence of hospital admissions per child with SH, A&D, ED?
What are the reasons for hospital admission?
How many inpatient, outpatient, attendances/ED are made annual by CYP with SH, A&D, ED?
What is the pattern of primary care attendances per child (subject to data availability from CPRD)
What is the relationship between primary care attendance and hospital admissions and outpatient attendances?
What is the relationship between GP or hospital attendance for SH, A&D or ED and overall mortality?

Expected Benefits:

This project is being conducted to identify problems in health care quality and delivery for CYP with AMH and neurodisability against those who are not affected by these conditions. It is a follow up study to the HQIP ‘Overview of Child Deaths in Four UK Countries’ report, which found that two thirds of children who died in England and Wales had chronic conditions, of which 30-40% were affected by neurological or sensory conditions and 30-40% were affected by mental health, learning difficulties or behavioural conditions.

It’s recognised that until the data has been processed and the outputs created, that there is difficulty in recognising the true benefits. However, exploration of the reasons why these CYP are admitted to hospital, the patterns of healthcare utilisation, as well as the assessment of quality of care indicators (e.g. waiting times) could highlight gaps in care and areas for improvement. The allocation of NHS funding to this timely project indicates its necessity and the findings, if disseminated appropriately (see previous section), could have a considerable impact.

A secondary benefit of this work is that the assessment of the feasibility of such a large, data-linkage study using population-based healthcare data will provide recommendations for improvements in data recording and data accessibility.

Furthermore, this work could be used to facilitate research questions. Since, although only AMH and CP will be addressed, the issues these populations face in terms of healthcare accessibility are similar to other conditions or diseases and the methods used in this study can be replicated and extended.

Ultimately however, it is hoped that through wide dissemination of the report - a launch day, press coverage, social media, summary sheets relevant to children and lay audiences, conferences, academic publications etc. - the report will act as a communicative platform for clinicians and policy makers to improve services for patients.

Outputs:

This project is the first government funded investigation into the extent to which routinely collected data may be used to inform and improve the health and social care received by CYP in the UK.

The proposed project outcomes for the CP and AMH work are as follows;

CP –

1) A descriptive analysis of the number, nature and reasons for hospital admissions, specifically admissions to Paediatric intensive care units (PICU) for children with CP compared for England, Northern Ireland, Scotland and Wales.
2) A description of primary and secondary care utilization for a representative population of children with CP in England and Wales
3) A description of healthcare utilization for cohorts of children with CP in each Nation to include outpatient, Emergency department attendeances where possible and severity where possible
4) A comparison of health care utilization at transition from children to adult services

AMH

1) A description of primary and secondary care utilization for a representative population of children with AMH (SH, A&D, ED) in England and Wales
2) A description of healthcare utilization for cohorts of children with AMH (SH, A&D, ED) in each Nation to include outpatient, Emergency department attendances where possible and severity where possible
3) A comparison of health care utilization at transition from children to adult services

Using these outcomes it is hoped that the key questions detailed within the objectives will be answered.

These outcomes will form the final national project report, that is due in December 2017. It will inform keystakeholders such as: service users, clinicians, health care commissioners, Departments of Health, Royal Colleges and policy makers, clinical standards and guideline development groups. The full report will be available on the HQIP and NCEPOD websites free of charge. The report will be clear and concise, and made readily available as a PDF for easy circulation.

The professional groups will be advised on areas which they can improve quality of care for adolescents with mental health disorders and CYP with chronic neurodisability. The recommendations made will be targeted at the most relevant group(s) to implement them. Key patient group stakeholders that will be informed and worked with on launch day include the children and young people’s patient safety expert group, the children and young peoples’ outcomes forum and The Child Health Intelligence Network.

In addition to the policy and public outputs, there will also be academic outputs. The purpose of these academic outputs are to report the methods used in order to answer the research question as well as the results of the study. Furthermore, reflection and feedback on the problems and barriers in obtaining and analysing UK wide data will be a valuable aid to future research. Academic outputs will be presented both in writing and at conferences for the purpose of sharing knowledge, to aid other researchers using these methods, these data, and these topic areas. Publishing in scientific journals will involve rigorous independent scientific peer review. This provides additional reassurance to the funder, the public and other researchers that the methods and results presented are of high quality, credible and scientifically robust.

The project plans to publish academic papers in peer reviewed journals such as, the Journal of Affective Disorders, BMJ Open, Archives of Disease of Child hood, Child; Care Health and Development. Academic conferences will include RCPCH Annual Conference and the Faculty of Public Health Annual Conference, and invited keynote talks will report on methodological, epidemiology and key findings for public health practitioners and CYP clinicians.

Its recognised that the provision for both groups of children require joined up multidisciplinary care and as such a communication and dissemination plan to deliver the findings to multidisciplinary audiences that include patient representation has been created .

All stakeholder groups will be asked to put links on their websites, add items into bulletins or newsletters regularly throughout the programme about the study. Alongside this, a project ‘discussion board’ where issues and problems can be posted and discussed has been created amongst the stakeholders. This has the added advantage of maintaining lines of communication across the project team.

Social media will be used following feedback from the stakeholders, and there are plans to include patient and participant leaflets, podcasts, and the use of social media such as YouTube where patients and carers will be asked to put the reports' findings in their own words.

NCEPOD sends out quarterly updates to medical directors where there are questionnaire or case note requests and this process will be adopted for this programme. A newsletter with study progress updates twice per year will be sent out to these directors.

Published results at a national level will only contain aggregated data with small numbers suppressed, in line with the HES analysis guide and the relevant Mental Health suppression guidelines.

Processing:

It is proposed that record level data from each participating nation (England, Wales, Scotland and Northern Ireland) will be released to the Secure Anonymised Information Linkage (SAIL) databank. SAIL have previous experience of hosting NHS Digital data and will be the safe haven for the datasets.

Researchers at Cardiff University have been approved by SAIL to access the data (via data access agreement, completion of Research Data and Confidentiality e­learning course, submission of CV). The Cardiff researchers will travel to SAIL and access the data on site at Swansea via the SAIL Gateway, but no record level data will be removed from the system. The SAIL Gateway is the remote data access system, and is the sole method by which approved data users are able to access any record level SAIL data. Access is restricted to only those individuals listed in the application to NHS Digital who are substantive employees of either Cardiff University or Swansea University. NCEPOD and HQIP will not have access to the data.

Only aggregated and anonymised data will small numbers suppressed, in line with the HES analysis guide and the relevant Mental Health suppression guidelines will be removed from the system.

SAIL will also hold data from Wales, Scotland and Northern Ireland and comparisons at a national aggregated and anonymised level will be made where possible. No record level data from the other countries, or any other source, will be linked to NHS Digital data.

Study population:
The study population will include HES data linked to ONS, and MHMDS/MHLDDS data, as well as a separate extract of IAPT data, for 0-25 year olds. The MHMDS/MHLDDS and ONS data supplied be only be for those patients who appear in the HES data.

All those aged 0-25 years within the HES, ONS, or MHMDS/MHLDDS are requested to be used as comparison groups. Data are requested for a 10 year time period - January 1st 2004 to 31st December 2014. The population of interest is children and young people (11-25 years) who have mental disorders (SH, A&D, ED which form the AMH) and CYP (0-25 years) with chronic neurodisability (CP) who are residents in England. For AMH, only patients 11 and older will be analysed as the number of cases below at start are low, whereas CP is often identified at a much younger age so the full age range can be used for that condition.

The population of interest will be identified and flagged by analysts in SAIL using algorithms based upon relevant ICD 10 codes (e.g. G80 for CP and X60-84 (SH), F32-39 (A & D) and F50 (ED) for AMH). Furthermore, age filters will be applied (those born after 01/01/1979) to the IAPT data, as NHS Digital cannot apply filters to this data set. The extraneous IAPT data not covering 0-25 year olds will be destroyed.

Cardiff staff will analyse both the AMH and the CP data, with Swansea staff analysing the AMH information.

Patient and Public Participation:
The service user groups for both themes form part of the study advisory groups and regularly inform project design. A further component of the NCEPOD lead project includes patient, parent and carer surveys. The findings from these are fed back into the study design.

Data Cleaning and preparation:
As with all analyses of routinely collected datasets there is potential for misclassification in coding diagnoses, linkage errors and missing data. Consistency checks will be made by SAIL analysts and missing data between the national datasets described.

Data Analysis:
This will be undertaken when datasets become available and is likely to take 12-14 months overall. The linked data will then be analysed by a team of researchers at both Cardiff and Swansea Universities who are experienced in data linkage and using large datasets.

Summary of the main analysis of the linked dataset:
Reasons for hospital admission, number of admissions and length of stay in CYP with AMH or CP will be described and compared to those without AMH or CP. Causes of death and patterns of healthcare utilisation prior to death will be examined. Reasons and frequency of admission will be compared, before, during and after the transition from paediatric to adult services. Patterns based on age, gender, demographics and area-based social deprivation will be investigated.

Descriptive statistics, chi-squared tests, rate ratios and regression analysis will be used as appropriate.
Comparison between countries will be made where possible using appropriate statistical methodologies e.g. Poisson regression.

Building Blocks:2-6 - Evaluating the long-term effectiveness, and the cost and consequences of the Family Nurse Partnership parenting support programme in reducing maltreatment in young children — DARS-NIC-333498-D1K7G

Type of data: information not disclosed for TRE projects

Opt outs honoured: Y, N, Anonymised - ICO Code Compliant (Section 251 NHS Act 2006)

Legal basis: Section 251 approval is in place for the flow of identifiable data, Section 42(4) of the Statistics and Registration Service Act (2007) as amended by section 287 of the Health and Social Care Act (2012), Health and Social Care Act 2012, Health and Social Care Act 2012 – s261(7)

Purposes: No (Academic)

Sensitive: Non Sensitive, and Sensitive, and Non-Sensitive

When:DSA runs 2019-04-13 — 2022-04-12 2018.03 — 2016.08.

Access method: One-Off

Data-controller type: CARDIFF UNIVERSITY

Sublicensing allowed: No

Datasets:

  1. Hospital Episode Statistics Outpatients
  2. Hospital Episode Statistics Accident and Emergency
  3. Office for National Statistics Mortality Data
  4. Bridge file: Hospital Episode Statistics to Mortality Data from the Office of National Statistics
  5. Hospital Episode Statistics Admitted Patient Care
  6. Civil Registration (Deaths) - Secondary Care Cut
  7. HES:Civil Registration (Deaths) bridge
  8. Civil Registrations of Death - Secondary Care Cut
  9. Hospital Episode Statistics Accident and Emergency (HES A and E)
  10. Hospital Episode Statistics Admitted Patient Care (HES APC)
  11. Hospital Episode Statistics Outpatients (HES OP)

Objectives:

The Family Nurse Partnership (FNP) has been developed and licenced by the University of Colorado. It is a voluntary, preventive programme for vulnerable young first time mothers. It offers intensive and structured home visiting, delivered by specially trained nurses, from early pregnancy until age two. Its three aims are: to improve pregnancy outcomes, improve child health and development and improve parents’ economic self-sufficiency.

A strong and rigorous US evidence base, developed over 30 years, has shown FNP benefits the most needy young families in the short, medium and long term across a wide range of outcomes helping improve social mobility and break the cycle of inter-generational disadvantage and poverty. Proven benefits include:
• improvements in antenatal health
• reductions in children’s injuries, neglect and abuse
• improved parenting practices and behaviour
• fewer subsequent pregnancies and greater intervals between births
• improved early language development, school readiness and academic achievement
• increased maternal employment and reduced welfare use
• increases in fathers’ involvement 

The University of Colorado (UC) has licensed the FNP to the Department of Health (DH). The model of international replication for FNP specified by UC follows four stages:
(i) adaptation to local context;
(ii) pilot testing of feasibility and acceptability;
(iii) randomised controlled trial, and
(iv) replication and expansion.

The programme was adapted for implementation and introduced in England in 2007. Due to the greatly differing nature of publicly funded health and social care service provision and socio-cultural context between England and the US, the relative benefits of the programme need to be replicated in England and costs determined before wide-spread implementation can be recommended.

DH is committed to strengthening the evidence base for FNP in an English context. To that end, DH commissioned the ‘Building Blocks’ randomised controlled trial (RCT) from Cardiff University to provide independent evidence on the effectiveness of the FNP programme in improving short term outcomes for young parents and their babies. The trial began in 2009 and the findings which cover the period from pregnancy to the child’s second birthday' were published in October 2015. The FNP described these as “important early findings and add to the evidence we have from the US, Netherlands and other early evaluation in England to help improve FNP in England.”

The National Institute of Health Research has now funded Cardiff University to undertake a follow up study to examine child outcomes to age six. This will build on the original study examining the longer term impact of FNP intervention. The study objectives are:
1. To determine the effectiveness of the FNP programme in reducing objectively measured long-term maltreatment outcomes when compared to usually provided health and social care alone. Using a multi-method multisource approach to maltreatment research main outcomes will be: Child in need status, child protection registration, referral to social care (overall; child protection; Child in Need)
2. To determine the long-term effectiveness of the FNP programme in reducing maltreatment when assessed using associated measures of injuries and ingestions, hospital DNA rates and immunisation rates.
3. To determine the long-term impact of the FNP programme upon intermediate programme outcomes, most notably subsequent pregnancies.
4. To explore the impact of theoretical moderators of programme effect, including domestic abuse and baseline client characteristics
5. To determine the costs and consequences of the FNP programme over the full period of available follow-up.

Cardiff University will follow up the mothers and children who took part in the first Building Blocks trial (BB:0-2) by obtaining health and mortality data from NHS Digital, data from Department for Education (DfE) and data from the Department of Health (abortion statistics) which will be linked with the original trial data.

The original Building Blocks trial (BB:0-2) provided evidence for the short-term effectiveness of the programme (up to 2 years after birth).

The Building Blocks: 2-6 (BB:2-6) study will provide evidence for the long-term effectiveness and costs of one of the most promising early intervention programmes for reducing risk of child maltreatment in a targeted vulnerable population. Specifically, data requested from NHS Digital will provide the basis for key study outcomes which are indicators of maltreatment.

The study will provide evidence to inform policy about whether to continue implementing a programme. The proposal presents a unique opportunity to extend learning from the trial by using existing trial outcome data in combination with newly arising routinely recorded data.

There are two waves of data requested and extracted from NHS Digital. The first wave was to allow the study to follow up participants until all the children reach six years of age, which was in March 2017. The second data request will request data up to and including dataset period 2016/17. This application covers the second wave and final request to NHS Digital. The study is following up participants until all the children reach six years of age, which was in March 2017.

The justification for the first application was firstly a proof of principle – that the data are sufficient to answer the research questions, to identify additional variables (if required) and to develop and validate a plan for cleaning and analysis for the second (and final) analysis. This will also ensure timely delivery of the results to the Department of Health and the policy makers for FNP in the UK. This has now been completed and the study will now conduct all planned cleaning and analysis on the final dataset where they have all participants and data years of follow up.






Yielded Benefits:

Following the "pilot phase" where data were received from data providers the study team has established a regulatory compliant model of linking health, social care and education data to clinical data. The pilot phase comprised 1) Writing and running data cleaning scripts for both the HES data (from NHS Digital) and NPD data (from Dept. for Education); 2) Assessing quality of data received and match rates; 3) A pilot report to document the governance; participant opt-out; linking to Information Centres; linkage within SAIL; preparing data for analysis; analysis considerations; and development of the consort flow diagram; 4) Development of the statistical and health economic analysis plans. All of the above work demonstrated the feasibility of delivering this study to achieve the intended outputs and measurable benefits. Since receiving the final data extract we have been cleaning and analysing the data. Results are being interpreted ready for reporting to the funder in Spring 2019. As described in the expected measurable benefits, Cardiff University have been meeting with DHSC and FNP National Unit to discuss the emerging results. This has influenced their decisions and recommendations made regarding funding for the FNP National Unit which remain confidential at this stage.

Expected Benefits:

This study will provide evidence for the long-term effectiveness and costs of one of the most promising early intervention programmes in a targeted vulnerable population. It will inform policy about whether to continue implementing a programme for which there is no existing UK evidence for effectiveness. The recognised potential programme benefits – in particular for child maltreatment have largely been evidenced in the longer term. This project presents a unique opportunity to extend learning from the trial by using existing outcome data in combination with newly arising data.

The original study (BB:0-2) has already been viewed by the relevant DH Policy team. Cardiff University expect the trial results will inform post-election decision-making regarding the implementation of this programme. The policy team within the Department of Health are fully aware of the follow-on study. The academic research team and DH policy team have a rolling joint dissemination meeting (to manage trial results dissemination) and Cardiff University will continue this relationship into the work of the follow on study, whilst being mindful of maintaining independence of the research team. The Department of Health hold the licence for the program in the UK. Its continued availability will depend upon evidence produced by the trial, and Cardiff University expect by the follow-on study.

This project is looking at the long term effects of a home-visiting intervention commissioned by the Department of Health. The study aims to determine the long-term effectiveness of this intervention in reducing objective and associated measures of maltreatment.

There was no pre-existing evidence for programme effectiveness in the UK (England). The completed trial provided definitive evidence of short-term impact. The current work will provide new evidence about longer-term impact on maltreatment. Existing evidence for programme effectiveness related to maltreatment exists in the US context only. US evidence (specifically for maltreatment) includes Olds et al JAMA 1997 278 (8); 637-643.

Both the original trial (Funder: DH PRP) and the follow-on study (Funder: NIHR-Public Health Research Programme NIHR PHR) are independent evaluations of the intervention. It is important for the research team to retain this independence and aim to produce high quality evidence to inform practice and policy. The Department of Health have indicated the importance of the work by funding both studies to a combined value of £6M. The evidence base for policy should comprise all relevant research and not the results of a single trial cohort, although Cardiff University’s trial cohort will produce the most directly relevant evidence.

The Department of Health funded the research team to run a large stakeholder event in January (2016) to which practitioners, policy leads and lay representatives from across the UK attended.

Following discussion with the Department of Health (23 March 2015), Cardiff University can also state that the intervention under investigation is currently embedded across 135 local authorities and that the results of the trial will influence not only policy but commissioning decisions in local authorities who now have responsibility for commissioning public health services for children aged 0-5. Cardiff University expect that the results accruing from the current data request will have similar reach and engagement from commissioners.

Outputs:

Results of the final analyses (following the second data extraction) will be reported to the Department of Health (NIHR-Public Health Research, and the DH Policy Research Programme), and to the FNP National Unit (FNPNU). The FNPNU is responsible for national delivery of FNP and is commissioned by the Department of Health and Public Health England who hold the license in England and have the lead role for its strategic policy direction. All local authorities in England will be notified of the results, as (since October 2015) they have responsibility for commissioning public health services for children aged 0-5. Participants will receive a summary of the results and all reports and publications will be made available in full in the public domain on the Cardiff University website. The research team have convened and met twice with a stakeholder group, including relevant policy leads from each country in the UK delivering FNP (England, Scotland, Northern Ireland). Cardiff University will stage a similar event to present and discuss the implications for practice and policy of the results of this longer-term follow up of participants. The reports are planned for summer 2018.

In addition to the policy and public outputs, there will also be academic outputs which are outlined below. The purpose of these academic outputs are to report the methods used in order to answer the research question as well as the results of the study. These will be presented both in writing and at conferences for the purpose of sharing knowledge to aid other researchers using these methods, these data, and these topic areas. Publishing in scientific journals will involve rigorous independent scientific peer review. This provides additional reassurance to the funder, the public and other researchers that the methods and results presented are of high quality, credible and scientifically robust.

The study plans the following academic publications:

1. A ‘protocol’ paper has been published in the by BMJ Open. This describes the aims, objectives and research design of the study. This exposes our approach to the scrutiny of other academics, raises awareness of the work taking place and provides an indication about when to expect the results.

2. A paper describing the piloting process of the study and describing data quality, the success of data matching at multiple information centres and the linkage conducted at SAIL. Academics will learn from the methodology of this work and use this to inform their own research. This is in draft and the aim is to submit in 2018 to publish in BMC Paediatrics.

3. A paper on main results of the study to be published in the Lancet is planned for late-2018. This is a high impact international journal which will reach academics across the UK and in other countries with the results of the long-term outcomes of FNP.

The individuals who will be cleaning and analysing the de-identified data include an individual who is also studying for a PhD. This individual will be discussing in their PhD thesis how a variety of data sources can be combined to build a clear picture of confirmed maltreatment, markers of maltreatment, and predictors of maltreatment for women and children recruited to the original trial and this long-term follow up. This differs from the main results of the study which focuses on confirmed cases of maltreatment only. Results presented for this PhD will be the availability and validity of linking fields that allow different data sources to be linked for measuring maltreatment. This work is currently being written up as a chapter in the thesis and is due to complete in March 2018 and will be using the data provided as part of the pilot data extract.

All access for analysis is to de-identified data, no identifiable data will be accessed. Published results will only contain aggregated data with small numbers suppressed. All outputs will be aggregated with small numbers suppressed in line with the HES Analysis Guide.

Processing:

Cardiff University provided the following fields to NHS Digital:
• Study ID
• NHS Number
• Date of Birth
• Sex
• Postcode

NHS Digital linked this information to the HES and ONS mortality data, stripped out the identifiers and returned a pseudonymised output to the data processor, the Swansea University’s Secure Anonymised Information Linkage (SAIL) Databank.

Cardiff University also sent SAIL a copy of the original study (BB:0-2) data and separately sent participant identifiers to the Department for Education which supplied linked pseudonymised data from its records to SAIL.


In addition data has been requested and approved from the Department of Health abortion statistics team to link to the mothers in this cohort. The study will be using the abortion data to calculate whether a mother had a subsequent pregnancy and the study will be creating a composite outcome so that if they have an abortion record and/or a record in their HES that indicates a pregnancy(including birth, termination, miscarriage) this will be used to report the rates of subsequent pregnancies between the two groups (control and intervention).


The only identifier provided to SAIL from each of the 4 sources was the Study ID. SAIL assign an anonymous linking field (ALF) to each individual to replace the study ID. The study ID-ALF key is encrypted and stored securely by SAIL. The individuals analysing the data will not have access to this key.

The key will be same for each of the datasets so that all data can be linked up without using identifiers to do so. The key will be retained by SAIL for two reasons – firstly, any individual who expresses at a later date a wish to be removed from the study can then be removed from the dataset. Secondly, Cardiff University are following up these individuals for four years therefore a refresh/update of the data will be required (this current request) when all children are aged 6 years and SAIL will need to assign the same ALF to these data.

The individuals who will be cleaning and analysing the de-identified data includes an individual who is also studying for a PhD. This individual will be discussing in their PhD thesis how a variety of data sources can be combined to build a clear picture of confirmed maltreatment, markers of maltreatment, and predictors of maltreatment for women and children recruited to the original trial and this long-term follow up. This differs from the main results of the study which focuses on confirmed cases of maltreatment only. Results presented for this PhD will be the availability and validity of linking fields that allow different data sources to be linked for measuring maltreatment. This work is currently being written up as a chapter in the thesis and is due to complete in March 2018 and will be using the data provided as part of the pilot data extract. This is the work required to complete as described above.

The research team work within the United Kingdom Clinical Research Collaboration (UKCRC) fully registered clinical trials unit - Centre for Trials Research at Cardiff University.

All data will be maintained in the safe data haven in Swansea [SAIL] within which all analyses will be undertaken. The research database will not be made available to other researchers (this will be a project specific resource). Access to the pseudonymised dataset will be via a secure remote portal. No data will leave the secure environment at SAIL. Approved, named data users (the research team - as listed in the application) can access the portal and defined data views remotely, subject to the appropriate access level being set and secure access keys being provided to them.

The contract between Cardiff University and Swansea University will confirm who will have access to the project database. Access is restricted to only those individuals listed in the Data Sharing Agreement with NHS Digital, who are either staff at Cardiff University or Swansea University.

In addition, data cannot be downloaded from the portal, all exports of data are approved by SAIL who ensure raw data and results with small numbers (and therefore a risk of identification) are not exported. Only graphs, statistical analyses outputs and aggregated tables will be exported out of the secure portal.

Baseline data and follow-up data from Building Blocks trial will be included in the main analysis, so that the total follow-up period for each participant will be just over six years.

The following is a high level summary of the main analysis of the linked pseudonymised dataset:

Participants who received FNP during the trial (intervention arm) will be compared with those who did not (Control arm). Rates of maltreatment will be compared between the two groups as well as describing differences in education, health and social care outcomes. Standard costings will be applied to episodes of healthcare and an economic analysis will compare costs between the groups.

More detail of the analysis plan is outlined below:
• Analyses will be conducted on an intention-to-treat basis and due emphasis placed on confidence intervals for the between-arm comparisons (FNP versus Usual Care).
• Descriptive statistics of demographic and outcome measures will be used to ascertain any marked imbalance between the arms at 2 years.
• The primary comparative analysis on Child in Need (CIN) status at any point between birth and 6 years will use logistic multilevel modelling to investigate differences between the groups. [Objective 1]
• Multilevel modelling will allow for clustering of effect within a site and family nurse and where this indicates little impact of clustering on effect, results from the single level model will be presented.
• Comparisons will be presented as adjusted risk differences and odds ratios, alongside 95% confidence intervals and p-values.
• Modelling the impact of key subgroups and different intervention elements (e.g. gestational age at programme entry, dosage) on outcome will be undertaken by extending the primary models and testing for interaction effects [Objectives 2 & 4].
• Logistic multilevel modelling will also be used to analyse the associated secondary outcomes (e.g. proportion of children with injuries and ingestions) [Objective 2].
• Counts data such as the number of emergency attendances will be analysed using Poisson multilevel regression modelling [Objective 2]
• Economic evaluation will consider costs and consequences of the FNP over the six years of follow up. The within trial cost consequences analysis will be extended from 0-2 to 0-6 years through collection of resource use data from medical and education records (including from the latter data related to social care usage). The nature of the data collected during the extended period will allow the long-term model to include additional predictors and hence produce more robust long-term estimates of costs and effects [Objective 5].

Following analysis, aggregated results / publishable information can be requested out of the secure environment for wider disclosure (subject to the data file being approved by data guardians at SAIL). Data guardians check for sensitive data, and small numbers that could risk disclosure before approving the file.

All processing of ONS data will be in line with the ONS standard conditions.


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 ie: employees, agents and contractors of the Data Recipient who may have access to that data).