University of Bristol - Longitudinal Linkage Collaboration - COPI Reg 3

University of Bristol (UoB) will continue to have legal basis under Regulation 3 of the COPI Regulations for processing CPI for Covid-19 Purpose. Where UoB will use the data for previously approved research related to a Covid-19 Purpose, and this research is still ongoing, UoB are able to continue to meet the requirements in Regs 3(1), 3(3) and 7 of the COPI Regulations. DARS have assessed that UoB do still meet the 3 limbs of the COPI regulation.

The UK Longitudinal Linkage Collaboration (UK LLC) project itself is cross-institutional and has staff based at University of Bristol and the University of Edinburgh. The University of Edinburgh are collaborators however they have no role in determining how the data will be processed, nor will they have any access to the record level data. For clarification regarding the distinct staff groups noted in the agreement:

Study staff = staff from the contributing cohorts;

University of Bristol (UoB) staff = staff of UK Longitudinal Linkage Collaboration (LLC) employed by UoB;

UK LLC staff = staff employed by UK LLC (UoB and University of Edinburgh (UoE) staff).

BACKGROUND

The UK Chief Scientific Advisor has established (October 2020) a programme of National Core Studies (NCS) for SARS-CoV-2 (Covid-19) research as a coordinated, long-term, national research initiative. This will consider Covid-19 in terms of a viral pandemic (including issues of cases, transmission, symptoms, and outcomes) and in terms of the health and social impacts of behavioural restrictions designed to mitigate the harms of the pandemic. The NCS has six different sub-programmes which are addressing major Covid-19 research areas; one of these is the Longitudinal Health and Well-being (LH&W) NCS which is designed to use data from longitudinal studies to address the impact of Covid-19 and of associated viral suppression measures on health and well-being. The LLC is the central hub component of the Longitudinal Health and Well-being NCS.

The UK LLC has been designed to underpin the LH&W NCS, although not exclusively as the NCS are designed to support each other, where for example: the LLC could form the infrastructure for the long-term patient follow-up of consenting trial participants (e.g. participants in the RECOVERY trial) or, where UK longitudinal population studies (LPS) are being used to collect specific new study data which is not available through routine records (e.g. the ALSPAC birth cohort is providing a frame to collect biological samples to inform the Immunology NCS research programme; the assayed results of which will be uploaded into the LLC). For this reason, users of the LLC may come from across the full range of NCS studies and the resource will be accessible to other legitimate UK-based researchers investigating Covid-19 through a sub-licence framework. The NCS is planned to be a two/three-year research programme commencing October 2020.

THE UK LLC & DATA SUBJECTS

UoB (for LLC purposes) will underpin the NCS programme by combining, within a Trusted Research Environment, study data from >15 major inter-disciplinary UK LPS with a combined total of 1 – 2 million UK participants, centrally link to a wide range of Covid-19 relevant and non-health administrative records to utilise alongside the UK Biobank, the Zoe Symptom Tracker cohort and other sources. The integrated data, infrastructure and accompanying governance aspects will collectively be known as the UK LLC.

The de-identified data from contributing studies will be used to establish a UK LLC participant and household register, this will be used within the UK LLC to index and catalogue available data. Researchers will be able request access to the integrated data via the Health Data Research UK Gateway (https://www.healthdatagateway.org/) which through the NCS programme is being established as the de facto application point for UK Data Science projects and as a systematic and coordinated infrastructure for assessing the legitimacy and ‘safe’ credentials of a project. The value of these LPS data assets will be significantly enhanced through linkage to Covid-19 relevant health and other routine records (e.g., employment status, occupation, and benefit receipt records). This will allow the use of objectively assessed diagnoses and other records and the collection of data not amenable to self-report (such as detailed records of care, or health service utilisation). It will also allow researchers to assess and quantify/document bias and error through triangulating data collected across different organisations and settings.

Throughout the development of the UK LLC, UoB have worked collaboratively with UK LPS to co-design and develop the protocol and methodology. One of the strengths of the collaborative approach championed by the UK LLC is that UK LPS are committed and enthusiastic partners in this work: this initiative comes from within the community. The UK LLC will provide infrastructure to support the LPS and researchers working within the UK to investigate high priority COVID-19 research questions. UK LLC/UoB will continue to seek input from the LPS into decision making, guidance in future development and operation. Crucially, the LPS’s all maintain and manage ‘their’ relationships with ‘their’ participants including making critical decisions about data use, and communicating with participants about the existence of, and work of, the UK LLC and providing reassurance that this does not alter the participants existing relationship.

The studies listed below are included in this Data Sharing Agreement (DSA) for the specified COVID-19 research programme and where the Data will flow under CV19: Regulation 3 (4) of the Health Service (Control of Patient Information) Regulations 2002. This will remain the case until the end of this DSA, or at the point in time at which they are migrated onto UK LLC DSAs DARS-NIC-748729-Z8B3M or DARS-NIC-420229-G9H9S where the Data will flow under either consent and/or Section 251 of NHS Act 2006.

At the point that any of the following studies are moved onto DARS-NIC-748729-Z8B3M or DARS-NIC-420229-G9H9, the study will be considered to have transitioned and will no longer be supported through this (DARS-NIC-420168-K4N1F) DSA.

1. The National Study of Health and Development (NSHD)

2. The Southall And Brent REvisited (SABRE)

3. Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”)

4. English Longitudinal Study of Ageing (ELSA)

5. Track-COVID - subset of INTERVAL COMPARE and STRIDES bioresource

6. NIHR (National Institute for Health Research) BioResouce

7. UK Household Longitudinal Study (aka “Understanding Society'')

8. Born in Bradford

9. European Prospective Investigation into Cancer - Norfolk (EPIC Norfolk)

10. The Fenland Study

11. Twins early Development Study (TEDS)

12. INTERVAL

13. COMPARE

14. STRIDES

University of Bristol anticipate the majority of the above studies will have completed the transfer to consent/s251 within the term of this DSA, and will review it's position on whether to further extend or close this DSA (DARS-NIC-420168-K4N1F) before expiry.

The following UK LLC studies have completed the transition of legal basis to consent under the UK LLC agreement DARS-NIC-748729-Z8B3M:

1. Genetic Links to Anxiety and Depression Study (GLAD)

2. 1958 National Child Development Study (NCDS)

3. 1970 British Cohort Study (BCS 70);

4. Next Steps

5. The Millennium Cohort Study (MCS)

6. Extended Cohort for E-health, environment, and DNA (EXCEED)

The following UK LLC study has completed the transition of legal basis to section 251 under the UK LLC agreement DARS-NIC-420229-G9H9S:

- TwinsUK

The UoB scientific programme requires retention and use of the NHS England data compiled with Covid-19 relevant datasets (primary care, secondary care, community, mental health care provision, Covid-19 testing and outcomes data, NHS service use interactions such as NHS 111 records, mortality, disease registry and demographics data) for the purposes of the UK LLC Data Sharing Agreements.

The UK LLC has conducted an assessment as to whether the data held and processed within the UK LLC is identifiable/potentially identifiable and therefore Personal Data which is subject to the Data Protection Act 2018 (DPA), UK General Data Protection Regulations (UK GDPR) and Common Law Duty of Confidentiality (Common Law) in accordance with guidance issued by the Information Commissioner’s Office. The assessment determined that the infrastructure, incorporating all data flows, should be considered as containing Personal Data and being subject to the DPA/UK GDPR and Common Law requirements. However, within this the assessment identified that the UK LLC sits at the threshold of being considered as holding pseudonymous data (Personal Data) and de-identified data where the risk of re-identification is not reasonably likely (not Personal Data), and that this differing status manifests when considering the data from the perspective of the different high-level groups of UK LLC data users and their distinct processing activities. It is evident that for the contributing studies, the use of data is either identifiable (for the provision of identifiers for linkage purposes) or pseudonymous (for the provision of data into the UK LLC and the processing of study data by study staff in their demarcated study processing area) and is therefore Personal Data. Yet, for the UK LLC staff and researcher users, the risk of re-identification of data within the UK LLC is not reasonably likely. This status, for these users, has been achieved through the deployment of data level as well as contextual controls which mitigate the risk of identification. The principal amongst these is the UK LLC operates as a ‘Trusted Research Environment’ which ensures that minimised data are used by minimised number of legitimate users within a secure environment for approved and proportionate purposes.

To contribute data into the UK LLC, LPS’s will be required to flow identifiers to the NHS to establish/filter a linkage between each participants' study record and their NHS registry record(s). Studies will update their privacy notices and notify participants of their involvement with UK LLC, explaining what this means and provide options to ‘opt-out’, in line with existing policy and practice.

UK LLC TRUSTED RESEARCH ENVIRONMENT (TRE)

The LLC will be built on a UK Secure eResearch Platform (UKSeRP). This is a well-established infrastructure for data science: it has ISO27001 and NHS Data Security Protection toolkit and Office for National Statistics accreditation under the Digital Economy Act 2017. Other instances of this environment are already used to host LPS and NHS England data and have been audited and commended as being well suited to the purpose and low risk.

The UK LLC TRE will be owned, operated, and managed by the LLC team (data management based at UoB). Under instruction from UoB, contributing LPS and data owners including NHS England will provide data to the UK LLC TRE using the ‘split file’ approach which is used across UKSeRP. This split file approach (details below) ensures the physical and processing operational separation of personal identifiers and participants de-identified study data and health records. This separation means that no party involved in this process can see both identifiers and data. This split file approach is coupled with the technological and socio-governance controls applied at the UK LLC (e.g., training, contracts, policy and procedures, penalties for misuse). The UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. For UK LLC staff and research users the risk of re-identification is controlled to the point that it is not reasonably likely.

UK LLC C-19 RESEARCH PROGRAMME

The UK LLC is specifically commissioned as a database resource to inform the LH&W NCS but is also designed to be a resource for the investigation of other Covid-19 questions as these are identified and prioritised through the Scientific Advisory Group for Emergencies (see the Health Data Research UK and SAGE research prioritisation funnel - current versions available through the HDRUK SAGE report, see - https://www.hdruk.ac.uk/covid-19/).

Given these questions are likely to change over time (e.g., the identification of ‘Long COVID’ as a potential new set of complex outcomes is introducing important new research questions), the research programme the LLC is designed to inform can be defined as the investigation of Covid-19 questions using data from longitudinal studies linked to routine records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing. For the use of linked longitudinal studies and NHS records this definition is refined to make clear the investigations primary outcomes must be focused on generating benefits to the health and social care system.

The investigation of Covid-19 questions using data from longitudinal studies linked to routine NHS records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing: where the primary research outcome is designed to benefit the health and social care system. These linked data will be of unique value in underpinning a programme of research on Covid-19 informed by the data assets within the UK LLC.

This program will seek to understand the patterns and predictors of infection, (including re-infection) and disease outcomes (such as ‘long covid’), and the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. It will also explore population level changes to physical and mental health, including hospital admission and mortality, in association with viral suppression measures and how these relate to changes in health behaviours.

Given the impact of the Covid-19 containment measures, it is important to investigate the role of socio-economic and neighbourhood/environmental factors in determining population level impacts to physical and mental health, to identify both groups at risk, and factors which offer resilience to adverse outcomes. Within this, UK LLC users will analyse changes in health care service use (using self-reported data and NHS health records) to determine if patterns in these have changed during the Covid-19 pandemic. Linked NHS records will also inform consideration of service use in relation to pre-pandemic health status and regular service interactions (e.g., screening, health reviews and routine service take up (such as annual seasonal flu vaccinations)).

This programme, through the LH&W NCS will specifically seek to answer the following questions:

i) What are the mental health consequences of Covid-19 infection, and of viral suppression measures? How do these differ by key demographics, socioeconomic status and by prior mental health status?

ii) What are the risks, determinants, and expressions of the physical health consequences of long-Covid?

iii) What are the disruptions to health care services (primary and secondary), and what are the health consequences of these disruptions? How do they vary by geography, demographics, and socioeconomic status?

iv) What are the risks of re-infection?

v) How do measures of historical health & physiology impact risk of C-19 infection and its consequences?

vi) How do we best identify cases in population studies, triangulating symptoms, antigen, and antibody testing? How do antibody profiles differ by case status (e.g., those with and without symptoms), how do antibody profiles vary over time? What are the health determinants of such variability?

vii) What are the long-term socioeconomic consequences of viral suppression measures? How do these in turn impact on health and on health care utilisation?

These research themes have been scientifically reviewed by the UK Chief Scientific Advisor and the UK Chief Medical Officer and are high priority questions of critical national interest.

OTHER DATA PROCESSING PURPOSES

In addition to the primary research themes, UoB will:

1) use linked NHS data to help inform descriptive and documentary analysis of the UK LLC dataset. This will describe the combined LPS population and make clear the LLC denominator, its characteristics and how this sub-set of the population relates to the wider population (through comparison with aggregated populations statistics). This will inform research users and those (e.g., policy makers) seeking to draw inferences from findings.

2) conduct feasibility assessments to help determine the viability of proposed projects. To do this, UK LLC staff will query the UK LLC integrated dataset (LPS data and linked NHS Records) to determine if UK LLC have sufficient case numbers to be able to inform specific research proposals. This purpose enhances the ethical and governance case for the resource as it means data are not shared where they cannot meaningfully inform the science, it also helps ensure the efficient use of research funding and resources.

DATA MINIMISATION

The datasets and data items requested have been restricted to only those directly relevant to the Covid-19 research programme that this infrastructure is designed to support. UoB fully adopt the standard NHS England definition of the Covid-19 relevant dataset.

As described, the UK LLC C-19 research programme is wide ranging and seeks to understand patterns and predictors of infection, disease outcomes, the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. Datasets requested include measure of Covid-19 related outcomes and pre-Covid-19 baseline datasets (allowing the value of these longitudinal data to be maximised).

The linked data requested are minimised to include only data covering the period each contributing study was in follow-up. The data requested are further minimised to include only records of participants of studies contributing to the UK LLC and those who have objected to this use of their data are excluded. Information on participants who have ‘opted-out’ will be uploaded to the UK LLC quarterly so that those participants’ data will be excluded from future studies.

UoB are unable to minimise datasets further based on fields or episodes at this time as work continues to be conducted to inform the UK’s Covid-19 research programme which aims to be responsive to health and government policy makers. For example in HES, UoB cannot consider only episodes with a specific diagnosis or procedure at this time, UoB do not yet understand the longer-term health implications from Covid-19 infection or whether there exist patterns and predictors of Covid-19 infection based on previous episodes in HES.

The study have minimised datasets so they only include records of participants of longitudinal population studies contributing to the UK LLC and to NHS England data that are directly relevant to the research programme. Whilst UoB cannot provide justification through reference to specific analysis that will be conducted using these data, given their purpose as a resource developed to enable researchers to respond to the emerging priorities as determined by the UK Scientific Advisory Group for Emergencies.

1) The UK LLC is a resource to enable the longitudinal perspective. The value of longitudinal research lies in the collection of a broad range of data, repeated at regular intervals allowing the investigation of interactions between things occurring over the lifetime of individuals. To enable this perspective to be applied to the Covid-19 research questions UoB require data from NHS England and other providers that covers the duration of the longitudinal population studies contributing their study data. For example, the National Study of Health and Development is still collecting data from babies born in one week of 1946. Emerging Covid-19 research for example around disease severity could feasibility be informed by patterns in childhood ill health and historic episodes of hospitalisation within these cohorts.

2) As above with the duration of NHS-linked data requested, the breadth of data within the LLC enable rapid response to emerging research priorities. UoB cannot predict what these priorities will be but can provide indicative examples:

The usefulness of a QOF exception code in GDPPR indicating that a patient has refused an annual health check are that this could be indicative of their behaviour around choosing to have vaccination; insights into vaccination and health behaviour have the potential to be of great importance. The risk is there are unforeseen needs which are then hard to fulfil if the data has been minimised.

The UK LLC is designed to support longitudinal research studies and research analysts working within the UK. Longitudinal research is geared to understand the longer-term outcomes of the pandemic and emerging interactions between health and social factors and including considerations of outcomes such as Long COVID whose duration is yet to be determined, and the impact of disruption of routine health service provision on wider mental and physical health.

LEGAL BASIS, ETHICS AND RISK OF POTENTIAL HARM

UoB will be sole Data Controller under this agreement.

UoB is establishing the UK LLC and using linked NHS records to inform the LH&W NCS research programme as a task carried out in the public interest (Article 6(1)(e) given the UK LLC is designed to answer specific high-priority research questions at a time of national crisis. Given that the requested health records include flows of sensitive information, the UK LLC also draws on Article 9(2)(j) in accordance with Article 89(1) where UoB's purpose is to conduct statistical scientific research. The case for this research being in the public interest is established through balancing the strong policy driver of improved understanding of the pandemic with consideration of potential harms to the patients/participants whose records are involved. Potential harms primarily relate to breach of confidentiality and the subsequent misuse of Personal Data or erosion of trust in longitudinal research/data science.

The lawful basis for the release and use of the confidential data being shared under this version of the agreement is Regulation 3(4) of the National Health Service (Control of Patient Information Regulations) 2002 (COPI) to require NHS England to share confidential patient information with organisations entitled to process this under COPI for COVID-19 purposes. The only permitted activities under this Data Sharing Agreement (DSA) are for COVID-19 purposes and within bounds of Reg 3(2) COPI. Reg 3 (2) COPI states that: "2) For the purposes of this regulation, “processing” includes any operations, or set of operations set out in regulation 2(2) which are undertaken for the purposes set out in paragraph (1)." The research relates to the monitoring and managing of COVID-19 and would therefore be covered by Reg 3(1)(d) of COPI.

The risk of this harm is minimised as this is:

1. an observational population-level research which will not result in a direct intervention to any participant;

2. the UK LLC uses tried and tested IT infrastructure and governance frameworks specifically designed to minimise risks to privacy during health data science;

3. the UK LLC is a fully de-identified research environment with sufficient controls that the risk of disclosure is not considered reasonably likely;

4. a reasonable expectation for the use of NHS data in this way has been established by the contributing studies, with studies making best endeavours to inform participants ahead of data extraction taking place, as well as continued fair processing (in relation to UK LLC) in line with usual policies and practices. Each stage of this fair processing will provide participants with the opportunity to object;

5. National Opt-Out will also be respected unless overridden by specific consent for the use of linked health records in the studies research programme;

6. all staff and users are vetted and approved professional researchers operating within controlled and auditable conditions.

This research is ethical given the strong scientific rationale, the process is transparent, best endeavours will be made to inform participants of the use of their data within UK LLC, with a right to object, that risks are mitigated and the participants (and wider public) are likely to directly benefit from the research through improved health care and government policy provision in response to the unprecedented Covid-19 crisis. Participants will be informed about how their data is used through ongoing fair processing communications including details of the process for ‘opting-out’.

OPERATIONAL MANAGEMENT

UoB is the study sponsor and Data Controller. UK LLC staff at UoB and UoE have responsibility for day to day running of the UK LLC databank: only UoB staff will have access to the data and TRE. The UK LLC Director, principle and co-investigators and data managers are employed by UoB. The UK LLC Co-Director and UK LLC administrative staff (responsible for administrative operations, communications, and participant/public engagement) are based at UoE). The Longitudinal Health & Wellbeing National Core Study is led from University College London (UCL) and the UK LLC reports into UCL on delivery of UK LLC objectives.

The UK LLC databank is configured to have two distinct classes of operational areas, which have three functions:

1) Data management and processing: All operational control and access for data processing and management of de-identified, including but not limited to data processing and storage, descriptive analysis and report/publication writing, subset, and provision of data to approved users is restricted to UoB staff. UoB staff are the only individuals who can access all data across all studies and sources in their raw and processed forms.

2a) Onward sharing to contributing studies: A secure operating partition will be created for each contributing study. Within these linked LPS - NHS data, restricted to the participants of that study, can be access by approved study staff through the ‘onward sharing’ mechanism. The study staff will use this area to conduct descriptive analysis on their participants (e.g., to provide documentation to describe the data asset, to inform bias assessments) and to assess feasibility of projects (e.g., to determine whether there are sufficient cases to scientifically justify inclusion in a proposed project) to inform potential users and to conduct study-level due diligence.

2b) Onward sharing to research analysts: Secure operating partitions will be created for each UK LLC approved project. The UK LLC approved users for these projects will have access to their folder which will contain a sub-set of data relevant to that study where onward sharing conditions of data providers are met. The UK LLC research users will control of data processing and analysis and leading publications.

The UoB UK LLC staff and UK LLC infrastructure system administrators will have access across all three operational areas.

The University of Swansea (UoS) is a Data Processor who provide data infrastructure and data management services to UoB and will maintain the integrity of the TRE, will assist with the data management processing (Under UoB’s direction) and will conduct the output disclosure assessments (to UoB’s specification).

The UoS have sub-contracted the Data Linkage functions of the UKSeRP (including all management of participant identifiers) to the Digital Health and Care Wales (DHCW), also referred to as the UK LLC Trusted Third Party (TTP) within this agreement. DHCW's role is limited to the the handling of participant identifiers prior to being submitted to NHS England for linkage to NHSE datasets. DHCW do not access or process any data disseminated by NHS England.

The UK LLC is funded by HM Treasury as part of the National Core Studies for Covid-19 research and is administered by the UK Medical Research Council (MRC). HM Treasury and MRC will not have access to NHS England data.

SUB-LICENSING

For the UK LLC to be a useful and sustainable resource for researchers to interrogate, and obtain meaningful data in a timely manner, contractual arrangements will need to facilitate high-volume and rapid turn-around of data requests. This will only be possible through removing the burden of the need for a high number of additional data sharing agreements across multiple institutions which would impede the speed of research. Therefore, a contract structure is needed to control the data flows, to control the purposes and way these data are processed, and to assign roles and responsibilities and to enforce the governance requirements of the individual studies and data owners and the legal basis under which they are permitted to operate.

These agreements are as follows:

- Data Deposit Agreement(s) established between UoB and the studies institution(s), ensuring fair processing will be carried out, and allowing study collected data to be made available for Covid-19 research to approved users within the UK LLC TRE.

- Shared data will be deposited into TRE where it will be processed, curated, and analysed.

- TRE established through research collaboration with University of Swansea (UoS) controlled through a Data Deposit

Agreement between UoB and UoS comprising:

1) a data deposit agreement for Welsh records and other records (SymptomTracker) to be deposited in the UK LLC and used for research

2) a collaborative contract where UoB procure a UKSeRP (UoS as Data Processor) and UoS data science expertise (an academic collaboration which involves Data Processing). Some aspects of the UoS contribution will be fulfilled by sub-contractor Data Health and Care Wales (DHCW) acting as a Trusted Third Party controlled through the Sub-Contract Agreement between UoS and DHCW.

The contract structure also permits the processes by which the participants of these studies can be identified by the NHS and their records selected and provided into the UK LLC for integration with the study data and onward sharing with researchers. The agreements for selecting and depositing these data has two distinct parts:

1) Record Linkage

• UoB will instruct the contributing studies to send a file of participant identifiers to Digital Health and Care Wales (DHCW) (controlled through the Data Deposit Agreement between UoB and each contributing studies institution(s)).

• UoB will instruct the DHCW (a process controlled through the TRE procurement contract between UoB and UoS, and as described in the Data Deposit Agreement between UoB and contributing studies) to send participant identifiers to NHS England for record linkage purposes

• NHS England will use these identifiers to link these to their patient register and select the longitudinal study participants data (controlled through the Data Sharing Agreement between UoB and NHS England).

2) Data Deposition & Processing

• NHS England will then de-identify and deposit the selected data into the UK LCC where UoB will become the data controller of this patient data, for the purpose of processing it, approving onward research use and providing managed access to relevant sub-sets of data for Covid-19 research (controlled through the Data Sharing Agreement between UoB and NHS England). The structure will then enable the UK LLC to determine the onward processing of the integrated data for Covid-19 research purposes by reviewing applications from researchers and onwardly sharing sub-sets of relevant data within the TRE to approved users (controlled through the Data Access Agreement between UoB and approved researchers’ institutions). The nature of the sublicensing arrangements above are in line with the NHS England standards - https://digital.nhs.uk/services/data-access-request-service-dars/dars-guidance/sub-licencing-and-onward-sharing-of-data. NHS England requirement to audit data use is a requirement in the onward sharing contract and remains feasible in the UK LLC framework.

The UK LLC is likely to generate diverse outputs, with these forming a mix of rapid insight reports to policy makers and substantive research reports. The NCS is specifically designed to establish a clear channel between decision makers and researchers and the UK LLC is designed to enable rapid response and up-to-date information needed to support this. The likely outputs include:

* Peer-reviewed scientific publication (and pre-print publications of initial drafts prior to review which are designed enable rapid awareness of the findings), outputs will be ongoing, preprints and publications to date:

Preprints | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

Publications | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London

* Conference presentations and workshop presentations (to academic, health service practitioners and planners, government policy makers, Scientific Advisory Group for Emergencies (SAGE) members). These will be ongoing as part of our communications and engagement strategy, events to date include: Specifically for UK Longitudinal Linkage Collaboration:

Presentation: Welcome Longitudinal Population Studies conference (2021)

Regular presentations to Health Data Research UK - Data & Connectivity National Core Studies Delivery Group (2021-2022)

Presentation: CLOSER Longitudinal Population Studies conference (2022) ‘UK LLC Overview and explaining the innovations made and challenges faced’

Presentation: LHW National Core Study - Joint EHR cohort leads meeting ‘how to access UK LLC’

Workshop 1: to National Core Studies Analysts ‘Overview of UK LLC and how to access UK LLC’

Workshop 2: to CLOSER, National Core Studies and Vanguard Members ‘overview of UK LLC and how to access data

Presentation to UK LLC Communications Network (2022)

Youtube: UK Longitudinal Linkage Collaboration Verbal YouTube Public Involvement Induction (2022)

Presentation: to Elizabeth Blackwell Institute Health Data Science Showcase (2021)

Presentation: ESRC Strategic Advisory Group ‘ Understanding Data for Population research’ (2021)

Presentation: to Longitudinal Area Network and Data-link (LAND) community of practice (2022)

Poster Presentation: to Welcome Trust Longitudinal Studies Conference ‘overview of UK LLC’ (2021)

Presentation and advice: to Welcome Longitudinal Population Study COVID-19 Steering Group ‘provided specialist advice on linkage opportunities’ (2021)

Presentation and advice: to Population Research UK Programme Group ‘provided specialist advice’ (2021)

Advice: to Department of Education Scientific Advisory Group ‘provided specialist advice on record linkages in longitudinal studies (2021)

* Reports to health service practitioners and planners, government policy makers and Committee’s (e.g., National Institute for Health and Care Excellence (NICE), SAGE), it is anticipated that these will be ongoing, briefings to date include:

Briefing Reports | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

* Rapid evidence synthesis reports as requested by health and government policy makers (through the entirety of the 3-year NCS programme);

* Lay summaries provided through websites, Blogs, social media posts, privacy notices and outputs designed to promote the transparent use of data and the wider research process, regularly updated on our data-use register as follows:

Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk)

All outputs from the TRE will be assessed for disclosure risk and will be anonymous population level findings. The framework for this is based on leading 'five safes’ approach and is closely adapted from equivalent processes accredited to standards of the Office for National Statistics and the UK Statistics Authority.

All ‘meta’ products of the UK LLC research (such as code list definitions, syntax, workings for derived variables) will be made available for reuse and interrogation/replication. Access to the UK LLC is ‘open’ in the sense that any legitimate researcher conducting Covid-19 research can apply for access using a transparent and consistent decision-making process. As such, the UK LLC enhanced resource should be considered as an important ‘output’ and a measurable benefit of the UK LLC/LH&W NCS.

The high profile of Covid-19 means that many findings may have media interest. Importantly, the UK LLC is establishing a Patient/Participant Involvement and Engagement strategy within its communication work-package package with a dedicated budget and experienced Communications PPIE Officer. The package will build on existing and established relationships with participants to involve public/participants in the co-design of mechanisms to engage and effectively communicate UK LLC findings more widely with the public. UoB are committed to public involvement in our design and communications. UoB have already included HDRUK PPIE panel members (public representatives) in a consultation exercise and will draw on the extensive network of participants to further explore and co-develop dissemination mechanisms, in developing the end-to-end animation in the first instance. UoB are also developing with the studies opportunities for a subgroup of participants to be involved, in the data access decision making process for example. UoB have ensured that the programme of work is adequately resourced to compensate the public for their time, this has been an integral part of the project planning.

The UK LLC is a novel research infrastructure. The creation of this for Covid-19 research may provide methodological insights and illustrations of a new way of working for linkage in longitudinal research. Insights from this will be provided to the longitudinal community and its funders/key stakeholders (through academic papers, presentations, and contributions to advisory groups).

UK LLC has successfully transformed the landscape for using patient records in longitudinal research in the UK and is now being sustained by UK Research and Innovation (UKRI) following its commissioning as part of the National Core Studies for COVID-19 research. This recognises UK LLC's improved scientific abilities for supporting longitudinal research, its cost efficiency for the UK taxpayer and NHS, its improved safeguards for participant and patient data, and its ability as a standing responsive research resource which is ready for crisis management and thus strengthens the UK’s crisis resilience capability.

The funders (UKRI), with support from NHS and other government departments, have identified UK LLC as an optimal model for providing linkage and a secure state of the art generic research database for the UK longitudinal research community, and recently reinvested for a further five years (until at least June 2028). To date, 25 LPS contribute data to the UK LLC TRE. There are 29 projects approved with data provisioned within the TRE to researchers across institutions with more projects in the pipeline: see UK LLC Data Use Register (https://ukllc.ac.uk/data-use-register/). With funding now secured and the benefits of the TRE way of working proven and communicated, UK LLC has been approached by 17 additional LPS wishing to join the collaboration (these will be added to this agreement through amendments as these arrangements are formalised)

UK LLC as an infrastructure is therefore a primary output of the National Core Studies LH&W programme. Reflecting this, the UK’s Chief Medical Officer has now added UK LLC to his list of strategic scientific infrastructure. Specific applied research outputs which exemplify the unique benefits of UK LLC include:

(1) Understanding the impacts of healthcare disruption (Lead Investigator: Green; UK LLC Project: https://ukllc.ac.uk/data-use-register/llc_0009): The pandemic resulted in unprecedented healthcare disruption. It would expected for disruption to increase hospital admission: yet, national GP data analysed in OpenSAFELY showed (currently unpublished) falls in absolute numbers and social inequalities in ambulatory care-sensitive conditions. This ecological paradox reflects reduced access to care and is misleading to the effects of disruption. Using UK LLC’s linked LPS and hospital admission records, Green and colleagues found that individuals’ reporting greater care disruption were more likely to be admitted to hospital, suggesting pandemic disruptions adversely influenced care management (BMJ, https://doi.org/10.1136/bmj-2023-075133). They also found social patterning in these outcomes. These findings were presented to NICE and SAGE and used in service planning considerations. This project demonstrates the key advantage of having LPS self-reported study measures of disruption (a more accurate measure to OpenSAFELY’s modelled disruption estimate), and that since hospital admission is relatively rare, both UK LLC’s linkage and pooled sample size were needed to inform this research.

(2) Comparing the burden of Long COVID in the community as measured by self-report and EHRs (Lead Investigator: Williams; UK LLC Project: https://ukllc.ac.uk/data-use-register/llc_0006): comparing OpenSAFELY national data with ONS COVID Infection Survey suggests large disparities between estimated Long COVID prevalence and individuals recorded with Long COVID in EHRs. Using UK LLC, researchers directly compared LPS self-reported Long COVID and NHS interactions with UK LLC’s linked GP diagnoses and referral codes. This identified the proportion of LPS Long COVID sufferers seeking GP care and coded as needing further treatment. Using eight UK LLC studies, researchers identified ~800 participants self-reporting Long COVID by spring 2021, <1.5% received a Long COVID diagnosis or referral code after >1 year of follow-up (see https://doi.org/10.1101/2023.02.10.23285717 for initial findings subject to peer review). With National Institute for Health and Care Excellence (NICE) UK LLC are exploring improved GP coding guidance and improvements to coding patterns (eg, GP system ‘pop ups’). This project demonstrates the ability of LPS and UK LLC to inform analysis of under-reported or under-coded conditions and assess sub-group effects. Through this, UK LLC can provide unique insights into population health and service/system requirements and insights into how to target upstream interventions and care.

(3) Multi-Longitudinal Cohort Study into occupational factors and COVID Risk as part of PROTECT National Core Study (Lead Investigator: Gittins; UK LLC Project: https://ukllc.ac.uk/data-use-register/llc_0007): assessed infection risks across job-role and work-related characteristics using self-reported employment status data from 14 LPS linked to NHS England health records, including COVID-19 diagnostic testing. SARS-CoV-2 infection and COVID-19 infection risk was greater in key-workers vs not; among non-home working or some home working vs all home working; while part-time workers vs full-time; and furlough vs not had reduced risk (see https://doi.org/10.1101/2023.12.19.23298502 for initial findings subject to peer review). This project demonstrates the near unique opportunities for researchers to investigate occupational health risk factors in the UK – given that job role is not recorded in any health or non-health routine system (outside of the national census programme). For this reason, with its national coverage, large and diverse sample size, and linkages to non-health records, UK LLC is emerging as a national resource for occupational health research. In collaboration with the Health & Safety Executive (HSE), UK LLC have been working with occupational health experts to enhance the resource through harmonising job-role and other work-based study data to support a breadth of occupational research and to inform HSE’s 10-year strategy to reduce work-based mental ill-health.

Findings from the LH&W NCS have led to Cabinet office and NHS England 'teach in' sessions, a national GP alert, an NHS enhanced service specification , rapid reports for SAGE, and they have provided key evidence to guide the NICE long COVID guidelines.

• Policy makers have been provided with information on long COVID burden of disease, risk factors, definition and long term outcomes via cabinet briefing reports and SAGE reports

• Low levels of long COVID GP coding were reported. This led to an NHS enhanced service specification, directed at GPs to drive an increase in long COVID coding

• NICE used long COVID findings in its evidence to review risk factors

• Findings on occupational risk fed into Health & Safety Executive policy decision making and UK LLC is being considered by HSE (who have provided funding to scope this) as a strategic resource for future mental ill-health policy research

• Briefing notes submitted addressing inequalities in healthcare disruption, mental health decline during lockdown and the impact of furlough on health behaviour and adverse outcomes.

Throughout the project UK LLC has worked with the public via our advisory groups, to shape and communicate our programme. Public contributors have helped to create material aimed at communicating the results to all stakeholders, including infographics, short videos and an animation. UK LLC is now providing a methodological exemplar for the emerging NHS Secure Data Environment’s (SDEs) at national and sub-national level with UK LLC contributing to SDE policy design, the accreditation framework and being identified as a TRE with best practice in public involvement, transparency of operations and as leading in governance design and practice. UK LLC governance materials and processes are directly contributing to NHS and Office for National Statistics practice.

There is now substantial and increasing demand for UK LLC to support non-COVID research, and for new LPS to join the collaboration - although neither of these purposes will be permitted through this DSA, it does indicate the interest and demand for UK LLC functionality given its successful COVID-19 research outputs.

University of Bristol (UoB) will continue to have legal basis under Regulation 3 of the COPI Regulations for processing CPI for Covid-19 Purpose. Where UoB will use the data for previously approved research related to a Covid-19 Purpose, and this research is still ongoing, UoB are able to continue to meet the requirements in Regs 3(1), 3(3) and 7 of the COPI Regulations. DARS have assessed that UoB do still meet the 3 limbs of the COPI regulation.

The UK Longitudinal Linkage Collaboration (UK LLC) project itself is cross-institutional and has staff based at University of Bristol and the University of Edinburgh. The University of Edinburgh are collaborators however they have no role in determining how the data will be processed, nor will they have any access to the record level data. For clarification regarding the distinct staff groups noted in the agreement:

Study staff = staff from the contributing cohorts;

University of Bristol (UoB) staff = staff of UK Longitudinal Linkage Collaboration (LLC) employed by UoB;

UK LLC staff = staff employed by UK LLC (UoB and University of Edinburgh (UoE) staff).

BACKGROUND

The UK Chief Scientific Advisor has established (October 2020) a programme of National Core Studies (NCS) for SARS-CoV-2 (Covid-19) research as a coordinated, long-term, national research initiative. This will consider Covid-19 in terms of a viral pandemic (including issues of cases, transmission, symptoms, and outcomes) and in terms of the health and social impacts of behavioural restrictions designed to mitigate the harms of the pandemic. The NCS has six different sub-programmes which are addressing major Covid-19 research areas; one of these is the Longitudinal Health and Well-being (LH&W) NCS which is designed to use data from longitudinal studies to address the impact of Covid-19 and of associated viral suppression measures on health and well-being. The LLC is the central hub component of the Longitudinal Health and Well-being NCS.

The UK LLC has been designed to underpin the LH&W NCS, although not exclusively as the NCS are designed to support each other, where for example: the LLC could form the infrastructure for the long-term patient follow-up of consenting trial participants (e.g. participants in the RECOVERY trial) or, where UK longitudinal population studies (LPS) are being used to collect specific new study data which is not available through routine records (e.g. the ALSPAC birth cohort is providing a frame to collect biological samples to inform the Immunology NCS research programme; the assayed results of which will be uploaded into the LLC). For this reason, users of the LLC may come from across the full range of NCS studies and the resource will be accessible to other legitimate UK-based researchers investigating Covid-19 through a sub-licence framework. The NCS is planned to be a two/three-year research programme commencing October 2020.

THE UK LLC & DATA SUBJECTS

UoB (for LLC purposes) will underpin the NCS programme by combining, within a Trusted Research Environment, study data from >15 major inter-disciplinary UK LPS with a combined total of 1 – 2 million UK participants, centrally link to a wide range of Covid-19 relevant and non-health administrative records to utilise alongside the UK Biobank, the Zoe Symptom Tracker cohort and other sources. The integrated data, infrastructure and accompanying governance aspects will collectively be known as the UK LLC.

The de-identified data from contributing studies will be used to establish a UK LLC participant and household register, this will be used within the UK LLC to index and catalogue available data. Researchers will be able request access to the integrated data via the Health Data Research UK Gateway (https://www.healthdatagateway.org/) which through the NCS programme is being established as the de facto application point for UK Data Science projects and as a systematic and coordinated infrastructure for assessing the legitimacy and ‘safe’ credentials of a project. The value of these LPS data assets will be significantly enhanced through linkage to Covid-19 relevant health and other routine records (e.g., employment status, occupation, and benefit receipt records). This will allow the use of objectively assessed diagnoses and other records and the collection of data not amenable to self-report (such as detailed records of care, or health service utilisation). It will also allow researchers to assess and quantify/document bias and error through triangulating data collected across different organisations and settings.

Throughout the development of the UK LLC, UoB have worked collaboratively with UK LPS to co-design and develop the protocol and methodology. One of the strengths of the collaborative approach championed by the UK LLC is that UK LPS are committed and enthusiastic partners in this work: this initiative comes from within the community. The UK LLC will provide infrastructure to support the LPS and researchers working within the UK to investigate high priority COVID-19 research questions. UK LLC/UoB will continue to seek input from the LPS into decision making, guidance in future development and operation. Crucially, the LPS’s all maintain and manage ‘their’ relationships with ‘their’ participants including making critical decisions about data use, and communicating with participants about the existence of, and work of, the UK LLC and providing reassurance that this does not alter the participants existing relationship.

The following studies are included in this agreement for the specified COVID-19 research programme and where the Data will flow under CV19: Regulation 3 (4) of the Health Service (Control of Patient Information) Regulations 2002. This will remain the case until the end of this agreement or at the point in time at which they are migrated onto a separate data sharing agreement where the Data will flow under either consent and/or Section 251 of NHS Act 2006 and which enables UK LLC to provide a generic research database service. At the point that any of the following studies are moved onto either of these new data sharing agreements, the study will be considered to have transitioned to this new arrangement and will no longer be supported through this agreement.

1. The National Study of Health and Development (NSHD)

2. The Southall And Brent REvisited (SABRE)

3. Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”)

4. Genetic Links to Anxiety and Depression Study (GLAD)

5. English Longitudinal Study of Ageing (ELSA)

6. 1958 National Child Development Study (NCDS)

7. 1970 British Cohort Study (BCS 70);

8. Next Steps

9. The Millennium Cohort Study (MCS)

10. Track-COVID - subset of INTERVAL COMPARE and STRIDES bioresource

11. NIHR (National Institute for Health Research) BioResouce

12. Extended Cohort for E-health, environment, and DNA (EXCEED)

13. UK Household Longitudinal Study (aka “Understanding Society'')

14. Born in Bradford

15. European Prospective Investigation into Cancer - Norfolk (EPIC Norfolk)

16. The Fenland Study

17. Twins early Development Study (TEDS)

18. INTERVAL

19. COMPARE

20. STRIDES

21. TwinsUK

The UoB scientific programme requires the extraction and use of the NHS England data compiled with Covid-19 relevant datasets (primary care, secondary care, community, mental health care provision, Covid-19 testing and outcomes data, NHS service use interactions such as NHS 111 records, mortality, disease registry and demographics data) for the purpose of establishing the UK LLC.

The UK LLC has conducted an assessment as to whether the data held and processed within the UK LLC is identifiable/potentially identifiable and therefore Personal Data which is subject to the Data Protection Act 2018 (DPA), UK General Data Protection Regulations (UK GDPR) and Common Law Duty of Confidentiality (Common Law) in accordance with guidance issued by the Information Commissioner’s Office. The assessment determined that the infrastructure, incorporating all data flows, should be considered as containing Personal Data and being subject to the DPA/UK GDPR and Common Law requirements. However, within this the assessment identified that the UK LLC sits at the threshold of being considered as holding pseudonymous data (Personal Data) and de-identified data where the risk of re-identification is not reasonably likely (not Personal Data), and that this differing status manifests when considering the data from the perspective of the different high-level groups of UK LLC data users and their distinct processing activities. It is evident that for the contributing studies, the use of data is either identifiable (for the provision of identifiers for linkage purposes) or pseudonymous (for the provision of data into the UK LLC and the processing of study data by study staff in their demarcated study processing area) and is therefore Personal Data. Yet, for the UK LLC staff and researcher users, the risk of re-identification of data within the UK LLC is not reasonably likely. This status, for these users, has been achieved through the deployment of data level as well as contextual controls which mitigate the risk of identification. The principal amongst these is the UK LLC operates as a ‘Trusted Research Environment’ which ensures that minimised data are used by minimised number of legitimate users within a secure environment for approved and proportionate purposes.

To contribute data into the UK LLC, LPS’s will be required to flow identifiers to the NHS to establish/filter a linkage between each participants' study record and their NHS registry record(s). Studies will update their privacy notices and notify participants of their involvement with UK LLC, explaining what this means and provide options to ‘opt-out’, in line with existing policy and practice.

UK LLC TRUSTED RESEARCH ENVIRONMENT (TRE)

The LLC will be built on a UK Secure eResearch Platform (UKSeRP). This is a well-established infrastructure for data science: it has ISO27001 and NHS Data Security Protection toolkit and Office for National Statistics accreditation under the Digital Economy Act 2017. Other instances of this environment are already used to host LPS and NHS England data and have been audited and commended as being well suited to the purpose and low risk.

The UK LLC TRE will be owned, operated, and managed by the LLC team (data management based at UoB). Under instruction from UoB, contributing LPS and data owners including NHS England will provide data to the UK LLC TRE using the ‘split file’ approach which is used across UKSeRP. This split file approach (details below) ensures the physical and processing operational separation of personal identifiers and participants de-identified study data and health records. This separation means that no party involved in this process can see both identifiers and data. This split file approach is coupled with the technological and socio-governance controls applied at the UK LLC (e.g., training, contracts, policy and procedures, penalties for misuse). The UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. For UK LLC staff and research users the risk of re-identification is controlled to the point that it is not reasonably likely.

UK LLC C-19 RESEARCH PROGRAMME

The UK LLC is specifically commissioned as a database resource to inform the LH&W NCS but is also designed to be a resource for the investigation of other Covid-19 questions as these are identified and prioritised through the Scientific Advisory Group for Emergencies (see the Health Data Research UK and SAGE research prioritisation funnel - current versions available through the HDRUK SAGE report, see - https://www.hdruk.ac.uk/covid-19/).

Given these questions are likely to change over time (e.g., the identification of ‘Long COVID’ as a potential new set of complex outcomes is introducing important new research questions), the research programme the LLC is designed to inform can be defined as the investigation of Covid-19 questions using data from longitudinal studies linked to routine records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing. For the use of linked longitudinal studies and NHS records this definition is refined to make clear the investigations primary outcomes must be focused on generating benefits to the health and social care system.

The investigation of Covid-19 questions using data from longitudinal studies linked to routine NHS records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing: where the primary research outcome is designed to benefit the health and social care system. These linked data will be of unique value in underpinning a programme of research on Covid-19 informed by the data assets within the UK LLC.

This program will seek to understand the patterns and predictors of infection, (including re-infection) and disease outcomes (such as ‘long covid’), and the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. It will also explore population level changes to physical and mental health, including hospital admission and mortality, in association with viral suppression measures and how these relate to changes in health behaviours.

Given the impact of the Covid-19 containment measures, it is important to investigate the role of socio-economic and neighbourhood/environmental factors in determining population level impacts to physical and mental health, to identify both groups at risk, and factors which offer resilience to adverse outcomes. Within this, UK LLC users will analyse changes in health care service use (using self-reported data and NHS health records) to determine if patterns in these have changed during the Covid-19 pandemic. Linked NHS records will also inform consideration of service use in relation to pre-pandemic health status and regular service interactions (e.g., screening, health reviews and routine service take up (such as annual seasonal flu vaccinations)).

This programme, through the LH&W NCS will specifically seek to answer the following questions:

i) What are the mental health consequences of Covid-19 infection, and of viral suppression measures? How do these differ by key demographics, socioeconomic status and by prior mental health status?

ii) What are the risks, determinants, and expressions of the physical health consequences of long-Covid?

iii) What are the disruptions to health care services (primary and secondary), and what are the health consequences of these disruptions? How do they vary by geography, demographics, and socioeconomic status?

iv) What are the risks of re-infection?

v) How do measures of historical health & physiology impact risk of C-19 infection and its consequences?

vi) How do we best identify cases in population studies, triangulating symptoms, antigen, and antibody testing? How do antibody profiles differ by case status (e.g., those with and without symptoms), how do antibody profiles vary over time? What are the health determinants of such variability?

vii) What are the long-term socioeconomic consequences of viral suppression measures? How do these in turn impact on health and on health care utilisation?

These research themes have been scientifically reviewed by the UK Chief Scientific Advisor and the UK Chief Medical Officer and are high priority questions of critical national interest.

OTHER DATA PROCESSING PURPOSES

In addition to the primary research themes, UoB will:

1) use linked NHS data to help inform descriptive and documentary analysis of the UK LLC dataset. This will describe the combined LPS population and make clear the LLC denominator, its characteristics and how this sub-set of the population relates to the wider population (through comparison with aggregated populations statistics). This will inform research users and those (e.g., policy makers) seeking to draw inferences from findings.

2) conduct feasibility assessments to help determine the viability of proposed projects. To do this, UK LLC staff will query the UK LLC integrated dataset (LPS data and linked NHS Records) to determine if UK LLC have sufficient case numbers to be able to inform specific research proposals. This purpose enhances the ethical and governance case for the resource as it means data are not shared where they cannot meaningfully inform the science, it also helps ensure the efficient use of research funding and resources.

DATA MINIMISATION

The datasets and data items requested have been restricted to only those directly relevant to the Covid-19 research programme that this infrastructure is designed to support. UoB fully adopt the standard NHS England definition of the Covid-19 relevant dataset.

As described, the UK LLC C-19 research programme is wide ranging and seeks to understand patterns and predictors of infection, disease outcomes, the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. Datasets requested include measure of Covid-19 related outcomes and pre-Covid-19 baseline datasets (allowing the value of these longitudinal data to be maximised).

The linked data requested are minimised to include only data covering the period each contributing study was in follow-up. The data requested are further minimised to include only records of participants of studies contributing to the UK LLC and those who have objected to this use of their data are excluded. Information on participants who have ‘opted-out’ will be uploaded to the UK LLC quarterly so that those participants’ data will be excluded from future studies.

UoB are unable to minimise datasets further based on fields or episodes at this time as work continues to be conducted to inform the UK’s Covid-19 research programme which aims to be responsive to health and government policy makers. For example in HES, UoB cannot consider only episodes with a specific diagnosis or procedure at this time, UoB do not yet understand the longer-term health implications from Covid-19 infection or whether there exist patterns and predictors of Covid-19 infection based on previous episodes in HES.

The study have minimised datasets so they only include records of participants of longitudinal population studies contributing to the UK LLC and to NHS England data that are directly relevant to the research programme. Whilst UoB cannot provide justification through reference to specific analysis that will be conducted using these data, given their purpose as a resource developed to enable researchers to respond to the emerging priorities as determined by the UK Scientific Advisory Group for Emergencies.

1) The UK LLC is a resource to enable the longitudinal perspective. The value of longitudinal research lies in the collection of a broad range of data, repeated at regular intervals allowing the investigation of interactions between things occurring over the lifetime of individuals. To enable this perspective to be applied to the Covid-19 research questions UoB require data from NHS England and other providers that covers the duration of the longitudinal population studies contributing their study data. For example, the National Study of Health and Development is still collecting data from babies born in one week of 1946. Emerging Covid-19 research for example around disease severity could feasibility be informed by patterns in childhood ill health and historic episodes of hospitalisation within these cohorts.

2) As above with the duration of NHS-linked data requested, the breadth of data within the LLC enable rapid response to emerging research priorities. UoB cannot predict what these priorities will be but can provide indicative examples:

The usefulness of a QOF exception code in GDPPR indicating that a patient has refused an annual health check are that this could be indicative of their behaviour around choosing to have vaccination; insights into vaccination and health behaviour have the potential to be of great importance. The risk is there are unforeseen needs which are then hard to fulfil if the data has been minimised.

The UK LLC is designed to support longitudinal research studies and research analysts working within the UK. Longitudinal research is geared to understand the longer-term outcomes of the pandemic and emerging interactions between health and social factors and including considerations of outcomes such as Long COVID whose duration is yet to be determined, and the impact of disruption of routine health service provision on wider mental and physical health.

LEGAL BASIS, ETHICS AND RISK OF POTENTIAL HARM

UoB will be sole Data Controller under this agreement.

UoB is establishing the UK LLC and using linked NHS records to inform the LH&W NCS research programme as a task carried out in the public interest (Article 6(1)(e) given the UK LLC is designed to answer specific high-priority research questions at a time of national crisis. Given that the requested health records include flows of sensitive information, the UK LLC also draws on Article 9(2)(j) in accordance with Article 89(1) where UoB's purpose is to conduct statistical scientific research. The case for this research being in the public interest is established through balancing the strong policy driver of improved understanding of the pandemic with consideration of potential harms to the patients/participants whose records are involved. Potential harms primarily relate to breach of confidentiality and the subsequent misuse of Personal Data or erosion of trust in longitudinal research/data science.

The lawful basis for the release and use of the confidential data being shared under this version of the agreement is Regulation 3(4) of the National Health Service (Control of Patient Information Regulations) 2002 (COPI) to require NHS England to share confidential patient information with organisations entitled to process this under COPI for COVID-19 purposes. The only permitted activities under this Data Sharing Agreement (DSA) are for COVID-19 purposes and within bounds of Reg 3(2) COPI. Reg 3 (2) COPI states that: "2) For the purposes of this regulation, “processing” includes any operations, or set of operations set out in regulation 2(2) which are undertaken for the purposes set out in paragraph (1)." The research relates to the monitoring and managing of COVID-19 and would therefore be covered by Reg 3(1)(d) of COPI.

The risk of this harm is minimised as this is:

1. an observational population-level research which will not result in a direct intervention to any participant;

2. the UK LLC uses tried and tested IT infrastructure and governance frameworks specifically designed to minimise risks to privacy during health data science;

3. the UK LLC is a fully de-identified research environment with sufficient controls that the risk of disclosure is not considered reasonably likely;

4. a reasonable expectation for the use of NHS data in this way has been established by the contributing studies, with studies making best endeavours to inform participants ahead of data extraction taking place, as well as continued fair processing (in relation to UK LLC) in line with usual policies and practices. Each stage of this fair processing will provide participants with the opportunity to object;

5. National Opt-Out will also be respected unless overridden by specific consent for the use of linked health records in the studies research programme;

6. all staff and users are vetted and approved professional researchers operating within controlled and auditable conditions.

This research is ethical given the strong scientific rationale, the process is transparent, best endeavours will be made to inform participants of the use of their data within UK LLC, with a right to object, that risks are mitigated and the participants (and wider public) are likely to directly benefit from the research through improved health care and government policy provision in response to the unprecedented Covid-19 crisis. Participants will be informed about how their data is used through ongoing fair processing communications including details of the process for ‘opting-out’.

OPERATIONAL MANAGEMENT

UoB is the study sponsor and Data Controller. UK LLC staff at UoB and UoE have responsibility for day to day running of the UK LLC databank: only UoB staff will have access to the data and TRE. The UK LLC Director, principle and co-investigators and data managers are employed by UoB. The UK LLC Co-Director and UK LLC administrative staff (responsible for administrative operations, communications, and participant/public engagement) are based at UoE). The Longitudinal Health & Wellbeing National Core Study is led from University College London (UCL) and the UK LLC reports into UCL on delivery of UK LLC objectives.

The UK LLC databank is configured to have two distinct classes of operational areas, which have three functions:

1) Data management and processing: All operational control and access for data processing and management of de-identified, including but not limited to data processing and storage, descriptive analysis and report/publication writing, subset, and provision of data to approved users is restricted to UoB staff. UoB staff are the only individuals who can access all data across all studies and sources in their raw and processed forms.

2a) Onward sharing to contributing studies: A secure operating partition will be created for each contributing study. Within these linked LPS - NHS data, restricted to the participants of that study, can be access by approved study staff through the ‘onward sharing’ mechanism. The study staff will use this area to conduct descriptive analysis on their participants (e.g., to provide documentation to describe the data asset, to inform bias assessments) and to assess feasibility of projects (e.g., to determine whether there are sufficient cases to scientifically justify inclusion in a proposed project) to inform potential users and to conduct study-level due diligence.

2b) Onward sharing to research analysts: Secure operating partitions will be created for each UK LLC approved project. The UK LLC approved users for these projects will have access to their folder which will contain a sub-set of data relevant to that study where onward sharing conditions of data providers are met. The UK LLC research users will control of data processing and analysis and leading publications.

The UoB UK LLC staff and UK LLC infrastructure system administrators will have access across all three operational areas.

The University of Swansea (UoS) is a Data Processor who provide data infrastructure and data management services to UoB and will maintain the integrity of the TRE, will assist with the data management processing (Under UoB’s direction) and will conduct the output disclosure assessments (to UoB’s specification).

The UoS have sub-contracted the Data Linkage functions of the UKSeRP (including all management of participant identifiers) to the Digital Health and Care Wales (DHCW), also referred to as the UK LLC Trusted Third Party (TTP) within this agreement. DHCW's role is limited to the the handling of participant identifiers prior to being submitted to NHS England for linkage to NHSE datasets. DHCW do not access or process any data disseminated by NHS England.

The UK LLC is funded by HM Treasury as part of the National Core Studies for Covid-19 research and is administered by the UK Medical Research Council (MRC). HM Treasury and MRC will not have access to NHS England data.

SUB-LICENSING

For the UK LLC to be a useful and sustainable resource for researchers to interrogate, and obtain meaningful data in a timely manner, contractual arrangements will need to facilitate high-volume and rapid turn-around of data requests. This will only be possible through removing the burden of the need for a high number of additional data sharing agreements across multiple institutions which would impede the speed of research. Therefore, a contract structure is needed to control the data flows, to control the purposes and way these data are processed, and to assign roles and responsibilities and to enforce the governance requirements of the individual studies and data owners and the legal basis under which they are permitted to operate.

These agreements are as follows:

- Data Deposit Agreement(s) established between UoB and the studies institution(s), ensuring fair processing will be carried out, and allowing study collected data to be made available for Covid-19 research to approved users within the UK LLC TRE.

- Shared data will be deposited into TRE where it will be processed, curated, and analysed.

- TRE established through research collaboration with University of Swansea (UoS) controlled through a Data Deposit

Agreement between UoB and UoS comprising:

1) a data deposit agreement for Welsh records and other records (SymptomTracker) to be deposited in the UK LLC and used for research

2) a collaborative contract where UoB procure a UKSeRP (UoS as Data Processor) and UoS data science expertise (an academic collaboration which involves Data Processing). Some aspects of the UoS contribution will be fulfilled by sub-contractor Data Health and Care Wales (DHCW) acting as a Trusted Third Party controlled through the Sub-Contract Agreement between UoS and DHCW.

The contract structure also permits the processes by which the participants of these studies can be identified by the NHS and their records selected and provided into the UK LLC for integration with the study data and onward sharing with researchers. The agreements for selecting and depositing these data has two distinct parts:

1) Record Linkage

• UoB will instruct the contributing studies to send a file of participant identifiers to Digital Health and Care Wales (DHCW) (controlled through the Data Deposit Agreement between UoB and each contributing studies institution(s)).

• UoB will instruct the DHCW (a process controlled through the TRE procurement contract between UoB and UoS, and as described in the Data Deposit Agreement between UoB and contributing studies) to send participant identifiers to NHS England for record linkage purposes

• NHS England will use these identifiers to link these to their patient register and select the longitudinal study participants data (controlled through the Data Sharing Agreement between UoB and NHS England).

2) Data Deposition & Processing

• NHS England will then de-identify and deposit the selected data into the UK LCC where UoB will become the data controller of this patient data, for the purpose of processing it, approving onward research use and providing managed access to relevant sub-sets of data for Covid-19 research (controlled through the Data Sharing Agreement between UoB and NHS England). The structure will then enable the UK LLC to determine the onward processing of the integrated data for Covid-19 research purposes by reviewing applications from researchers and onwardly sharing sub-sets of relevant data within the TRE to approved users (controlled through the Data Access Agreement between UoB and approved researchers’ institutions). The nature of the sublicensing arrangements above are in line with the NHS England standards - https://digital.nhs.uk/services/data-access-request-service-dars/dars-guidance/sub-licencing-and-onward-sharing-of-data. NHS England requirement to audit data use is a requirement in the onward sharing contract and remains feasible in the UK LLC framework.

The UK LLC is likely to generate diverse outputs, with these forming a mix of rapid insight reports to policy makers and substantive research reports. The NCS is specifically designed to establish a clear channel between decision makers and researchers and the UK LLC is designed to enable rapid response and up-to-date information needed to support this. The likely outputs include:

* Peer-reviewed scientific publication (and pre-print publications of initial drafts prior to review which are designed enable rapid awareness of the findings), outputs will be ongoing, preprints and publications to date:

Preprints | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

Publications | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London

* Conference presentations and workshop presentations (to academic, health service practitioners and planners, government policy makers, Scientific Advisory Group for Emergencies (SAGE) members). These will be ongoing as part of our communications and engagement strategy, events to date include: Specifically for UK Longitudinal Linkage Collaboration:

Presentation: Welcome Longitudinal Population Studies conference (2021)

Regular presentations to Health Data Research UK - Data & Connectivity National Core Studies Delivery Group (2021-2022)

Presentation: CLOSER Longitudinal Population Studies conference (2022) ‘UK LLC Overview and explaining the innovations made and challenges faced’

Presentation: LHW National Core Study - Joint EHR cohort leads meeting ‘how to access UK LLC’

Workshop 1: to National Core Studies Analysts ‘Overview of UK LLC and how to access UK LLC’

Workshop 2: to CLOSER, National Core Studies and Vanguard Members ‘overview of UK LLC and how to access data

Presentation to UK LLC Communications Network (2022)

Youtube: UK Longitudinal Linkage Collaboration Verbal YouTube Public Involvement Induction (2022)

Presentation: to Elizabeth Blackwell Institute Health Data Science Showcase (2021)

Presentation: ESRC Strategic Advisory Group ‘ Understanding Data for Population research’ (2021)

Presentation: to Longitudinal Area Network and Data-link (LAND) community of practice (2022)

Poster Presentation: to Welcome Trust Longitudinal Studies Conference ‘overview of UK LLC’ (2021)

Presentation and advice: to Welcome Longitudinal Population Study COVID-19 Steering Group ‘provided specialist advice on linkage opportunities’ (2021)

Presentation and advice: to Population Research UK Programme Group ‘provided specialist advice’ (2021)

Advice: to Department of Education Scientific Advisory Group ‘provided specialist advice on record linkages in longitudinal studies (2021)

* Reports to health service practitioners and planners, government policy makers and Committee’s (e.g., National Institute for Health and Care Excellence (NICE), SAGE), it is anticipated that these will be ongoing, briefings to date include:

Briefing Reports | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

* Rapid evidence synthesis reports as requested by health and government policy makers (through the entirety of the 3-year NCS programme);

* Lay summaries provided through websites, Blogs, social media posts, privacy notices and outputs designed to promote the transparent use of data and the wider research process, regularly updated on our data-use register as follows:

Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk)

All outputs from the TRE will be assessed for disclosure risk and will be anonymous population level findings. The framework for this is based on leading 'five safes’ approach and is closely adapted from equivalent processes accredited to standards of the Office for National Statistics and the UK Statistics Authority.

All ‘meta’ products of the UK LLC research (such as code list definitions, syntax, workings for derived variables) will be made available for reuse and interrogation/replication. Access to the UK LLC is ‘open’ in the sense that any legitimate researcher conducting Covid-19 research can apply for access using a transparent and consistent decision-making process. As such, the UK LLC enhanced resource should be considered as an important ‘output’ and a measurable benefit of the UK LLC/LH&W NCS.

The high profile of Covid-19 means that many findings may have media interest. Importantly, the UK LLC is establishing a Patient/Participant Involvement and Engagement strategy within its communication work-package package with a dedicated budget and experienced Communications PPIE Officer. The package will build on existing and established relationships with participants to involve public/participants in the co-design of mechanisms to engage and effectively communicate UK LLC findings more widely with the public. UoB are committed to public involvement in our design and communications. UoB have already included HDRUK PPIE panel members (public representatives) in a consultation exercise and will draw on the extensive network of participants to further explore and co-develop dissemination mechanisms, in developing the end-to-end animation in the first instance. UoB are also developing with the studies opportunities for a subgroup of participants to be involved, in the data access decision making process for example. UoB have ensured that the programme of work is adequately resourced to compensate the public for their time, this has been an integral part of the project planning.

The UK LLC is a novel research infrastructure. The creation of this for Covid-19 research may provide methodological insights and illustrations of a new way of working for linkage in longitudinal research. Insights from this will be provided to the longitudinal community and its funders/key stakeholders (through academic papers, presentations, and contributions to advisory groups).

UK LLC has been a real success and is being sustained by UK Research and Innovation (UKRI) following its commissioning as part of the National Core Studies for COVID-19 research due to the improved scientific abilities for supporting longitudinal research, its cost efficiency for the UK taxpayer and NHS, its improved safeguards for participant and patient data, and its ability as a standing responsive research resource which is ready for crisis management and thus strengthens the UK’s crisis resilience capability.

The funders (UKRI), with support from NHS and other government departments, have identified UK LLC as an optimal model for providing linkage and a secure state of the art generic research database for the UK longitudinal research community, and recently reinvested for a further five years (until at least June 2028). To date, 25 LPS contribute data to the UK LLC TRE. There are 29 projects approved with data provisioned within the TRE to researchers across institutions with more projects in the pipeline: see UK LLC Data Use Register (https://ukllc.ac.uk/data-use-register/). With funding now secured and the benefits of the TRE way of working proven and communicated, UK LLC has been approached by 17 additional LPS wishing to join the collaboration (these will be added to this agreement through amendments as these arrangements are formalised)

UK LLC as an infrastructure is therefore a primary output of the National Core Studies LH&W programme. Reflecting this, the UK’s Chief Medical Officer has now added UK LLC to his list of strategic scientific infrastructure. Specific applied research outputs which exemplify the unique benefits of UK LLC include:

(1) Understanding the impacts of healthcare disruption (Lead Investigator: Green; UK LLC Project: https://ukllc.ac.uk/data-use-register/llc_0009): The pandemic resulted in unprecedented healthcare disruption. It would expected for disruption to increase hospital admission: yet, national GP data analysed in OpenSAFELY showed (currently unpublished) falls in absolute numbers and social inequalities in ambulatory care-sensitive conditions. This ecological paradox reflects reduced access to care and is misleading to the effects of disruption. Using UK LLC’s linked LPS and hospital admission records, Green and colleagues found that individuals’ reporting greater care disruption were more likely to be admitted to hospital, suggesting pandemic disruptions adversely influenced care management (BMJ, https://doi.org/10.1136/bmj-2023-075133). They also found social patterning in these outcomes. These findings were presented to NICE and SAGE and used in service planning considerations. This project demonstrates the key advantage of having LPS self-reported study measures of disruption (a more accurate measure to OpenSAFELY’s modelled disruption estimate), and that since hospital admission is relatively rare, both UK LLC’s linkage and pooled sample size were needed to inform this research.

(2) Comparing the burden of Long COVID in the community as measured by self-report and EHRs (Lead Investigator: Williams; UK LLC Project: https://ukllc.ac.uk/data-use-register/llc_0006): comparing OpenSAFELY national data with ONS COVID Infection Survey suggests large disparities between estimated Long COVID prevalence and individuals recorded with Long COVID in EHRs. Using UK LLC, researchers directly compared LPS self-reported Long COVID and NHS interactions with UK LLC’s linked GP diagnoses and referral codes. This identified the proportion of LPS Long COVID sufferers seeking GP care and coded as needing further treatment. Using eight UK LLC studies, researchers identified ~800 participants self-reporting Long COVID by spring 2021, <1.5% received a Long COVID diagnosis or referral code after >1 year of follow-up (see https://doi.org/10.1101/2023.02.10.23285717 for initial findings subject to peer review). With National Institute for Health and Care Excellence (NICE) UK LLC are exploring improved GP coding guidance and improvements to coding patterns (eg, GP system ‘pop ups’). This project demonstrates the ability of LPS and UK LLC to inform analysis of under-reported or under-coded conditions and assess sub-group effects. Through this, UK LLC can provide unique insights into population health and service/system requirements and insights into how to target upstream interventions and care.

(3) Multi-Longitudinal Cohort Study into occupational factors and COVID Risk as part of PROTECT National Core Study (Lead Investigator: Gittins; UK LLC Project: https://ukllc.ac.uk/data-use-register/llc_0007): assessed infection risks across job-role and work-related characteristics using self-reported employment status data from 14 LPS linked to NHS England health records, including COVID-19 diagnostic testing. SARS-CoV-2 infection and COVID-19 infection risk was greater in key-workers vs not; among non-home working or some home working vs all home working; while part-time workers vs full-time; and furlough vs not had reduced risk (see https://doi.org/10.1101/2023.12.19.23298502 for initial findings subject to peer review). This project demonstrates the near unique opportunities for researchers to investigate occupational health risk factors in the UK – given that job role is not recorded in any health or non-health routine system (outside of the national census programme). For this reason, with its national coverage, large and diverse sample size, and linkages to non-health records, UK LLC is emerging as a national resource for occupational health research. In collaboration with the Health & Safety Executive (HSE), UK LLC have been working with occupational health experts to enhance the resource through harmonising job-role and other work-based study data to support a breadth of occupational research and to inform HSE’s 10-year strategy to reduce work-based mental ill-health.

Findings from the LH&W NCS have led to Cabinet office and NHS England 'teach in' sessions, a national GP alert, an NHS enhanced service specification , rapid reports for SAGE, and they have provided key evidence to guide the NICE long COVID guidelines.

• Policy makers have been provided with information on long COVID burden of disease, risk factors, definition and long term outcomes via cabinet briefing reports and SAGE reports

• Low levels of long COVID GP coding were reported. This led to an NHS enhanced service specification, directed at GPs to drive an increase in long COVID coding

• NICE used long COVID findings in its evidence to review risk factors

• Findings on occupational risk fed into Health & Safety Executive policy decision making and UK LLC is being considered by HSE (who have provided funding to scope this) as a strategic resource for future mental ill-health policy research

• Briefing notes submitted addressing inequalities in healthcare disruption, mental health decline during lockdown and the impact of furlough on health behaviour and adverse outcomes.

Throughout the project UK LLC has worked with the public via our advisory groups, to shape and communicate our programme. Public contributors have helped to create material aimed at communicating the results to all stakeholders, including infographics, short videos and an animation. UK LLC is now providing a methodological exemplar for the emerging NHS Secure Data Environment’s (SDEs) at national and sub-national level with UK LLC contributing to SDE policy design, the accreditation framework and being identified as a TRE with best practice in public involvement, transparency of operations and as leading in governance design and practice. UK LLC governance materials and processes are directly contributing to NHS and Office for National Statistics practice.

There is now substantial and increasing demand for UK LLC to support non-COVID research, and for new LPS to join the collaboration - although neither of these purposes will be permitted through this DSA, it does indicate the interest and demand for UK LLC functionality given its successful COVID-19 research outputs.

University of Bristol (UoB) will continue to have legal basis under Regulation 3 of the COPI Regulations for processing CPI for Covid-19 Purpose. Where UoB will use the data for previously approved research related to a Covid-19 Purpose, and this research is still ongoing, UoB are able to continue to meet the requirements in Regs 3(1), 3(3) and 7 of the COPI Regulations. DARS have assessed that UoB do still meet the 3 limbs of the COPI regulation.

The UK Longitudinal Linkage Collaboration (UK LLC) project itself is cross-institutional and has staff based at University of Bristol and the University of Edinburgh. The University of Edinburgh are collaborators however they have no role in determining how the data will be processed, nor will they have any access to the record level data. For clarification regarding the distinct staff groups noted in the agreement:

Study staff = staff from the contributing cohorts;

University of Bristol (UoB) staff = staff of UK Longitudinal Linkage Collaboration (LLC) employed by UoB;

UK LLC staff = staff employed by UK LLC (UoB and University of Edinburgh (UoE) staff).

BACKGROUND

The UK Chief Scientific Advisor has established (October 2020) a programme of National Core Studies (NCS) for SARS-CoV-2 (Covid-19) research as a coordinated, long-term, national research initiative. This will consider Covid-19 in terms of a viral pandemic (including issues of cases, transmission, symptoms, and outcomes) and in terms of the health and social impacts of behavioural restrictions designed to mitigate the harms of the pandemic. The NCS has six different sub-programmes which are addressing major Covid-19 research areas; one of these is the Longitudinal Health and Well-being (LH&W) NCS which is designed to use data from longitudinal studies to address the impact of Covid-19 and of associated viral suppression measures on health and well-being. The LLC is the central hub component of the Longitudinal Health and Well-being NCS.

The UK LLC has been designed to underpin the LH&W NCS, although not exclusively as the NCS are designed to support each other, where for example: the LLC could form the infrastructure for the long-term patient follow-up of consenting trial participants (e.g. participants in the RECOVERY trial) or, where UK longitudinal population studies (LPS) are being used to collect specific new study data which is not available through routine records (e.g. the ALSPAC birth cohort is providing a frame to collect biological samples to inform the Immunology NCS research programme; the assayed results of which will be uploaded into the LLC). For this reason, users of the LLC may come from across the full range of NCS studies and the resource will be accessible to other legitimate UK-based researchers investigating Covid-19 through a sub-licence framework. The NCS is planned to be a two/three-year research programme commencing October 2020.

THE UK LLC & DATA SUBJECTS

UoB (for LLC purposes) will underpin the NCS programme by combining, within a Trusted Research Environment, study data from >15 major inter-disciplinary UK LPS with a combined total of 1 – 2 million UK participants, centrally link to a wide range of Covid-19 relevant and non-health administrative records to utilise alongside the UK Biobank, the Zoe Symptom Tracker cohort and other sources. The integrated data, infrastructure and accompanying governance aspects will collectively be known as the UK LLC.

The de-identified data from contributing studies will be used to establish a UK LLC participant and household register, this will be used within the UK LLC to index and catalogue available data. Researchers will be able request access to the integrated data via the Health Data Research UK Gateway (https://www.healthdatagateway.org/) which through the NCS programme is being established as the de facto application point for UK Data Science projects and as a systematic and coordinated infrastructure for assessing the legitimacy and ‘safe’ credentials of a project. The value of these LPS data assets will be significantly enhanced through linkage to Covid-19 relevant health and other routine records (e.g., employment status, occupation, and benefit receipt records). This will allow the use of objectively assessed diagnoses and other records and the collection of data not amenable to self-report (such as detailed records of care, or health service utilisation). It will also allow researchers to assess and quantify/document bias and error through triangulating data collected across different organisations and settings.

Throughout the development of the UK LLC, UoB have worked collaboratively with UK LPS to co-design and develop the protocol and methodology. One of the strengths of the collaborative approach championed by the UK LLC is that UK LPS are committed and enthusiastic partners in this work: this initiative comes from within the community. The UK LLC will provide infrastructure to support the LPS and researchers working within the UK to investigate high priority COVID-19 research questions. UK LLC/UoB will continue to seek input from the LPS into decision making, guidance in future development and operation. Crucially, the LPS’s all maintain and manage ‘their’ relationships with ‘their’ participants including making critical decisions about data use, and communicating with participants about the existence of, and work of, the UK LLC and providing reassurance that this does not alter the participants existing relationship.

The following studies are included in this agreement for the specified COVID-19 research programme and where the Data will flow under CV19: Regulation 3 (4) of the Health Service (Control of Patient Information) Regulations 2002. This will remain the case until the end of this agreement or at the point in time at which they are migrated onto a separate data sharing agreement where the Data will flow under either consent and/or Section 251 of NHS Act 2006 and which enables UK LLC to provide a generic research database service. At the point that any of the following studies are moved onto either of these new data sharing agreements, the study will be considered to have transitioned to this new arrangement and will no longer be supported through this agreement.

1. The National Study of Health and Development (NSHD)

2. The Southall And Brent REvisited (SABRE)

3. Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”)

4. Genetic Links to Anxiety and Depression Study (GLAD)

5. English Longitudinal Study of Ageing (ELSA)

6. 1958 National Child Development Study (NCDS)

7. 1970 British Cohort Study (BCS 70);

8. Next Steps

9. The Millennium Cohort Study (MCS)

10. Track-COVID - subset of INTERVAL COMPARE and STRIDES bioresource

11. NIHR (National Institute for Health Research) BioResouce

12. Extended Cohort for E-health, environment, and DNA (EXCEED)

13. UK Household Longitudinal Study (aka “Understanding Society'')

14. Born in Bradford

15. European Prospective Investigation into Cancer - Norfolk (EPIC Norfolk)

16. The Fenland Study

17. Twins early Development Study (TEDS)

18. INTERVAL

19. COMPARE

20. STRIDES

21. TwinsUK

The UoB scientific programme requires the extraction and use of the NHS England data compiled with Covid-19 relevant datasets (primary care, secondary care, community, mental health care provision, Covid-19 testing and outcomes data, NHS service use interactions such as NHS 111 records, mortality, disease registry and demographics data) for the purpose of establishing the UK LLC.

The UK LLC has conducted an assessment as to whether the data held and processed within the UK LLC is identifiable/potentially identifiable and therefore Personal Data which is subject to the Data Protection Act 2018 (DPA), UK General Data Protection Regulations (UK GDPR) and Common Law Duty of Confidentiality (Common Law) in accordance with guidance issued by the Information Commissioner’s Office. The assessment determined that the infrastructure, incorporating all data flows, should be considered as containing Personal Data and being subject to the DPA/UK GDPR and Common Law requirements. However, within this the assessment identified that the UK LLC sits at the threshold of being considered as holding pseudonymous data (Personal Data) and de-identified data where the risk of re-identification is not reasonably likely (not Personal Data), and that this differing status manifests when considering the data from the perspective of the different high-level groups of UK LLC data users and their distinct processing activities. It is evident that for the contributing studies, the use of data is either identifiable (for the provision of identifiers for linkage purposes) or pseudonymous (for the provision of data into the UK LLC and the processing of study data by study staff in their demarcated study processing area) and is therefore Personal Data. Yet, for the UK LLC staff and researcher users, the risk of re-identification of data within the UK LLC is not reasonably likely. This status, for these users, has been achieved through the deployment of data level as well as contextual controls which mitigate the risk of identification. The principal amongst these is the UK LLC operates as a ‘Trusted Research Environment’ which ensures that minimised data are used by minimised number of legitimate users within a secure environment for approved and proportionate purposes.

To contribute data into the UK LLC, LPS’s will be required to flow identifiers to the NHS to establish/filter a linkage between each participants' study record and their NHS registry record(s). Studies will update their privacy notices and notify participants of their involvement with UK LLC, explaining what this means and provide options to ‘opt-out’, in line with existing policy and practice.

UK LLC TRUSTED RESEARCH ENVIRONMENT (TRE)

The LLC will be built on a UK Secure eResearch Platform (UKSeRP). This is a well-established infrastructure for data science: it has ISO27001 and NHS Data Security Protection toolkit and Office for National Statistics accreditation under the Digital Economy Act 2017. Other instances of this environment are already used to host LPS and NHS England data and have been audited and commended as being well suited to the purpose and low risk.

The UK LLC TRE will be owned, operated, and managed by the LLC team (data management based at UoB). Under instruction from UoB, contributing LPS and data owners including NHS England will provide data to the UK LLC TRE using the ‘split file’ approach which is used across UKSeRP. This split file approach (details below) ensures the physical and processing operational separation of personal identifiers and participants de-identified study data and health records. This separation means that no party involved in this process can see both identifiers and data. This split file approach is coupled with the technological and socio-governance controls applied at the UK LLC (e.g., training, contracts, policy and procedures, penalties for misuse). The UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. For UK LLC staff and research users the risk of re-identification is controlled to the point that it is not reasonably likely.

UK LLC C-19 RESEARCH PROGRAMME

The UK LLC is specifically commissioned as a database resource to inform the LH&W NCS but is also designed to be a resource for the investigation of other Covid-19 questions as these are identified and prioritised through the Scientific Advisory Group for Emergencies (see the Health Data Research UK and SAGE research prioritisation funnel - current versions available through the HDRUK SAGE report, see - https://www.hdruk.ac.uk/covid-19/).

Given these questions are likely to change over time (e.g., the identification of ‘Long COVID’ as a potential new set of complex outcomes is introducing important new research questions), the research programme the LLC is designed to inform can be defined as the investigation of Covid-19 questions using data from longitudinal studies linked to routine records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing. For the use of linked longitudinal studies and NHS records this definition is refined to make clear the investigations primary outcomes must be focused on generating benefits to the health and social care system.

The investigation of Covid-19 questions using data from longitudinal studies linked to routine NHS records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing: where the primary research outcome is designed to benefit the health and social care system. These linked data will be of unique value in underpinning a programme of research on Covid-19 informed by the data assets within the UK LLC.

This program will seek to understand the patterns and predictors of infection, (including re-infection) and disease outcomes (such as ‘long covid’), and the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. It will also explore population level changes to physical and mental health, including hospital admission and mortality, in association with viral suppression measures and how these relate to changes in health behaviours.

Given the impact of the Covid-19 containment measures, it is important to investigate the role of socio-economic and neighbourhood/environmental factors in determining population level impacts to physical and mental health, to identify both groups at risk, and factors which offer resilience to adverse outcomes. Within this, UK LLC users will analyse changes in health care service use (using self-reported data and NHS health records) to determine if patterns in these have changed during the Covid-19 pandemic. Linked NHS records will also inform consideration of service use in relation to pre-pandemic health status and regular service interactions (e.g., screening, health reviews and routine service take up (such as annual seasonal flu vaccinations)).

This programme, through the LH&W NCS will specifically seek to answer the following questions:

i) What are the mental health consequences of Covid-19 infection, and of viral suppression measures? How do these differ by key demographics, socioeconomic status and by prior mental health status?

ii) What are the risks, determinants, and expressions of the physical health consequences of long-Covid?

iii) What are the disruptions to health care services (primary and secondary), and what are the health consequences of these disruptions? How do they vary by geography, demographics, and socioeconomic status?

iv) What are the risks of re-infection?

v) How do measures of historical health & physiology impact risk of C-19 infection and its consequences?

vi) How do we best identify cases in population studies, triangulating symptoms, antigen, and antibody testing? How do antibody profiles differ by case status (e.g., those with and without symptoms), how do antibody profiles vary over time? What are the health determinants of such variability?

vii) What are the long-term socioeconomic consequences of viral suppression measures? How do these in turn impact on health and on health care utilisation?

These research themes have been scientifically reviewed by the UK Chief Scientific Advisor and the UK Chief Medical Officer and are high priority questions of critical national interest.

OTHER DATA PROCESSING PURPOSES

In addition to the primary research themes, UoB will:

1) use linked NHS data to help inform descriptive and documentary analysis of the UK LLC dataset. This will describe the combined LPS population and make clear the LLC denominator, its characteristics and how this sub-set of the population relates to the wider population (through comparison with aggregated populations statistics). This will inform research users and those (e.g., policy makers) seeking to draw inferences from findings.

2) conduct feasibility assessments to help determine the viability of proposed projects. To do this, UK LLC staff will query the UK LLC integrated dataset (LPS data and linked NHS Records) to determine if UK LLC have sufficient case numbers to be able to inform specific research proposals. This purpose enhances the ethical and governance case for the resource as it means data are not shared where they cannot meaningfully inform the science, it also helps ensure the efficient use of research funding and resources.

DATA MINIMISATION

The datasets and data items requested have been restricted to only those directly relevant to the Covid-19 research programme that this infrastructure is designed to support. UoB fully adopt the standard NHS England definition of the Covid-19 relevant dataset.

As described, the UK LLC C-19 research programme is wide ranging and seeks to understand patterns and predictors of infection, disease outcomes, the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. Datasets requested include measure of Covid-19 related outcomes and pre-Covid-19 baseline datasets (allowing the value of these longitudinal data to be maximised).

The linked data requested are minimised to include only data covering the period each contributing study was in follow-up. The data requested are further minimised to include only records of participants of studies contributing to the UK LLC and those who have objected to this use of their data are excluded. Information on participants who have ‘opted-out’ will be uploaded to the UK LLC quarterly so that those participants’ data will be excluded from future studies.

UoB are unable to minimise datasets further based on fields or episodes at this time as work continues to be conducted to inform the UK’s Covid-19 research programme which aims to be responsive to health and government policy makers. For example in HES, UoB cannot consider only episodes with a specific diagnosis or procedure at this time, UoB do not yet understand the longer-term health implications from Covid-19 infection or whether there exist patterns and predictors of Covid-19 infection based on previous episodes in HES.

The study have minimised datasets so they only include records of participants of longitudinal population studies contributing to the UK LLC and to NHS England data that are directly relevant to the research programme. Whilst UoB cannot provide justification through reference to specific analysis that will be conducted using these data, given their purpose as a resource developed to enable researchers to respond to the emerging priorities as determined by the UK Scientific Advisory Group for Emergencies.

1) The UK LLC is a resource to enable the longitudinal perspective. The value of longitudinal research lies in the collection of a broad range of data, repeated at regular intervals allowing the investigation of interactions between things occurring over the lifetime of individuals. To enable this perspective to be applied to the Covid-19 research questions UoB require data from NHS England and other providers that covers the duration of the longitudinal population studies contributing their study data. For example, the National Study of Health and Development is still collecting data from babies born in one week of 1946. Emerging Covid-19 research for example around disease severity could feasibility be informed by patterns in childhood ill health and historic episodes of hospitalisation within these cohorts.

2) As above with the duration of NHS-linked data requested, the breadth of data within the LLC enable rapid response to emerging research priorities. UoB cannot predict what these priorities will be but can provide indicative examples:

The usefulness of a QOF exception code in GDPPR indicating that a patient has refused an annual health check are that this could be indicative of their behaviour around choosing to have vaccination; insights into vaccination and health behaviour have the potential to be of great importance. The risk is there are unforeseen needs which are then hard to fulfil if the data has been minimised.

The UK LLC is designed to support longitudinal research studies and research analysts working within the UK. Longitudinal research is geared to understand the longer-term outcomes of the pandemic and emerging interactions between health and social factors and including considerations of outcomes such as Long COVID whose duration is yet to be determined, and the impact of disruption of routine health service provision on wider mental and physical health.

LEGAL BASIS, ETHICS AND RISK OF POTENTIAL HARM

UoB will be sole Data Controller under this agreement.

UoB is establishing the UK LLC and using linked NHS records to inform the LH&W NCS research programme as a task carried out in the public interest (Article 6(1)(e) given the UK LLC is designed to answer specific high-priority research questions at a time of national crisis. Given that the requested health records include flows of sensitive information, the UK LLC also draws on Article 9(2)(j) in accordance with Article 89(1) where UoB's purpose is to conduct statistical scientific research. The case for this research being in the public interest is established through balancing the strong policy driver of improved understanding of the pandemic with consideration of potential harms to the patients/participants whose records are involved. Potential harms primarily relate to breach of confidentiality and the subsequent misuse of Personal Data or erosion of trust in longitudinal research/data science.

The lawful basis for the release and use of the confidential data being shared under this version of the agreement is Regulation 3(4) of the National Health Service (Control of Patient Information Regulations) 2002 (COPI) to require NHS England to share confidential patient information with organisations entitled to process this under COPI for COVID-19 purposes. The only permitted activities under this Data Sharing Agreement (DSA) are for COVID-19 purposes and within bounds of Reg 3(2) COPI. Reg 3 (2) COPI states that: "2) For the purposes of this regulation, “processing” includes any operations, or set of operations set out in regulation 2(2) which are undertaken for the purposes set out in paragraph (1)." The research relates to the monitoring and managing of COVID-19 and would therefore be covered by Reg 3(1)(d) of COPI.

The risk of this harm is minimised as this is:

1. an observational population-level research which will not result in a direct intervention to any participant;

2. the UK LLC uses tried and tested IT infrastructure and governance frameworks specifically designed to minimise risks to privacy during health data science;

3. the UK LLC is a fully de-identified research environment with sufficient controls that the risk of disclosure is not considered reasonably likely;

4. a reasonable expectation for the use of NHS data in this way has been established by the contributing studies, with studies making best endeavours to inform participants ahead of data extraction taking place, as well as continued fair processing (in relation to UK LLC) in line with usual policies and practices. Each stage of this fair processing will provide participants with the opportunity to object;

5. National Opt-Out will also be respected unless overridden by specific consent for the use of linked health records in the studies research programme;

6. all staff and users are vetted and approved professional researchers operating within controlled and auditable conditions.

This research is ethical given the strong scientific rationale, the process is transparent, best endeavours will be made to inform participants of the use of their data within UK LLC, with a right to object, that risks are mitigated and the participants (and wider public) are likely to directly benefit from the research through improved health care and government policy provision in response to the unprecedented Covid-19 crisis. Participants will be informed about how their data is used through ongoing fair processing communications including details of the process for ‘opting-out’.

OPERATIONAL MANAGEMENT

UoB is the study sponsor and Data Controller. UK LLC staff at UoB and UoE have responsibility for day to day running of the UK LLC databank: only UoB staff will have access to the data and TRE. The UK LLC Director, principle and co-investigators and data managers are employed by UoB. The UK LLC Co-Director and UK LLC administrative staff (responsible for administrative operations, communications, and participant/public engagement) are based at UoE). The Longitudinal Health & Wellbeing National Core Study is led from University College London (UCL) and the UK LLC reports into UCL on delivery of UK LLC objectives.

The UK LLC databank is configured to have two distinct classes of operational areas, which have three functions:

1) Data management and processing: All operational control and access for data processing and management of de-identified, including but not limited to data processing and storage, descriptive analysis and report/publication writing, subset, and provision of data to approved users is restricted to UoB staff. UoB staff are the only individuals who can access all data across all studies and sources in their raw and processed forms.

2a) Onward sharing to contributing studies: A secure operating partition will be created for each contributing study. Within these linked LPS - NHS data, restricted to the participants of that study, can be access by approved study staff through the ‘onward sharing’ mechanism. The study staff will use this area to conduct descriptive analysis on their participants (e.g., to provide documentation to describe the data asset, to inform bias assessments) and to assess feasibility of projects (e.g., to determine whether there are sufficient cases to scientifically justify inclusion in a proposed project) to inform potential users and to conduct study-level due diligence.

2b) Onward sharing to research analysts: Secure operating partitions will be created for each UK LLC approved project. The UK LLC approved users for these projects will have access to their folder which will contain a sub-set of data relevant to that study where onward sharing conditions of data providers are met. The UK LLC research users will control of data processing and analysis and leading publications.

The UoB UK LLC staff and UK LLC infrastructure system administrators will have access across all three operational areas.

The University of Swansea (UoS) is a Data Processor who provide data infrastructure and data management services to UoB and will maintain the integrity of the TRE, will assist with the data management processing (Under UoB’s direction) and will conduct the output disclosure assessments (to UoB’s specification).

The UoS have sub-contracted the Data Linkage functions of the UKSeRP (including all management of participant identifiers) to the Digital Health and Care Wales (DHCW), also referred to as the UK LLC Trusted Third Party (TTP) within this agreement. DHCW's role is limited to the the handling of participant identifiers prior to being submitted to NHS England for linkage to NHSE datasets. DHCW do not access or process any data disseminated by NHS England.

The UK LLC is funded by HM Treasury as part of the National Core Studies for Covid-19 research and is administered by the UK Medical Research Council (MRC). HM Treasury and MRC will not have access to NHS England data.

SUB-LICENSING

For the UK LLC to be a useful and sustainable resource for researchers to interrogate, and obtain meaningful data in a timely manner, contractual arrangements will need to facilitate high-volume and rapid turn-around of data requests. This will only be possible through removing the burden of the need for a high number of additional data sharing agreements across multiple institutions which would impede the speed of research. Therefore, a contract structure is needed to control the data flows, to control the purposes and way these data are processed, and to assign roles and responsibilities and to enforce the governance requirements of the individual studies and data owners and the legal basis under which they are permitted to operate.

These agreements are as follows:

- Data Deposit Agreement(s) established between UoB and the studies institution(s), ensuring fair processing will be carried out, and allowing study collected data to be made available for Covid-19 research to approved users within the UK LLC TRE.

- Shared data will be deposited into TRE where it will be processed, curated, and analysed.

- TRE established through research collaboration with University of Swansea (UoS) controlled through a Data Deposit

Agreement between UoB and UoS comprising:

1) a data deposit agreement for Welsh records and other records (SymptomTracker) to be deposited in the UK LLC and used for research

2) a collaborative contract where UoB procure a UKSeRP (UoS as Data Processor) and UoS data science expertise (an academic collaboration which involves Data Processing). Some aspects of the UoS contribution will be fulfilled by sub-contractor Data Health and Care Wales (DHCW) acting as a Trusted Third Party controlled through the Sub-Contract Agreement between UoS and DHCW.

The contract structure also permits the processes by which the participants of these studies can be identified by the NHS and their records selected and provided into the UK LLC for integration with the study data and onward sharing with researchers. The agreements for selecting and depositing these data has two distinct parts:

1) Record Linkage

• UoB will instruct the contributing studies to send a file of participant identifiers to Digital Health and Care Wales (DHCW) (controlled through the Data Deposit Agreement between UoB and each contributing studies institution(s)).

• UoB will instruct the DHCW (a process controlled through the TRE procurement contract between UoB and UoS, and as described in the Data Deposit Agreement between UoB and contributing studies) to send participant identifiers to NHS England for record linkage purposes

• NHS England will use these identifiers to link these to their patient register and select the longitudinal study participants data (controlled through the Data Sharing Agreement between UoB and NHS England).

2) Data Deposition & Processing

• NHS England will then de-identify and deposit the selected data into the UK LCC where UoB will become the data controller of this patient data, for the purpose of processing it, approving onward research use and providing managed access to relevant sub-sets of data for Covid-19 research (controlled through the Data Sharing Agreement between UoB and NHS England). The structure will then enable the UK LLC to determine the onward processing of the integrated data for Covid-19 research purposes by reviewing applications from researchers and onwardly sharing sub-sets of relevant data within the TRE to approved users (controlled through the Data Access Agreement between UoB and approved researchers’ institutions). The nature of the sublicensing arrangements above are in line with the NHS England standards - https://digital.nhs.uk/services/data-access-request-service-dars/dars-guidance/sub-licencing-and-onward-sharing-of-data. NHS England requirement to audit data use is a requirement in the onward sharing contract and remains feasible in the UK LLC framework.

The UK LLC is likely to generate diverse outputs, with these forming a mix of rapid insight reports to policy makers and substantive research reports. The NCS is specifically designed to establish a clear channel between decision makers and researchers and the UK LLC is designed to enable rapid response and up-to-date information needed to support this. The likely outputs include:

* Peer-reviewed scientific publication (and pre-print publications of initial drafts prior to review which are designed enable rapid awareness of the findings), outputs will be ongoing, preprints and publications to date:

Preprints | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

Publications | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London

* Conference presentations and workshop presentations (to academic, health service practitioners and planners, government policy makers, Scientific Advisory Group for Emergencies (SAGE) members). These will be ongoing as part of our communications and engagement strategy, events to date include: Specifically for UK Longitudinal Linkage Collaboration:

Presentation: Welcome Longitudinal Population Studies conference (2021)

Regular presentations to Health Data Research UK - Data & Connectivity National Core Studies Delivery Group (2021-2022)

Presentation: CLOSER Longitudinal Population Studies conference (2022) ‘UK LLC Overview and explaining the innovations made and challenges faced’

Presentation: LHW National Core Study - Joint EHR cohort leads meeting ‘how to access UK LLC’

Workshop 1: to National Core Studies Analysts ‘Overview of UK LLC and how to access UK LLC’

Workshop 2: to CLOSER, National Core Studies and Vanguard Members ‘overview of UK LLC and how to access data

Presentation to UK LLC Communications Network (2022)

Youtube: UK Longitudinal Linkage Collaboration Verbal YouTube Public Involvement Induction (2022)

Presentation: to Elizabeth Blackwell Institute Health Data Science Showcase (2021)

Presentation: ESRC Strategic Advisory Group ‘ Understanding Data for Population research’ (2021)

Presentation: to Longitudinal Area Network and Data-link (LAND) community of practice (2022)

Poster Presentation: to Welcome Trust Longitudinal Studies Conference ‘overview of UK LLC’ (2021)

Presentation and advice: to Welcome Longitudinal Population Study COVID-19 Steering Group ‘provided specialist advice on linkage opportunities’ (2021)

Presentation and advice: to Population Research UK Programme Group ‘provided specialist advice’ (2021)

Advice: to Department of Education Scientific Advisory Group ‘provided specialist advice on record linkages in longitudinal studies (2021)

* Reports to health service practitioners and planners, government policy makers and Committee’s (e.g., National Institute for Health and Care Excellence (NICE), SAGE), it is anticipated that these will be ongoing, briefings to date include:

Briefing Reports | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

* Rapid evidence synthesis reports as requested by health and government policy makers (through the entirety of the 3-year NCS programme);

* Lay summaries provided through websites, Blogs, social media posts, privacy notices and outputs designed to promote the transparent use of data and the wider research process, regularly updated on our data-use register as follows:

Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk)

All outputs from the TRE will be assessed for disclosure risk and will be anonymous population level findings. The framework for this is based on leading 'five safes’ approach and is closely adapted from equivalent processes accredited to standards of the Office for National Statistics and the UK Statistics Authority.

All ‘meta’ products of the UK LLC research (such as code list definitions, syntax, workings for derived variables) will be made available for reuse and interrogation/replication. Access to the UK LLC is ‘open’ in the sense that any legitimate researcher conducting Covid-19 research can apply for access using a transparent and consistent decision-making process. As such, the UK LLC enhanced resource should be considered as an important ‘output’ and a measurable benefit of the UK LLC/LH&W NCS.

The high profile of Covid-19 means that many findings may have media interest. Importantly, the UK LLC is establishing a Patient/Participant Involvement and Engagement strategy within its communication work-package package with a dedicated budget and experienced Communications PPIE Officer. The package will build on existing and established relationships with participants to involve public/participants in the co-design of mechanisms to engage and effectively communicate UK LLC findings more widely with the public. UoB are committed to public involvement in our design and communications. UoB have already included HDRUK PPIE panel members (public representatives) in a consultation exercise and will draw on the extensive network of participants to further explore and co-develop dissemination mechanisms, in developing the end-to-end animation in the first instance. UoB are also developing with the studies opportunities for a subgroup of participants to be involved, in the data access decision making process for example. UoB have ensured that the programme of work is adequately resourced to compensate the public for their time, this has been an integral part of the project planning.

The UK LLC is a novel research infrastructure. The creation of this for Covid-19 research may provide methodological insights and illustrations of a new way of working for linkage in longitudinal research. Insights from this will be provided to the longitudinal community and its funders/key stakeholders (through academic papers, presentations, and contributions to advisory groups).

UK LLC has been a real success and is being sustained by UK Research and Innovation (UKRI) following its commissioning as part of the National Core Studies for COVID-19 research due to the improved scientific abilities for supporting longitudinal research, its cost efficiency for the UK taxpayer and NHS, its improved safeguards for participant and patient data, and its ability as a standing responsive research resource which is ready for crisis management and thus strengthens the UK’s crisis resilience capability.

The funders (UKRI), with support from NHS and other government departments, have identified UK LLC as an optimal model for providing linkage and a secure state of the art generic research database for the UK longitudinal research community, and recently reinvested for a further five years (until at least June 2028). To date, 25 LPS contribute data to the UK LLC TRE. There are 29 projects approved with data provisioned within the TRE to researchers across institutions with more projects in the pipeline: see UK LLC Data Use Register (https://ukllc.ac.uk/data-use-register/). With funding now secured and the benefits of the TRE way of working proven and communicated, UK LLC has been approached by 17 additional LPS wishing to join the collaboration (these will be added to this agreement through amendments as these arrangements are formalised)

UK LLC as an infrastructure is therefore a primary output of the National Core Studies LH&W programme. Reflecting this, the UK’s Chief Medical Officer has now added UK LLC to his list of strategic scientific infrastructure. Specific applied research outputs which exemplify the unique benefits of UK LLC include:

(1) Understanding the impacts of healthcare disruption (Lead Investigator: Green; UK LLC Project: https://ukllc.ac.uk/data-use-register/llc_0009): The pandemic resulted in unprecedented healthcare disruption. It would expected for disruption to increase hospital admission: yet, national GP data analysed in OpenSAFELY showed (currently unpublished) falls in absolute numbers and social inequalities in ambulatory care-sensitive conditions. This ecological paradox reflects reduced access to care and is misleading to the effects of disruption. Using UK LLC’s linked LPS and hospital admission records, Green and colleagues found that individuals’ reporting greater care disruption were more likely to be admitted to hospital, suggesting pandemic disruptions adversely influenced care management (BMJ, https://doi.org/10.1136/bmj-2023-075133). They also found social patterning in these outcomes. These findings were presented to NICE and SAGE and used in service planning considerations. This project demonstrates the key advantage of having LPS self-reported study measures of disruption (a more accurate measure to OpenSAFELY’s modelled disruption estimate), and that since hospital admission is relatively rare, both UK LLC’s linkage and pooled sample size were needed to inform this research.

(2) Comparing the burden of Long COVID in the community as measured by self-report and EHRs (Lead Investigator: Williams; UK LLC Project: https://ukllc.ac.uk/data-use-register/llc_0006): comparing OpenSAFELY national data with ONS COVID Infection Survey suggests large disparities between estimated Long COVID prevalence and individuals recorded with Long COVID in EHRs. Using UK LLC, researchers directly compared LPS self-reported Long COVID and NHS interactions with UK LLC’s linked GP diagnoses and referral codes. This identified the proportion of LPS Long COVID sufferers seeking GP care and coded as needing further treatment. Using eight UK LLC studies, researchers identified ~800 participants self-reporting Long COVID by spring 2021, <1.5% received a Long COVID diagnosis or referral code after >1 year of follow-up (see https://doi.org/10.1101/2023.02.10.23285717 for initial findings subject to peer review). With National Institute for Health and Care Excellence (NICE) UK LLC are exploring improved GP coding guidance and improvements to coding patterns (eg, GP system ‘pop ups’). This project demonstrates the ability of LPS and UK LLC to inform analysis of under-reported or under-coded conditions and assess sub-group effects. Through this, UK LLC can provide unique insights into population health and service/system requirements and insights into how to target upstream interventions and care.

(3) Multi-Longitudinal Cohort Study into occupational factors and COVID Risk as part of PROTECT National Core Study (Lead Investigator: Gittins; UK LLC Project: https://ukllc.ac.uk/data-use-register/llc_0007): assessed infection risks across job-role and work-related characteristics using self-reported employment status data from 14 LPS linked to NHS England health records, including COVID-19 diagnostic testing. SARS-CoV-2 infection and COVID-19 infection risk was greater in key-workers vs not; among non-home working or some home working vs all home working; while part-time workers vs full-time; and furlough vs not had reduced risk (see https://doi.org/10.1101/2023.12.19.23298502 for initial findings subject to peer review). This project demonstrates the near unique opportunities for researchers to investigate occupational health risk factors in the UK – given that job role is not recorded in any health or non-health routine system (outside of the national census programme). For this reason, with its national coverage, large and diverse sample size, and linkages to non-health records, UK LLC is emerging as a national resource for occupational health research. In collaboration with the Health & Safety Executive (HSE), UK LLC have been working with occupational health experts to enhance the resource through harmonising job-role and other work-based study data to support a breadth of occupational research and to inform HSE’s 10-year strategy to reduce work-based mental ill-health.

Findings from the LH&W NCS have led to Cabinet office and NHS England 'teach in' sessions, a national GP alert, an NHS enhanced service specification , rapid reports for SAGE, and they have provided key evidence to guide the NICE long COVID guidelines.

• Policy makers have been provided with information on long COVID burden of disease, risk factors, definition and long term outcomes via cabinet briefing reports and SAGE reports

• Low levels of long COVID GP coding were reported. This led to an NHS enhanced service specification, directed at GPs to drive an increase in long COVID coding

• NICE used long COVID findings in its evidence to review risk factors

• Findings on occupational risk fed into Health & Safety Executive policy decision making and UK LLC is being considered by HSE (who have provided funding to scope this) as a strategic resource for future mental ill-health policy research

• Briefing notes submitted addressing inequalities in healthcare disruption, mental health decline during lockdown and the impact of furlough on health behaviour and adverse outcomes.

Throughout the project UK LLC has worked with the public via our advisory groups, to shape and communicate our programme. Public contributors have helped to create material aimed at communicating the results to all stakeholders, including infographics, short videos and an animation. UK LLC is now providing a methodological exemplar for the emerging NHS Secure Data Environment’s (SDEs) at national and sub-national level with UK LLC contributing to SDE policy design, the accreditation framework and being identified as a TRE with best practice in public involvement, transparency of operations and as leading in governance design and practice. UK LLC governance materials and processes are directly contributing to NHS and Office for National Statistics practice.

There is now substantial and increasing demand for UK LLC to support non-COVID research, and for new LPS to join the collaboration - although neither of these purposes will be permitted through this DSA, it does indicate the interest and demand for UK LLC functionality given its successful COVID-19 research outputs.

University of Bristol (UoB) will continue to have legal basis under Regulation 3 of the COPI Regulations for processing CPI for Covid-19 Purpose. Where UoB will use the data for previously approved research related to a Covid-19 Purpose, and this research is still ongoing, UoB are able to continue to meet the requirements in Regs 3(1), 3(3) and 7 of the COPI Regulations. DARS have assessed that UoB do still meet the 3 limbs of the COPI regulation.

The UK Longitudinal Linkage Collaboration (UK LLC) project itself is cross-institutional and has staff based at University of Bristol and the University of Edinburgh. The University of Edinburgh are collaborators however they have no role in determining how the data will be processed, nor will they have any access to the record level data. For clarification regarding the distinct staff groups noted in the agreement:

Study staff = staff from the contributing cohorts;

University of Bristol (UoB) staff = staff of UK Longitudinal Linkage Collaboration (LLC) employed by UoB;

UK LLC staff = staff employed by UK LLC (UoB and University of Edinburgh (UoE) staff).

BACKGROUND

The UK Chief Scientific Advisor has established (October 2020) a programme of National Core Studies (NCS) for SARS-CoV-2 (Covid-19) research as a coordinated, long-term, national research initiative. This will consider Covid-19 in terms of a viral pandemic (including issues of cases, transmission, symptoms, and outcomes) and in terms of the health and social impacts of behavioural restrictions designed to mitigate the harms of the pandemic. The NCS has six different sub-programmes which are addressing major Covid-19 research areas; one of these is the Longitudinal Health and Well-being (LH&W) NCS which is designed to use data from longitudinal studies to address the impact of Covid-19 and of associated viral suppression measures on health and well-being. The LLC is the central hub component of the Longitudinal Health and Well-being NCS.

The UK LLC has been designed to underpin the LH&W NCS, although not exclusively as the NCS are designed to support each other, where for example: the LLC could form the infrastructure for the long-term patient follow-up of consenting trial participants (e.g. participants in the RECOVERY trial) or, where UK longitudinal population studies (LPS) are being used to collect specific new study data which is not available through routine records (e.g. the ALSPAC birth cohort is providing a frame to collect biological samples to inform the Immunology NCS research programme; the assayed results of which will be uploaded into the LLC). For this reason, users of the LLC may come from across the full range of NCS studies and the resource will be accessible to other legitimate UK-based researchers investigating Covid-19 through a sub-licence framework. The NCS is planned to be a two/three-year research programme commencing October 2020.

THE UK LLC & DATA SUBJECTS

UoB (for LLC purposes) will underpin the NCS programme by combining, within a Trusted Research Environment, study data from >15 major inter-disciplinary UK LPS with a combined total of 1 – 2 million UK participants, centrally link to a wide range of Covid-19 relevant and non-health administrative records to utilise alongside the UK Biobank, the Zoe Symptom Tracker cohort and other sources. The integrated data, infrastructure and accompanying governance aspects will collectively be known as the UK LLC.

The de-identified data from contributing studies will be used to establish a UK LLC participant and household register, this will be used within the UK LLC to index and catalogue available data. Researchers will be able request access to the integrated data via the Health Data Research UK Gateway (https://www.healthdatagateway.org/) which through the NCS programme is being established as the de facto application point for UK Data Science projects and as a systematic and coordinated infrastructure for assessing the legitimacy and ‘safe’ credentials of a project. The value of these LPS data assets will be significantly enhanced through linkage to Covid-19 relevant health and other routine records (e.g., employment status, occupation, and benefit receipt records). This will allow the use of objectively assessed diagnoses and other records and the collection of data not amenable to self-report (such as detailed records of care, or health service utilisation). It will also allow researchers to assess and quantify/document bias and error through triangulating data collected across different organisations and settings.

Throughout the development of the UK LLC, UoB have worked collaboratively with UK LPS to co-design and develop the protocol and methodology. One of the strengths of the collaborative approach championed by the UK LLC is that UK LPS are committed and enthusiastic partners in this work: this initiative comes from within the community. The UK LLC will provide infrastructure to support the LPS and researchers working within the UK to investigate high priority COVID-19 research questions. UK LLC/UoB will continue to seek input from the LPS into decision making, guidance in future development and operation. Crucially, the LPS’s all maintain and manage ‘their’ relationships with ‘their’ participants including making critical decisions about data use, and communicating with participants about the existence of, and work of, the UK LLC and providing reassurance that this does not alter the participants existing relationship.

Data for the following studies will flow under COPI:

1. The National Study of Health and Development (NSHD)

2. The Southall And Brent REvisited (SABRE)

3. Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”)

4. Genetic Links to Anxiety and Depression Study (GLAD)

5. English Longitudinal Study of Ageing (ELSA)

6. 1958 National Child Development Study (NCDS)

7. 1970 British Cohort Study (BCS 70);

8. Next Steps

9. The Millenium Cohort Study (MCS)

10. Track-COVID - subset of INTERVAL COMPARE and STRIDES bioresource

11. NIHR (National Institute for Health Research) BioResouce

12. Extended Cohort for E-health, environment, and DNA (EXCEED)

13. Understanding Society

14. Born in Bradford

15. European Prospective Investigation into Cancer (EPIC)

16. The Fenland Study

17. Twins early Development Study (TEDS)

18. INTERVAL

19. COMPARE

20. STRIDES

21. TwinsUK

There is a longer-term objective to include a larger number of studies into the UK LLC to boost statistical power, increase participant heterogeneity and sample size in population and outcome sub-groups and to increase the range of self-reported information. Any additional study would be included as an amendment to this application.

The UoB scientific programme requires the extraction and use of the NHS England data compiled with Covid-19 relevant datasets (primary care, secondary care, community, mental health care provision, Covid-19 testing and outcomes data, NHS service use interactions such as NHS 111 records, mortality, disease registry and demographics data) for the purpose of establishing the UK LLC.

The UK LLC has conducted an assessment as to whether the data held and processed within the UK LLC is identifiable/potentially identifiable and therefore Personal Data which is subject to the Data Protection Act 2018 (DPA), UK General Data Protection Regulations (UK GDPR) and Common Law Duty of Confidentiality (Common Law) in accordance with guidance issued by the Information Commissioner’s Office. The assessment determined that the infrastructure, incorporating all data flows, should be considered as containing Personal Data and being subject to the DPA/UK GDPR and Common Law requirements. However, within this the assessment identified that the UK LLC sits at the threshold of being considered as holding pseudonymous data (Personal Data) and de-identified data where the risk of re-identification is not reasonably likely (not Personal Data), and that this differing status manifests when considering the data from the perspective of the different high-level groups of UK LLC data users and their distinct processing activities. It is evident that for the contributing studies, the use of data is either identifiable (for the provision of identifiers for linkage purposes) or pseudonymous (for the provision of data into the UK LLC and the processing of study data by study staff in their demarcated study processing area) and is therefore Personal Data. Yet, for the UK LLC staff and researcher users, the risk of re-identification of data within the UK LLC is not reasonably likely. This status, for these users, has been achieved through the deployment of data level as well as contextual controls which mitigate the risk of identification. The principal amongst these is the UK LLC operates as a ‘Trusted Research Environment’ which ensures that minimised data are used by minimised number of legitimate users within a secure environment for approved and proportionate purposes.

To contribute data into the UK LLC, LPS’s will be required to flow identifiers to the NHS to establish/filter a linkage between each participants' study record and their NHS registry record(s). Studies will update their privacy notices and notify participants of their involvement with UK LLC, explaining what this means and provide options to ‘opt-out’, in line with existing policy and practice.

UK LLC TRUSTED RESEARCH ENVIRONMENT (TRE)

The LLC will be built on a UK Secure eResearch Platform (UKSeRP). This is a well-established infrastructure for data science: it has ISO27001 and NHS Data Security Protection toolkit and Office for National Statistics accreditation under the Digital Economy Act 2017. Other instances of this environment are already used to host LPS and NHS England data and have been audited and commended as being well suited to the purpose and low risk.

The UK LLC TRE will be owned, operated, and managed by the LLC team (data management based at UoB). Under instruction from UoB, contributing LPS and data owners including NHS England will provide data to the UK LLC TRE using the ‘split file’ approach which is used across UKSeRP. This split file approach (details below) ensures the physical and processing operational separation of personal identifiers and participants de-identified study data and health records. This separation means that no party involved in this process can see both identifiers and data. This split file approach is coupled with the technological and socio-governance controls applied at the UK LLC (e.g., training, contracts, policy and procedures, penalties for misuse). The UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. For UK LLC staff and research users the risk of re-identification is controlled to the point that it is not reasonably likely.

UK LLC C-19 RESEARCH PROGRAMME

The UK LLC is specifically commissioned as a database resource to inform the LH&W NCS but is also designed to be a resource for the investigation of other Covid-19 questions as these are identified and prioritised through the Scientific Advisory Group for Emergencies (see the Health Data Research UK and SAGE research prioritisation funnel - current versions available through the HDRUK SAGE report, see - https://www.hdruk.ac.uk/covid-19/).

Given these questions are likely to change over time (e.g., the identification of ‘Long COVID’ as a potential new set of complex outcomes is introducing important new research questions), the research programme the LLC is designed to inform can be defined as the investigation of Covid-19 questions using data from longitudinal studies linked to routine records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing. For the use of linked longitudinal studies and NHS records this definition is refined to make clear the investigations primary outcomes must be focused on generating benefits to the health and social care system.

The investigation of Covid-19 questions using data from longitudinal studies linked to routine NHS records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing: where the primary research outcome is designed to benefit the health and social care system. These linked data will be of unique value in underpinning a programme of research on Covid-19 informed by the data assets within the UK LLC.

This program will seek to understand the patterns and predictors of infection, (including re-infection) and disease outcomes (such as ‘long covid’), and the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. It will also explore population level changes to physical and mental health, including hospital admission and mortality, in association with viral suppression measures and how these relate to changes in health behaviours.

Given the impact of the Covid-19 containment measures, it is important to investigate the role of socio-economic and neighbourhood/environmental factors in determining population level impacts to physical and mental health, to identify both groups at risk, and factors which offer resilience to adverse outcomes. Within this, UK LLC users will analyse changes in health care service use (using self-reported data and NHS health records) to determine if patterns in these have changed during the Covid-19 pandemic. Linked NHS records will also inform consideration of service use in relation to pre-pandemic health status and regular service interactions (e.g., screening, health reviews and routine service take up (such as annual seasonal flu vaccinations)).

This programme, through the LH&W NCS will specifically seek to answer the following questions:

i) What are the mental health consequences of Covid-19 infection, and of viral suppression measures? How do these differ by key demographics, socioeconomic status and by prior mental health status?

ii) What are the risks, determinants, and expressions of the physical health consequences of long-Covid?

iii) What are the disruptions to health care services (primary and secondary), and what are the health consequences of these disruptions? How do they vary by geography, demographics, and socioeconomic status?

iv) What are the risks of re-infection?

v) How do measures of historical health & physiology impact risk of C-19 infection and its consequences?

vi) How do we best identify cases in population studies, triangulating symptoms, antigen, and antibody testing? How do antibody profiles differ by case status (e.g., those with and without symptoms), how do antibody profiles vary over time? What are the health determinants of such variability?

vii) What are the long-term socioeconomic consequences of viral suppression measures? How do these in turn impact on health and on health care utilisation?

These research themes have been scientifically reviewed by the UK Chief Scientific Advisor and the UK Chief Medical Officer and are high priority questions of critical national interest.

OTHER DATA PROCESSING PURPOSES

In addition to the primary research themes, UoB will:

1) use linked NHS data to help inform descriptive and documentary analysis of the UK LLC dataset. This will describe the combined LPS population and make clear the LLC denominator, its characteristics and how this sub-set of the population relates to the wider population (through comparison with aggregated populations statistics). This will inform research users and those (e.g., policy makers) seeking to draw inferences from findings.

2) conduct feasibility assessments to help determine the viability of proposed projects. To do this, UK LLC staff will query the UK LLC integrated dataset (LPS data and linked NHS Records) to determine if UK LLC have sufficient case numbers to be able to inform specific research proposals. This purpose enhances the ethical and governance case for the resource as it means data are not shared where they cannot meaningfully inform the science, it also helps ensure the efficient use of research funding and resources.

DATA MINIMISATION

The datasets and data items requested have been restricted to only those directly relevant to the Covid-19 research programme that this infrastructure is designed to support. UoB fully adopt the standard NHS England definition of the Covid-19 relevant dataset.

As described, the UK LLC C-19 research programme is wide ranging and seeks to understand patterns and predictors of infection, disease outcomes, the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. Datasets requested include measure of Covid-19 related outcomes and pre-Covid-19 baseline datasets (allowing the value of these longitudinal data to be maximised).

The linked data requested are minimised to include only data covering the period each contributing study was in follow-up. The data requested are further minimised to include only records of participants of studies contributing to the UK LLC and those who have objected to this use of their data are excluded. Information on participants who have ‘opted-out’ will be uploaded to the UK LLC quarterly so that those participants’ data will be excluded from future studies.

UoB are unable to minimise datasets further based on fields or episodes at this time as work continues to be conducted to inform the UK’s Covid-19 research programme which aims to be responsive to health and government policy makers. For example in HES, UoB cannot consider only episodes with a specific diagnosis or procedure at this time, UoB do not yet understand the longer-term health implications from Covid-19 infection or whether there exist patterns and predictors of Covid-19 infection based on previous episodes in HES.

The study have minimised datasets so they only include records of participants of longitudinal population studies contributing to the UK LLC and to NHS England data that are directly relevant to the research programme. Whilst UoB cannot provide justification through reference to specific analysis that will be conducted using these data, given their purpose as a resource developed to enable researchers to respond to the emerging priorities as determined by the UK Scientific Advisory Group for Emergencies.

1) The UK LLC is a resource to enable the longitudinal perspective. The value of longitudinal research lies in the collection of a broad range of data, repeated at regular intervals allowing the investigation of interactions between things occurring over the lifetime of individuals. To enable this perspective to be applied to the Covid-19 research questions UoB require data from NHS England and other providers that covers the duration of the longitudinal population studies contributing their study data. For example, the National Study of Health and Development is still collecting data from babies born in one week of 1946. Emerging Covid-19 research for example around disease severity could feasibility be informed by patterns in childhood ill health and historic episodes of hospitalisation within these cohorts.

2) As above with the duration of NHS-linked data requested, the breadth of data within the LLC enable rapid response to emerging research priorities. UoB cannot predict what these priorities will be but can provide indicative examples:

The usefulness of a QOF exception code in GDPPR indicating that a patient has refused an annual health check are that this could be indicative of their behaviour around choosing to have vaccination; insights into vaccination and health behaviour have the potential to be of great importance. The risk is there are unforeseen needs which are then hard to fulfil if the data has been minimised.

The UK LLC is designed to support longitudinal research studies and research analysts working within the UK. Longitudinal research is geared to understand the longer-term outcomes of the pandemic and emerging interactions between health and social factors and including considerations of outcomes such as Long COVID whose duration is yet to be determined, and the impact of disruption of routine health service provision on wider mental and physical health.

LEGAL BASIS, ETHICS AND RISK OF POTENTIAL HARM

UoB will be sole Data Controller under this agreement.

UoB is establishing the UK LLC and using linked NHS records to inform the LH&W NCS research programme as a task carried out in the public interest (Article 6(1)(e) given the UK LLC is designed to answer specific high-priority research questions at a time of national crisis. Given that the requested health records include flows of sensitive information, the UK LLC also draws on Article 9(2)(j) in accordance with Article 89(1) where UoB's purpose is to conduct statistical scientific research. The case for this research being in the public interest is established through balancing the strong policy driver of improved understanding of the pandemic with consideration of potential harms to the patients/participants whose records are involved. Potential harms primarily relate to breach of confidentiality and the subsequent misuse of Personal Data or erosion of trust in longitudinal research/data science.

The lawful basis for the release and use of the confidential data being shared under this version of the agreement is Regulation 3(4) of the National Health Service (Control of Patient Information Regulations) 2002 (COPI) to require NHS England to share confidential patient information with organisations entitled to process this under COPI for COVID-19 purposes. The only permitted activities under this Data Sharing Agreement (DSA) are for COVID-19 purposes and within bounds of Reg 3(2) COPI. Reg 3 (2) COPI states that: "2) For the purposes of this regulation, “processing” includes any operations, or set of operations set out in regulation 2(2) which are undertaken for the purposes set out in paragraph (1)." The research relates to the monitoring and managing of COVID-19 and would therefore be covered by Reg 3(1)(d) of COPI.

The risk of this harm is minimised as this is:

1. an observational population-level research which will not result in a direct intervention to any participant;

2. the UK LLC uses tried and tested IT infrastructure and governance frameworks specifically designed to minimise risks to privacy during health data science;

3. the UK LLC is a fully de-identified research environment with sufficient controls that the risk of disclosure is not considered reasonably likely;

4. a reasonable expectation for the use of NHS data in this way has been established by the contributing studies, with studies making best endeavours to inform participants ahead of data extraction taking place, as well as continued fair processing (in relation to UK LLC) in line with usual policies and practices. Each stage of this fair processing will provide participants with the opportunity to object;

5. National Opt-Out will also be respected unless overridden by specific consent for the use of linked health records in the studies research programme;

6. all staff and users are vetted and approved professional researchers operating within controlled and auditable conditions.

This research is ethical given the strong scientific rationale, the process is transparent, best endeavours will be made to inform participants of the use of their data within UK LLC, with a right to object, that risks are mitigated and the participants (and wider public) are likely to directly benefit from the research through improved health care and government policy provision in response to the unprecedented Covid-19 crisis. Participants will be informed about how their data is used through ongoing fair processing communications including details of the process for ‘opting-out’.

OPERATIONAL MANAGEMENT

UoB is the study sponsor and Data Controller. UK LLC staff at UoB and UoE have responsibility for day to day running of the UK LLC databank: only UoB staff will have access to the data and TRE. The UK LLC Director, principle and co-investigators and data managers are employed by UoB. The UK LLC Co-Director and UK LLC administrative staff (responsible for administrative operations, communications, and participant/public engagement) are based at UoE). The Longitudinal Health & Wellbeing National Core Study is led from University College London (UCL) and the UK LLC reports into UCL on delivery of UK LLC objectives.

The UK LLC databank is configured to have two distinct classes of operational areas, which have three functions:

1) Data management and processing: All operational control and access for data processing and management of de-identified, including but not limited to data processing and storage, descriptive analysis and report/publication writing, subset, and provision of data to approved users is restricted to UoB staff. UoB staff are the only individuals who can access all data across all studies and sources in their raw and processed forms.

2a) Onward sharing to contributing studies: A secure operating partition will be created for each contributing study. Within these linked LPS - NHS data, restricted to the participants of that study, can be access by approved study staff through the ‘onward sharing’ mechanism. The study staff will use this area to conduct descriptive analysis on their participants (e.g., to provide documentation to describe the data asset, to inform bias assessments) and to assess feasibility of projects (e.g., to determine whether there are sufficient cases to scientifically justify inclusion in a proposed project) to inform potential users and to conduct study-level due diligence.

2b) Onward sharing to research analysts: Secure operating partitions will be created for each UK LLC approved project. The UK LLC approved users for these projects will have access to their folder which will contain a sub-set of data relevant to that study where onward sharing conditions of data providers are met. The UK LLC research users will control of data processing and analysis and leading publications.

The UoB UK LLC staff and UK LLC infrastructure system administrators will have access across all three operational areas.

The University of Swansea (UoS) is a Data Processor who provide data infrastructure and data management services to UoB and will maintain the integrity of the TRE, will assist with the data management processing (Under UoB’s direction) and will conduct the output disclosure assessments (to UoB’s specification). The UoS have sub-contracted the Data Linkage functions of the UKSeRP (including all management of participant identifiers) to the Digital Health and Care Wales (DHCW) (part of Velindre University NHS Trust) who as such, are Data Processors for UoB.

The UK LLC is funded by HM Treasury as part of the National Core Studies for Covid-19 research and is administered by the UK Medical Research Council (MRC). HM Treasury and MRC will not have access to NHS England data.

SUB-LICENSING

For the UK LLC to be a useful and sustainable resource for researchers to interrogate, and obtain meaningful data in a timely manner, contractual arrangements will need to facilitate high-volume and rapid turn-around of data requests. This will only be possible through removing the burden of the need for a high number of additional data sharing agreements across multiple institutions which would impede the speed of research. Therefore, a contract structure is needed to control the data flows, to control the purposes and way these data are processed, and to assign roles and responsibilities and to enforce the governance requirements of the individual studies and data owners and the legal basis under which they are permitted to operate.

These agreements are as follows:

- Data Deposit Agreement(s) established between UoB and the studies institution(s), ensuring fair processing will be carried out, and allowing study collected data to be made available for Covid-19 research to approved users within the UK LLC TRE.

- Shared data will be deposited into TRE where it will be processed, curated, and analysed.

- TRE established through research collaboration with University of Swansea (UoS) controlled through a Data Deposit

Agreement between UoB and UoS comprising:

1) a data deposit agreement for Welsh records and other records (SymptomTracker) to be deposited in the UK LLC and used for research

2) a collaborative contract where UoB procure a UKSeRP (UoS as Data Processor) and UoS data science expertise (an academic collaboration which involves Data Processing). Some aspects of the UoS contribution will be fulfilled by sub-contractor Data Health and Care Wales (DHCW) acting as a Trusted Third Party controlled through the Sub-Contract Agreement between UoS and DHCW.

The contract structure also permits the processes by which the participants of these studies can be identified by the NHS and their records selected and provided into the UK LLC for integration with the study data and onward sharing with researchers. The agreements for selecting and depositing these data has two distinct parts:

1) Record Linkage

• UoB will instruct the contributing studies to send a file of participant identifiers to NHS Wales Informatic Service (controlled through the Data Deposit Agreement between UoB and each contributing studies institution(s)).

• UoB will instruct the DHCW (a process controlled through the TRE procurement contract between UoB and UoS, and as described in the Data Deposit Agreement between UoB and contributing studies) to send participant identifiers to NHS England for record linkage purposes

• NHS England will use these identifiers to link these to their patient register and select the longitudinal study participants data (controlled through the Data Sharing Agreement between UoB and NHS England).

2) Data Deposition & Processing

• NHS England will then de-identify and deposit the selected data into the UK LCC where UoB will become the data controller of this patient data, for the purpose of processing it, approving onward research use and providing managed access to relevant sub-sets of data for Covid-19 research (controlled through the Data Sharing Agreement between UoB and NHS England). The structure will then enable the UK LLC to determine the onward processing of the integrated data for Covid-19 research purposes by reviewing applications from researchers and onwardly sharing sub-sets of relevant data within the TRE to approved users (controlled through the Data Access Agreement between UoB and approved researchers’ institutions). The nature of the sublicensing arrangements above are in line with the NHS England standards - https://digital.nhs.uk/services/data-access-request-service-dars/dars-guidance/sub-licencing-and-onward-sharing-of-data. NHS England requirement to audit data use is a requirement in the onward sharing contract and remains feasible in the UK LLC framework.

The UK LLC is likely to generate diverse outputs, with these forming a mix of rapid insight reports to policy makers and substantive research reports. The NCS is specifically designed to establish a clear channel between decision makers and researchers and the UK LLC is designed to enable rapid response and up-to-date information needed to support this. The likely outputs include:

* Peer-reviewed scientific publication (and pre-print publications of initial drafts prior to review which are designed enable rapid awareness of the findings), outputs will be ongoing, preprints and publications to date:

Preprints | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

Publications | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London

* Conference presentations and workshop presentations (to academic, health service practitioners and planners, government policy makers, Scientific Advisory Group for Emergencies (SAGE) members). These will be ongoing as part of our communications and engagement strategy, events to date include: Specifically for UK Longitudinal Linkage Collaboration:

Presentation: Welcome Longitudinal Population Studies conference (2021)

Regular presentations to Health Data Research UK - Data & Connectivity National Core Studies Delivery Group (2021-2022)

Presentation: CLOSER Longitudinal Population Studies conference (2022) ‘UK LLC Overview and explaining the innovations made and challenges faced’

Presentation: LHW National Core Study - Joint EHR cohort leads meeting ‘how to access UK LLC’

Workshop 1: to National Core Studies Analysts ‘Overview of UK LLC and how to access UK LLC’

Workshop 2: to CLOSER, National Core Studies and Vanguard Members ‘overview of UK LLC and how to access data

Presentation to UK LLC Communications Network (2022)

Youtube: UK Longitudinal Linkage Collaboration Verbal YouTube Public Involvement Induction (2022)

Presentation: to Elizabeth Blackwell Institute Health Data Science Showcase (2021)

Presentation: ESRC Strategic Advisory Group ‘ Understanding Data for Population research’ (2021)

Presentation: to Longitudinal Area Network and Data-link (LAND) community of practice (2022)

Poster Presentation: to Welcome Trust Longitudinal Studies Conference ‘overview of UK LLC’ (2021)

Presentation and advice: to Welcome Longitudinal Population Study COVID-19 Steering Group ‘provided specialist advice on linkage opportunities’ (2021)

Presentation and advice: to Population Research UK Programme Group ‘provided specialist advice’ (2021)

Advice: to Department of Education Scientific Advisory Group ‘provided specialist advice on record linkages in longitudinal studies (2021)

* Reports to health service practitioners and planners, government policy makers and Committee’s (e.g., National Institute for Health and Care Excellence (NICE), SAGE), it is anticipated that these will be ongoing, briefings to date include:

Briefing Reports | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

* Rapid evidence synthesis reports as requested by health and government policy makers (through the entirety of the 3-year NCS programme);

* Lay summaries provided through websites, Blogs, social media posts, privacy notices and outputs designed to promote the transparent use of data and the wider research process, regularly updated on our data-use register as follows:

Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk)

All outputs from the TRE will be assessed for disclosure risk and will be anonymous population level findings. The framework for this is based on leading 'five safes’ approach and is closely adapted from equivalent processes accredited to standards of the Office for National Statistics and the UK Statistics Authority.

All ‘meta’ products of the UK LLC research (such as code list definitions, syntax, workings for derived variables) will be made available for reuse and interrogation/replication. Access to the UK LLC is ‘open’ in the sense that any legitimate researcher conducting Covid-19 research can apply for access using a transparent and consistent decision-making process. As such, the UK LLC enhanced resource should be considered as an important ‘output’ and a measurable benefit of the UK LLC/LH&W NCS.

The high profile of Covid-19 means that many findings may have media interest. Importantly, the UK LLC is establishing a Patient/Participant Involvement and Engagement strategy within its communication work-package package with a dedicated budget and experienced Communications PPIE Officer. The package will build on existing and established relationships with participants to involve public/participants in the co-design of mechanisms to engage and effectively communicate UK LLC findings more widely with the public. UoB are committed to public involvement in our design and communications. UoB have already included HDRUK PPIE panel members (public representatives) in a consultation exercise and will draw on the extensive network of participants to further explore and co-develop dissemination mechanisms, in developing the end-to-end animation in the first instance. UoB are also developing with the studies opportunities for a subgroup of participants to be involved, in the data access decision making process for example. UoB have ensured that the programme of work is adequately resourced to compensate the public for their time, this has been an integral part of the project planning.

The UK LLC is a novel research infrastructure. The creation of this for Covid-19 research may provide methodological insights and illustrations of a new way of working for linkage in longitudinal research. Insights from this will be provided to the longitudinal community and its funders/key stakeholders (through academic papers, presentations, and contributions to advisory groups).

There are ten projects approved with data provisioned within the TRE to 20 researchers across seven institutions with six more projects in the pipeline. Key questions being addressed are: comparing the burden of long COVID in the community; mental health and wellbeing consequences of contracting COVID19;impacts of the COVID-19 pandemic on access to elective care in the UK. A full list of approved projects can be found here: Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk).

Findings from the LH&W NCS have led to Cabinet office and NHS England 'teach in' sessions, a national GP alert, an NHS enhanced service specification, rapid reports for SAGE, and they have provided key evidence to guide the NICE long COVID guidelines.

• During lockdown the LHW Team identified instances of inappropriate switching of blood thinners, triggering a national alert to GPs

• Policy makers have been provided with information on long COVID burden of disease, risk factors, definition and long term outcomes via cabinet briefing reports and SAGE reports

• Low levels of long COVID GP coding were reported. This led to a NHS enhanced service specification, directed at GPs to drive an increase in long COVID coding

• NICE used long COVID findings in its evidence to review risk factors

• Briefing notes submitted addressing inequalities in healthcare disruption, mental health decline during lockdown and the impact of furlough on health behaviour an mental health reasons for a low antibody response.

• LHW work was amongst the first to provide Chief Medical Officer with a quantification of the first vaccine side effects and identify differences in vaccine update among ethnic groups, and point out that people with learning disabilities should be prioritised.

The team have also presented early results from the first wave of the convalescence study qualitative interviews to NHS England, highlighting that long COVID participants are failing at the first hurdle of accessing healthcare. The team also communicated that there is an increased risk of developing cardiovascular complications up to a year after covid-19 infection, even in the absence of persistent symptom reporting.

Throughout the project we have worked with the public via our advisory group, to shape and communicate our programme. Public contributors have helped to create material aimed at communicating the results to all stakeholders, including infographics, short videos and an animation. We have also created a long COVID forum to host a diverse and inclusive public/patient conversation on defining and researching long COVID.

UoB will continue to have legal basis under Regulation 3 of the COPI Regulations for processing CPI for Covid-19 Purpose. Where University of Bristol (UoB) will use the data for previously approved research related to a Covid-19 Purpose, and this research is still ongoing, UoB are able to continue to meet the requirements in Regs 3(1), 3(3) and 7 of the COPI Regulations. DARS have assessed that UoB do still meet the 3 limbs of the COPI regulation.

The UK Longitudinal Linkage Collaboration (UK LLC) project itself is cross-institutional and has staff based at University of Bristol and the University of Edinburgh. The University of Edinburgh are collaborators however they have no role in determining how the data will be processed, nor will they have any access to the record level data. For clarification regarding the distinct staff groups noted in the agreement:

Study staff = staff from the contributing cohorts;

University of Bristol (UoB) staff = staff of UK Longitudinal Linkage Collaboration (LLC) employed by UoB;

UK LLC staff = staff employed by UK LLC (UoB and University of Edinburgh (UoE) staff).

BACKGROUND

The UK Chief Scientific Advisor has established (October 2020) a programme of National Core Studies (NCS) for SARS-CoV-2 (Covid-19) research as a coordinated, long-term, national research initiative. This will consider Covid-19 in terms of a viral pandemic (including issues of cases, transmission, symptoms, and outcomes) and in terms of the health and social impacts of behavioural restrictions designed to mitigate the harms of the pandemic. The NCS has six different sub-programmes which are addressing major Covid-19 research areas; one of these is the Longitudinal Health and Well-being (LH&W) NCS which is designed to use data from longitudinal studies to address the impact of Covid-19 and of associated viral suppression measures on health and well-being. The LLC is the central hub component of the Longitudinal Health and Well-being NCS.

The UK LLC has been designed to underpin the LH&W NCS, although not exclusively as the NCS are designed to support each other, where for example: the LLC could form the infrastructure for the long-term patient follow-up of consenting trial participants (e.g. participants in the RECOVERY trial) or, where UK longitudinal population studies (LPS) are being used to collect specific new study data which is not available through routine records (e.g. the ALSPAC birth cohort is providing a frame to collect biological samples to inform the Immunology NCS research programme; the assayed results of which will be uploaded into the LLC). For this reason, users of the LLC may come from across the full range of NCS studies and the resource will be accessible to other legitimate UK-based researchers investigating Covid-19 through a sub-licence framework. The NCS is planned to be a two/three-year research programme commencing October 2020.

THE UK LLC & DATA SUBJECTS

UoB (for LLC purposes) will underpin the NCS programme by combining, within a Trusted Research Environment, study data from >15 major inter-disciplinary UK LPS with a combined total of 1 – 2 million UK participants, centrally link to a wide range of Covid-19 relevant and non-health administrative records to utilise alongside the UK Biobank, the Zoe Symptom Tracker cohort and other sources. The integrated data, infrastructure and accompanying governance aspects will collectively be known as the UK LLC.

The de-identified data from contributing studies will be used to establish a UK LLC participant and household register, this will be used within the UK LLC to index and catalogue available data. Researchers will be able request access to the integrated data via the Health Data Research UK Gateway (https://www.healthdatagateway.org/) which through the NCS programme is being established as the de facto application point for UK Data Science projects and as a systematic and coordinated infrastructure for assessing the legitimacy and ‘safe’ credentials of a project. The value of these LPS data assets will be significantly enhanced through linkage to Covid-19 relevant health and other routine records (e.g., employment status, occupation, and benefit receipt records). This will allow the use of objectively assessed diagnoses and other records and the collection of data not amenable to self-report (such as detailed records of care, or health service utilisation). It will also allow researchers to assess and quantify/document bias and error through triangulating data collected across different organisations and settings.

Throughout the development of the UK LLC, UoB have worked collaboratively with UK LPS to co-design and develop the protocol and methodology. One of the strengths of the collaborative approach championed by the UK LLC is that UK LPS are committed and enthusiastic partners in this work: this initiative comes from within the community. The UK LLC will provide infrastructure to support the LPS and researchers working within the UK to investigate high priority COVID-19 research questions. UK LLC/UoB will continue to seek input from the LPS into decision making, guidance in future development and operation. Crucially, the LPS’s all maintain and manage ‘their’ relationships with ‘their’ participants including making critical decisions about data use, and communicating with participants about the existence of, and work of, the UK LLC and providing reassurance that this does not alter the participants existing relationship.

Data for the following studies will flow under COPI:

1. The National Study of Health and Development (NSHD)

2. The Southall And Brent REvisited (SABRE)

3. Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”)

4. Genetic Links to Anxiety and Depression Study (GLAD)

5. English Longitudinal Study of Ageing (ELSA)

6. 1958 National Child Development Study (NCDS)

7. 1970 British Cohort Study (BCS 70);

8. Next Steps

9. The Millenium Cohort Study (MCS)

10. Track-COVID - subset of INTERVAL COMPARE and STRIDES bioresource

11. NIHR (National Institute for Health Research) BioResouce

12. Extended Cohort for E-health, environment, and DNA (EXCEED)

13. Understanding Society

14. Born in Bradford

15. European Prospective Investigation into Cancer (EPIC)

16. The Fenland Study

17. Twins early Development Study (TEDS)

18. INTERVAL

19. COMPARE

20. STRIDES

21. TwinsUK

There is a longer-term objective to include a larger number of studies into the UK LLC to boost statistical power, increase participant heterogeneity and sample size in population and outcome sub-groups and to increase the range of self-reported information. Any additional study would be included as an amendment to this application.

The UoB scientific programme requires the extraction and use of the NHS England data compiled with Covid-19 relevant datasets (primary care, secondary care, community, mental health care provision, Covid-19 testing and outcomes data, NHS service use interactions such as NHS 111 records, mortality, disease registry and demographics data) for the purpose of establishing the UK LLC.

The UK LLC has conducted an assessment as to whether the data held and processed within the UK LLC is identifiable/potentially identifiable and therefore Personal Data which is subject to the Data Protection Act 2018 (DPA), UK General Data Protection Regulations (UK GDPR) and Common Law Duty of Confidentiality (Common Law) in accordance with guidance issued by the Information Commissioner’s Office. The assessment determined that the infrastructure, incorporating all data flows, should be considered as containing Personal Data and being subject to the DPA/UK GDPR and Common Law requirements. However, within this the assessment identified that the UK LLC sits at the threshold of being considered as holding pseudonymous data (Personal Data) and de-identified data where the risk of re-identification is not reasonably likely (not Personal Data), and that this differing status manifests when considering the data from the perspective of the different high-level groups of UK LLC data users and their distinct processing activities. It is evident that for the contributing studies, the use of data is either identifiable (for the provision of identifiers for linkage purposes) or pseudonymous (for the provision of data into the UK LLC and the processing of study data by study staff in their demarcated study processing area) and is therefore Personal Data. Yet, for the UK LLC staff and researcher users, the risk of re-identification of data within the UK LLC is not reasonably likely. This status, for these users, has been achieved through the deployment of data level as well as contextual controls which mitigate the risk of identification. The principal amongst these is the UK LLC operates as a ‘Trusted Research Environment’ which ensures that minimised data are used by minimised number of legitimate users within a secure environment for approved and proportionate purposes.

To contribute data into the UK LLC, LPS’s will be required to flow identifiers to the NHS to establish/filter a linkage between each participants' study record and their NHS registry record(s). Studies will update their privacy notices and notify participants of their involvement with UK LLC, explaining what this means and provide options to ‘opt-out’, in line with existing policy and practice.

UK LLC TRUSTED RESEARCH ENVIRONMENT (TRE)

The LLC will be built on a UK Secure eResearch Platform (UKSeRP). This is a well-established infrastructure for data science: it has ISO27001 and NHS Data Security Protection toolkit and Office for National Statistics accreditation under the Digital Economy Act 2017. Other instances of this environment are already used to host LPS and NHS England data and have been audited and commended as being well suited to the purpose and low risk.

The UK LLC TRE will be owned, operated, and managed by the LLC team (data management based at UoB). Under instruction from UoB, contributing LPS and data owners including NHS England will provide data to the UK LLC TRE using the ‘split file’ approach which is used across UKSeRP. This split file approach (details below) ensures the physical and processing operational separation of personal identifiers and participants de-identified study data and health records. This separation means that no party involved in this process can see both identifiers and data. This split file approach is coupled with the technological and socio-governance controls applied at the UK LLC (e.g., training, contracts, policy and procedures, penalties for misuse). The UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. For UK LLC staff and research users the risk of re-identification is controlled to the point that it is not reasonably likely.

UK LLC C-19 RESEARCH PROGRAMME

The UK LLC is specifically commissioned as a database resource to inform the LH&W NCS but is also designed to be a resource for the investigation of other Covid-19 questions as these are identified and prioritised through the Scientific Advisory Group for Emergencies (see the Health Data Research UK and SAGE research prioritisation funnel - current versions available through the HDRUK SAGE report, see - https://www.hdruk.ac.uk/covid-19/).

Given these questions are likely to change over time (e.g., the identification of ‘Long COVID’ as a potential new set of complex outcomes is introducing important new research questions), the research programme the LLC is designed to inform can be defined as the investigation of Covid-19 questions using data from longitudinal studies linked to routine records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing. For the use of linked longitudinal studies and NHS records this definition is refined to make clear the investigations primary outcomes must be focused on generating benefits to the health and social care system.

The investigation of Covid-19 questions using data from longitudinal studies linked to routine NHS records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing: where the primary research outcome is designed to benefit the health and social care system. These linked data will be of unique value in underpinning a programme of research on Covid-19 informed by the data assets within the UK LLC.

This program will seek to understand the patterns and predictors of infection, (including re-infection) and disease outcomes (such as ‘long covid’), and the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. It will also explore population level changes to physical and mental health, including hospital admission and mortality, in association with viral suppression measures and how these relate to changes in health behaviours.

Given the impact of the Covid-19 containment measures, it is important to investigate the role of socio-economic and neighbourhood/environmental factors in determining population level impacts to physical and mental health, to identify both groups at risk, and factors which offer resilience to adverse outcomes. Within this, UK LLC users will analyse changes in health care service use (using self-reported data and NHS health records) to determine if patterns in these have changed during the Covid-19 pandemic. Linked NHS records will also inform consideration of service use in relation to pre-pandemic health status and regular service interactions (e.g., screening, health reviews and routine service take up (such as annual seasonal flu vaccinations)).

This programme, through the LH&W NCS will specifically seek to answer the following questions:

i) What are the mental health consequences of Covid-19 infection, and of viral suppression measures? How do these differ by key demographics, socioeconomic status and by prior mental health status?

ii) What are the risks, determinants, and expressions of the physical health consequences of long-Covid?

iii) What are the disruptions to health care services (primary and secondary), and what are the health consequences of these disruptions? How do they vary by geography, demographics, and socioeconomic status?

iv) What are the risks of re-infection?

v) How do measures of historical health & physiology impact risk of C-19 infection and its consequences?

vi) How do we best identify cases in population studies, triangulating symptoms, antigen, and antibody testing? How do antibody profiles differ by case status (e.g., those with and without symptoms), how do antibody profiles vary over time? What are the health determinants of such variability?

vii) What are the long-term socioeconomic consequences of viral suppression measures? How do these in turn impact on health and on health care utilisation?

These research themes have been scientifically reviewed by the UK Chief Scientific Advisor and the UK Chief Medical Officer and are high priority questions of critical national interest.

OTHER DATA PROCESSING PURPOSES

In addition to the primary research themes, UoB will:

1) use linked NHS data to help inform descriptive and documentary analysis of the UK LLC dataset. This will describe the combined LPS population and make clear the LLC denominator, its characteristics and how this sub-set of the population relates to the wider population (through comparison with aggregated populations statistics). This will inform research users and those (e.g., policy makers) seeking to draw inferences from findings.

2) conduct feasibility assessments to help determine the viability of proposed projects. To do this, UK LLC staff will query the UK LLC integrated dataset (LPS data and linked NHS Records) to determine if UK LLC have sufficient case numbers to be able to inform specific research proposals. This purpose enhances the ethical and governance case for the resource as it means data are not shared where they cannot meaningfully inform the science, it also helps ensure the efficient use of research funding and resources.

DATA MINIMISATION

The datasets and data items requested have been restricted to only those directly relevant to the Covid-19 research programme that this infrastructure is designed to support. UoB fully adopt the standard NHS England definition of the Covid-19 relevant dataset.

As described, the UK LLC C-19 research programme is wide ranging and seeks to understand patterns and predictors of infection, disease outcomes, the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. Datasets requested include measure of Covid-19 related outcomes and pre-Covid-19 baseline datasets (allowing the value of these longitudinal data to be maximised).

The linked data requested are minimised to include only data covering the period each contributing study was in follow-up. The data requested are further minimised to include only records of participants of studies contributing to the UK LLC and those who have objected to this use of their data are excluded. Information on participants who have ‘opted-out’ will be uploaded to the UK LLC quarterly so that those participants’ data will be excluded from future studies.

UoB are unable to minimise datasets further based on fields or episodes at this time as work continues to be conducted to inform the UK’s Covid-19 research programme which aims to be responsive to health and government policy makers. For example in HES, UoB cannot consider only episodes with a specific diagnosis or procedure at this time, UoB do not yet understand the longer-term health implications from Covid-19 infection or whether there exist patterns and predictors of Covid-19 infection based on previous episodes in HES.

The study have minimised datasets so they only include records of participants of longitudinal population studies contributing to the UK LLC and to NHS England data that are directly relevant to the research programme. Whilst UoB cannot provide justification through reference to specific analysis that will be conducted using these data, given their purpose as a resource developed to enable researchers to respond to the emerging priorities as determined by the UK Scientific Advisory Group for Emergencies.

1) The UK LLC is a resource to enable the longitudinal perspective. The value of longitudinal research lies in the collection of a broad range of data, repeated at regular intervals allowing the investigation of interactions between things occurring over the lifetime of individuals. To enable this perspective to be applied to the Covid-19 research questions UoB require data from NHS England and other providers that covers the duration of the longitudinal population studies contributing their study data. For example, the National Study of Health and Development is still collecting data from babies born in one week of 1946. Emerging Covid-19 research for example around disease severity could feasibility be informed by patterns in childhood ill health and historic episodes of hospitalisation within these cohorts.

2) As above with the duration of NHS-linked data requested, the breadth of data within the LLC enable rapid response to emerging research priorities. UoB cannot predict what these priorities will be but can provide indicative examples:

The usefulness of a QOF exception code in GDPPR indicating that a patient has refused an annual health check are that this could be indicative of their behaviour around choosing to have vaccination; insights into vaccination and health behaviour have the potential to be of great importance. The risk is there are unforeseen needs which are then hard to fulfil if the data has been minimised.

The UK LLC is designed to support longitudinal research studies and research analysts working within the UK. Longitudinal research is geared to understand the longer-term outcomes of the pandemic and emerging interactions between health and social factors and including considerations of outcomes such as Long COVID whose duration is yet to be determined, and the impact of disruption of routine health service provision on wider mental and physical health.

LEGAL BASIS, ETHICS AND RISK OF POTENTIAL HARM

UoB will be sole Data Controller under this agreement.

UoB is establishing the UK LLC and using linked NHS records to inform the LH&W NCS research programme as a task carried out in the public interest (Article 6(1)(e) given the UK LLC is designed to answer specific high-priority research questions at a time of national crisis. Given that the requested health records include flows of sensitive information, the UK LLC also draws on Article 9(2)(j) in accordance with Article 89(1) where UoB's purpose is to conduct statistical scientific research. The case for this research being in the public interest is established through balancing the strong policy driver of improved understanding of the pandemic with consideration of potential harms to the patients/participants whose records are involved. Potential harms primarily relate to breach of confidentiality and the subsequent misuse of Personal Data or erosion of trust in longitudinal research/data science.

The lawful basis for the release and use of the confidential data being shared under this version of the agreement is Regulation 3(4) of the National Health Service (Control of Patient Information Regulations) 2002 (COPI) to require NHS England to share confidential patient information with organisations entitled to process this under COPI for COVID-19 purposes. The only permitted activities under this Data Sharing Agreement (DSA) are for COVID-19 purposes and within bounds of Reg 3(2) COPI. Reg 3 (2) COPI states that: "2) For the purposes of this regulation, “processing” includes any operations, or set of operations set out in regulation 2(2) which are undertaken for the purposes set out in paragraph (1)." The research relates to the monitoring and managing of COVID-19 and would therefore be covered by Reg 3(1)(d) of COPI.

The risk of this harm is minimised as this is:

1. an observational population-level research which will not result in a direct intervention to any participant;

2. the UK LLC uses tried and tested IT infrastructure and governance frameworks specifically designed to minimise risks to privacy during health data science;

3. the UK LLC is a fully de-identified research environment with sufficient controls that the risk of disclosure is not considered reasonably likely;

4. a reasonable expectation for the use of NHS data in this way has been established by the contributing studies, with studies making best endeavours to inform participants ahead of data extraction taking place, as well as continued fair processing (in relation to UK LLC) in line with usual policies and practices. Each stage of this fair processing will provide participants with the opportunity to object;

5. National Opt-Out will also be respected unless overridden by specific consent for the use of linked health records in the studies research programme;

6. all staff and users are vetted and approved professional researchers operating within controlled and auditable conditions.

This research is ethical given the strong scientific rationale, the process is transparent, best endeavours will be made to inform participants of the use of their data within UK LLC, with a right to object, that risks are mitigated and the participants (and wider public) are likely to directly benefit from the research through improved health care and government policy provision in response to the unprecedented Covid-19 crisis. Participants will be informed about how their data is used through ongoing fair processing communications including details of the process for ‘opting-out’.

OPERATIONAL MANAGEMENT

UoB is the study sponsor and Data Controller. UK LLC staff at UoB and UoE have responsibility for day to day running of the UK LLC databank: only UoB staff will have access to the data and TRE. The UK LLC Director, principle and co-investigators and data managers are employed by UoB. The UK LLC Co-Director and UK LLC administrative staff (responsible for administrative operations, communications, and participant/public engagement) are based at UoE). The Longitudinal Health & Wellbeing National Core Study is led from University College London (UCL) and the UK LLC reports into UCL on delivery of UK LLC objectives.

The UK LLC databank is configured to have two distinct classes of operational areas, which have three functions:

1) Data management and processing: All operational control and access for data processing and management of de-identified, including but not limited to data processing and storage, descriptive analysis and report/publication writing, subset, and provision of data to approved users is restricted to UoB staff. UoB staff are the only individuals who can access all data across all studies and sources in their raw and processed forms.

2a) Onward sharing to contributing studies: A secure operating partition will be created for each contributing study. Within these linked LPS - NHS data, restricted to the participants of that study, can be access by approved study staff through the ‘onward sharing’ mechanism. The study staff will use this area to conduct descriptive analysis on their participants (e.g., to provide documentation to describe the data asset, to inform bias assessments) and to assess feasibility of projects (e.g., to determine whether there are sufficient cases to scientifically justify inclusion in a proposed project) to inform potential users and to conduct study-level due diligence.

2b) Onward sharing to research analysts: Secure operating partitions will be created for each UK LLC approved project. The UK LLC approved users for these projects will have access to their folder which will contain a sub-set of data relevant to that study where onward sharing conditions of data providers are met. The UK LLC research users will control of data processing and analysis and leading publications.

The UoB UK LLC staff and UK LLC infrastructure system administrators will have access across all three operational areas.

The University of Swansea (UoS) is a Data Processor who provide data infrastructure and data management services to UoB and will maintain the integrity of the TRE, will assist with the data management processing (Under UoB’s direction) and will conduct the output disclosure assessments (to UoB’s specification). The UoS have sub-contracted the Data Linkage functions of the UKSeRP (including all management of participant identifiers) to the Digital Health and Care Wales (DHCW) (part of Velindre University NHS Trust) who as such, are Data Processors for UoB.

The UK LLC is funded by HM Treasury as part of the National Core Studies for Covid-19 research and is administered by the UK Medical Research Council (MRC). HM Treasury and MRC will not have access to NHS England data.

SUB-LICENSING

For the UK LLC to be a useful and sustainable resource for researchers to interrogate, and obtain meaningful data in a timely manner, contractual arrangements will need to facilitate high-volume and rapid turn-around of data requests. This will only be possible through removing the burden of the need for a high number of additional data sharing agreements across multiple institutions which would impede the speed of research. Therefore, a contract structure is needed to control the data flows, to control the purposes and way these data are processed, and to assign roles and responsibilities and to enforce the governance requirements of the individual studies and data owners and the legal basis under which they are permitted to operate.

These agreements are as follows:

- Data Deposit Agreement(s) established between UoB and the studies institution(s), ensuring fair processing will be carried out, and allowing study collected data to be made available for Covid-19 research to approved users within the UK LLC TRE.

- Shared data will be deposited into TRE where it will be processed, curated, and analysed.

- TRE established through research collaboration with University of Swansea (UoS) controlled through a Data Deposit

Agreement between UoB and UoS comprising:

1) a data deposit agreement for Welsh records and other records (SymptomTracker) to be deposited in the UK LLC and used for research

2) a collaborative contract where UoB procure a UKSeRP (UoS as Data Processor) and UoS data science expertise (an academic collaboration which involves Data Processing). Some aspects of the UoS contribution will be fulfilled by sub-contractor Data Health and Care Wales (DHCW) acting as a Trusted Third Party controlled through the Sub-Contract Agreement between UoS and DHCW.

The contract structure also permits the processes by which the participants of these studies can be identified by the NHS and their records selected and provided into the UK LLC for integration with the study data and onward sharing with researchers. The agreements for selecting and depositing these data has two distinct parts:

1) Record Linkage

• UoB will instruct the contributing studies to send a file of participant identifiers to NHS Wales Informatic Service (controlled through the Data Deposit Agreement between UoB and each contributing studies institution(s)).

• UoB will instruct the DHCW (a process controlled through the TRE procurement contract between UoB and UoS, and as described in the Data Deposit Agreement between UoB and contributing studies) to send participant identifiers to NHS England for record linkage purposes

• NHS England will use these identifiers to link these to their patient register and select the longitudinal study participants data (controlled through the Data Sharing Agreement between UoB and NHS England).

2) Data Deposition & Processing

• NHS England will then de-identify and deposit the selected data into the UK LCC where UoB will become the data controller of this patient data, for the purpose of processing it, approving onward research use and providing managed access to relevant sub-sets of data for Covid-19 research (controlled through the Data Sharing Agreement between UoB and NHS England). The structure will then enable the UK LLC to determine the onward processing of the integrated data for Covid-19 research purposes by reviewing applications from researchers and onwardly sharing sub-sets of relevant data within the TRE to approved users (controlled through the Data Access Agreement between UoB and approved researchers’ institutions). The nature of the sublicensing arrangements above are in line with the NHS England standards - https://digital.nhs.uk/services/data-access-request-service-dars/dars-guidance/sub-licencing-and-onward-sharing-of-data. NHS England requirement to audit data use is a requirement in the onward sharing contract and remains feasible in the UK LLC framework.

The UK LLC is likely to generate diverse outputs, with these forming a mix of rapid insight reports to policy makers and substantive research reports. The NCS is specifically designed to establish a clear channel between decision makers and researchers and the UK LLC is designed to enable rapid response and up-to-date information needed to support this. The likely outputs include:

* Peer-reviewed scientific publication (and pre-print publications of initial drafts prior to review which are designed enable rapid awareness of the findings), outputs will be ongoing, preprints and publications to date:

Preprints | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

Publications | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London

* Conference presentations and workshop presentations (to academic, health service practitioners and planners, government policy makers, Scientific Advisory Group for Emergencies (SAGE) members). These will be ongoing as part of our communications and engagement strategy, events to date include: Specifically for UK Longitudinal Linkage Collaboration:

Presentation: Welcome Longitudinal Population Studies conference (2021)

Regular presentations to Health Data Research UK - Data & Connectivity National Core Studies Delivery Group (2021-2022)

Presentation: CLOSER Longitudinal Population Studies conference (2022) ‘UK LLC Overview and explaining the innovations made and challenges faced’

Presentation: LHW National Core Study - Joint EHR cohort leads meeting ‘how to access UK LLC’

Workshop 1: to National Core Studies Analysts ‘Overview of UK LLC and how to access UK LLC’

Workshop 2: to CLOSER, National Core Studies and Vanguard Members ‘overview of UK LLC and how to access data

Presentation to UK LLC Communications Network (2022)

Youtube: UK Longitudinal Linkage Collaboration Verbal YouTube Public Involvement Induction (2022)

Presentation: to Elizabeth Blackwell Institute Health Data Science Showcase (2021)

Presentation: ESRC Strategic Advisory Group ‘ Understanding Data for Population research’ (2021)

Presentation: to Longitudinal Area Network and Data-link (LAND) community of practice (2022)

Poster Presentation: to Welcome Trust Longitudinal Studies Conference ‘overview of UK LLC’ (2021)

Presentation and advice: to Welcome Longitudinal Population Study COVID-19 Steering Group ‘provided specialist advice on linkage opportunities’ (2021)

Presentation and advice: to Population Research UK Programme Group ‘provided specialist advice’ (2021)

Advice: to Department of Education Scientific Advisory Group ‘provided specialist advice on record linkages in longitudinal studies (2021)

* Reports to health service practitioners and planners, government policy makers and Committee’s (e.g., National Institute for Health and Care Excellence (NICE), SAGE), it is anticipated that these will be ongoing, briefings to date include:

Briefing Reports | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

* Rapid evidence synthesis reports as requested by health and government policy makers (through the entirety of the 3-year NCS programme);

* Lay summaries provided through websites, Blogs, social media posts, privacy notices and outputs designed to promote the transparent use of data and the wider research process, regularly updated on our data-use register as follows:

Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk)

All outputs from the TRE will be assessed for disclosure risk and will be anonymous population level findings. The framework for this is based on leading 'five safes’ approach and is closely adapted from equivalent processes accredited to standards of the Office for National Statistics and the UK Statistics Authority.

All ‘meta’ products of the UK LLC research (such as code list definitions, syntax, workings for derived variables) will be made available for reuse and interrogation/replication. Access to the UK LLC is ‘open’ in the sense that any legitimate researcher conducting Covid-19 research can apply for access using a transparent and consistent decision-making process. As such, the UK LLC enhanced resource should be considered as an important ‘output’ and a measurable benefit of the UK LLC/LH&W NCS.

The high profile of Covid-19 means that many findings may have media interest. Importantly, the UK LLC is establishing a Patient/Participant Involvement and Engagement strategy within its communication work-package package with a dedicated budget and experienced Communications PPIE Officer. The package will build on existing and established relationships with participants to involve public/participants in the co-design of mechanisms to engage and effectively communicate UK LLC findings more widely with the public. UoB are committed to public involvement in our design and communications. UoB have already included HDRUK PPIE panel members (public representatives) in a consultation exercise and will draw on the extensive network of participants to further explore and co-develop dissemination mechanisms, in developing the end-to-end animation in the first instance. UoB are also developing with the studies opportunities for a subgroup of participants to be involved, in the data access decision making process for example. UoB have ensured that the programme of work is adequately resourced to compensate the public for their time, this has been an integral part of the project planning.

The UK LLC is a novel research infrastructure. The creation of this for Covid-19 research may provide methodological insights and illustrations of a new way of working for linkage in longitudinal research. Insights from this will be provided to the longitudinal community and its funders/key stakeholders (through academic papers, presentations, and contributions to advisory groups).

There are ten projects approved with data provisioned within the TRE to 20 researchers across seven institutions with six more projects in the pipeline. Key questions being addressed are: comparing the burden of long COVID in the community; mental health and wellbeing consequences of contracting COVID19;impacts of the COVID-19 pandemic on access to elective care in the UK. A full list of approved projects can be found here: Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk).

Findings from the LH&W NCS have led to Cabinet office and NHS England 'teach in' sessions, a national GP alert, an NHS enhanced service specification , rapid reports for SAGE, and they have provided key evidence to guide the NICE long COVID guidelines.

• During lockdown the LHW Team identified instances of inappropriate switching of blood thinners, triggering a national alert to GPs

• Policy makers have been provided with information on long COVID burden of disease, risk factors, definition and long term outcomes via cabinet briefing reports and SAGE reports

• Low levels of long COVID GP coding were reported. This led to a NHS enhanced service specification, directed at GPs to drive an increase in long COVID coding

• NICE used long COVID findings in its evidence to review risk factors

• Briefing notes submitted addressing inequalities in healthcare disruption, mental health decline during lockdown and the impact of furlough on health behaviour an mental health reasons for a low antibody response.

• LHW work was amongst the first to provide Chief Medical Officer with a quantification of the first vaccine side effects and identify differences in vaccine update among ethnic groups, and point out that people with learning disabilities should be prioritised.

The team have also presented early results from the first wave of the convalescence study qualitative interviews to NHS England, highlighting that long COVID participants are failing at the first hurdle of accessing healthcare. The team also communicated that there is an increased risk of developing cardiovascular complications up to a year after covid-19 infection, even in the absence of persistent symptom reporting.

Throughout the project we have worked with the public via our advisory group, to shape and communicate our programme. Public contributors have helped to create material aimed at communicating the results to all stakeholders, including infographics, short videos and an animation. We have also created a long COVID forum to host a diverse and inclusive public/patient conversation on defining and researching long COVID.

V3 -

UoB will continue to have legal basis under Regulation 3 of the COPI Regulations for processing CPI for Covid-19 Purpose. Where University of Bristol (UoB) will use the data for previously approved research related to a Covid-19 Purpose, and this research is still ongoing, UoB are able to continue to meet the requirements in Regs 3(1), 3(3) and 7 of the COPI Regulations. DARS have assessed that UoB do still meet the 3 limbs of the COPI regulation.

The UK Longitudinal Linkage Collaboration (UK LLC) project itself is cross-institutional and has staff based at University of Bristol and the University of Edinburgh. The University of Edinburgh are collaborators however they have no role in determining how the data will be processed, nor will they have any access to the record level data. For clarification regarding the distinct staff groups noted in the agreement:

Study staff = staff from the contributing cohorts;

University of Bristol (UoB) staff = staff of UK Longitudinal Linkage Collaboration (LLC) employed by UoB;

UK LLC staff = staff employed by UK LLC (UoB and University of Edinburgh (UoE) staff).

BACKGROUND

The UK Chief Scientific Advisor has established (October 2020) a programme of National Core Studies (NCS) for SARS-CoV-2 (Covid-19) research as a coordinated, long-term, national research initiative. This will consider Covid-19 in terms of a viral pandemic (including issues of cases, transmission, symptoms, and outcomes) and in terms of the health and social impacts of behavioural restrictions designed to mitigate the harms of the pandemic. The NCS has six different sub-programmes which are addressing major Covid-19 research areas; one of these is the Longitudinal Health and Well-being (LH&W) NCS which is designed to use data from longitudinal studies to address the impact of Covid-19 and of associated viral suppression measures on health and well-being. The LLC is the central hub component of the Longitudinal Health and Well-being NCS.

The UK LLC has been designed to underpin the LH&W NCS, although not exclusively as the NCS are designed to support each other, where for example: the LLC could form the infrastructure for the long-term patient follow-up of consenting trial participants (e.g. participants in the RECOVERY trial) or, where UK longitudinal population studies (LPS) are being used to collect specific new study data which is not available through routine records (e.g. the ALSPAC birth cohort is providing a frame to collect biological samples to inform the Immunology NCS research programme; the assayed results of which will be uploaded into the LLC). For this reason, users of the LLC may come from across the full range of NCS studies and the resource will be accessible to other legitimate UK-based researchers investigating Covid-19 through a sub-licence framework. The NCS is planned to be a two/three-year research programme commencing October 2020.

THE UK LLC & DATA SUBJECTS

UoB (for LLC purposes) will underpin the NCS programme by combining, within a Trusted Research Environment, study data from >15 major inter-disciplinary UK LPS with a combined total of 1 – 2 million UK participants, centrally link to a wide range of Covid-19 relevant and non-health administrative records to utilise alongside the UK Biobank, the Zoe Symptom Tracker cohort and other sources. The integrated data, infrastructure and accompanying governance aspects will collectively be known as the UK LLC.

The de-identified data from contributing studies will be used to establish a UK LLC participant and household register, this will be used within the UK LLC to index and catalogue available data. Researchers will be able request access to the integrated data via the Health Data Research UK Gateway (https://www.healthdatagateway.org/) which through the NCS programme is being established as the de facto application point for UK Data Science projects and as a systematic and coordinated infrastructure for assessing the legitimacy and ‘safe’ credentials of a project. The value of these LPS data assets will be significantly enhanced through linkage to Covid-19 relevant health and other routine records (e.g., employment status, occupation, and benefit receipt records). This will allow the use of objectively assessed diagnoses and other records and the collection of data not amenable to self-report (such as detailed records of care, or health service utilisation). It will also allow researchers to assess and quantify/document bias and error through triangulating data collected across different organisations and settings.

Throughout the development of the UK LLC, UoB have worked collaboratively with UK LPS to co-design and develop the protocol and methodology. One of the strengths of the collaborative approach championed by the UK LLC is that UK LPS are committed and enthusiastic partners in this work: this initiative comes from within the community. The UK LLC will provide infrastructure to support the LPS and researchers working within the UK to investigate high priority COVID-19 research questions. UK LLC/UoB will continue to seek input from the LPS into decision making, guidance in future development and operation. Crucially, the LPS’s all maintain and manage ‘their’ relationships with ‘their’ participants including making critical decisions about data use, and communicating with participants about the existence of, and work of, the UK LLC and providing reassurance that this does not alter the participants existing relationship.

Data for the following studies will flow under COPI:

1. The National Study of Health and Development (NSHD)

2. The Southall And Brent REvisited (SABRE)

3. Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”)

4. Genetic Links to Anxiety and Depression Study (GLAD)

5. English Longitudinal Study of Ageing (ELSA)

6. 1958 National Child Development Study (NCDS)

7. 1970 British Cohort Study (BCS 70);

8. Next Steps

9. The Millenium Cohort Study (MCS)

10. Track-COVID - subset of INTERVAL COMPARE and STRIDES bioresource

11. NIHR (National Institute for Health Research) BioResouce

12. Extended Cohort for E-health, environment, and DNA (EXCEED)

13. Understanding Society

14. Born in Bradford

15. European Prospective Investigation into Cancer (EPIC)

16. The Fenland Study

17. Twins early Development Study (TEDS)

18. INTERVAL

19. COMPARE

20. STRIDES

21. TwinsUK

There is a longer-term objective to include a larger number of studies into the UK LLC to boost statistical power, increase participant heterogeneity and sample size in population and outcome sub-groups and to increase the range of self-reported information. Any additional study would be included as an amendment to this application.

The UoB scientific programme requires the extraction and use of the NHS Digital data compiled with Covid-19 relevant datasets (primary care, secondary care, community, mental health care provision, Covid-19 testing and outcomes data, NHS service use interactions such as NHS 111 records, mortality, disease registry and demographics data) for the purpose of establishing the UK LLC.

The UK LLC has conducted an assessment as to whether the data held and processed within the UK LLC is identifiable/potentially identifiable and therefore Personal Data which is subject to the Data Protection Act 2018 (DPA), UK General Data Protection Regulations (UK GDPR) and Common Law Duty of Confidentiality (Common Law) in accordance with guidance issued by the Information Commissioner’s Office. The assessment determined that the infrastructure, incorporating all data flows, should be considered as containing Personal Data and being subject to the DPA/UK GDPR and Common Law requirements. However, within this the assessment identified that the UK LLC sits at the threshold of being considered as holding pseudonymous data (Personal Data) and de-identified data where the risk of re-identification is not reasonably likely (not Personal Data), and that this differing status manifests when considering the data from the perspective of the different high-level groups of UK LLC data users and their distinct processing activities. It is evident that for the contributing studies, the use of data is either identifiable (for the provision of identifiers for linkage purposes) or pseudonymous (for the provision of data into the UK LLC and the processing of study data by study staff in their demarcated study processing area) and is therefore Personal Data. Yet, for the UK LLC staff and researcher users, the risk of re-identification of data within the UK LLC is not reasonably likely. This status, for these users, has been achieved through the deployment of data level as well as contextual controls which mitigate the risk of identification. The principal amongst these is the UK LLC operates as a ‘Trusted Research Environment’ which ensures that minimised data are used by minimised number of legitimate users within a secure environment for approved and proportionate purposes.

To contribute data into the UK LLC, LPS’s will be required to flow identifiers to the NHS to establish/filter a linkage between each participants' study record and their NHS registry record(s). Studies will update their privacy notices and notify participants of their involvement with UK LLC, explaining what this means and provide options to ‘opt-out’, in line with existing policy and practice.

UK LLC TRUSTED RESEARCH ENVIRONMENT (TRE)

The LLC will be built on a UK Secure eResearch Platform (UKSeRP). This is a well-established infrastructure for data science: it has ISO27001 and NHS Data Security Protection toolkit and Office for National Statistics accreditation under the Digital Economy Act 2017. Other instances of this environment are already used to host LPS and NHS Digital data and have been audited and commended as being well suited to the purpose and low risk.

The UK LLC TRE will be owned, operated, and managed by the LLC team (data management based at UoB). Under instruction from UoB, contributing LPS and data owners including NHS Digital will provide data to the UK LLC TRE using the ‘split file’ approach which is used across UKSeRP. This split file approach (details below) ensures the physical and processing operational separation of personal identifiers and participants de-identified study data and health records. This separation means that no party involved in this process can see both identifiers and data. This split file approach is coupled with the technological and socio-governance controls applied at the UK LLC (e.g., training, contracts, policy and procedures, penalties for misuse). The UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. For UK LLC staff and research users the risk of re-identification is controlled to the point that it is not reasonably likely.

UK LLC C-19 RESEARCH PROGRAMME

The UK LLC is specifically commissioned as a database resource to inform the LH&W NCS but is also designed to be a resource for the investigation of other Covid-19 questions as these are identified and prioritised through the Scientific Advisory Group for Emergencies (see the Health Data Research UK and SAGE research prioritisation funnel - current versions available through the HDRUK SAGE report, see - https://www.hdruk.ac.uk/covid-19/).

Given these questions are likely to change over time (e.g., the identification of ‘Long COVID’ as a potential new set of complex outcomes is introducing important new research questions), the research programme the LLC is designed to inform can be defined as the investigation of Covid-19 questions using data from longitudinal studies linked to routine records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing. For the use of linked longitudinal studies and NHS records this definition is refined to make clear the investigations primary outcomes must be focused on generating benefits to the health and social care system.

The investigation of Covid-19 questions using data from longitudinal studies linked to routine NHS records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing: where the primary research outcome is designed to benefit the health and social care system. These linked data will be of unique value in underpinning a programme of research on Covid-19 informed by the data assets within the UK LLC.

This program will seek to understand the patterns and predictors of infection, (including re-infection) and disease outcomes (such as ‘long covid’), and the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. It will also explore population level changes to physical and mental health, including hospital admission and mortality, in association with viral suppression measures and how these relate to changes in health behaviours.

Given the impact of the Covid-19 containment measures, it is important to investigate the role of socio-economic and neighbourhood/environmental factors in determining population level impacts to physical and mental health, to identify both groups at risk, and factors which offer resilience to adverse outcomes. Within this, UK LLC users will analyse changes in health care service use (using self-reported data and NHS health records) to determine if patterns in these have changed during the Covid-19 pandemic. Linked NHS records will also inform consideration of service use in relation to pre-pandemic health status and regular service interactions (e.g., screening, health reviews and routine service take up (such as annual seasonal flu vaccinations)).

This programme, through the LH&W NCS will specifically seek to answer the following questions:

i) What are the mental health consequences of Covid-19 infection, and of viral suppression measures? How do these differ by key demographics, socioeconomic status and by prior mental health status?

ii) What are the risks, determinants, and expressions of the physical health consequences of long-Covid?

iii) What are the disruptions to health care services (primary and secondary), and what are the health consequences of these disruptions? How do they vary by geography, demographics, and socioeconomic status?

iv) What are the risks of re-infection?

v) How do measures of historical health & physiology impact risk of C-19 infection and its consequences?

vi) How do we best identify cases in population studies, triangulating symptoms, antigen, and antibody testing? How do antibody profiles differ by case status (e.g., those with and without symptoms), how do antibody profiles vary over time? What are the health determinants of such variability?

vii) What are the long-term socioeconomic consequences of viral suppression measures? How do these in turn impact on health and on health care utilisation?

These research themes have been scientifically reviewed by the UK Chief Scientific Advisor and the UK Chief Medical Officer and are high priority questions of critical national interest.

OTHER DATA PROCESSING PURPOSES

In addition to the primary research themes, UoB will:

1) use linked NHS data to help inform descriptive and documentary analysis of the UK LLC dataset. This will describe the combined LPS population and make clear the LLC denominator, its characteristics and how this sub-set of the population relates to the wider population (through comparison with aggregated populations statistics). This will inform research users and those (e.g., policy makers) seeking to draw inferences from findings.

2) conduct feasibility assessments to help determine the viability of proposed projects. To do this, UK LLC staff will query the UK LLC integrated dataset (LPS data and linked NHS Records) to determine if UK LLC have sufficient case numbers to be able to inform specific research proposals. This purpose enhances the ethical and governance case for the resource as it means data are not shared where they cannot meaningfully inform the science, it also helps ensure the efficient use of research funding and resources.

DATA MINIMISATION

The datasets and data items requested have been restricted to only those directly relevant to the Covid-19 research programme that this infrastructure is designed to support. UoB fully adopt the standard NHS Digital definition of the Covid-19 relevant dataset.

As described, the UK LLC C-19 research programme is wide ranging and seeks to understand patterns and predictors of infection, disease outcomes, the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. Datasets requested include measure of Covid-19 related outcomes and pre-Covid-19 baseline datasets (allowing the value of these longitudinal data to be maximised).

The linked data requested are minimised to include only data covering the period each contributing study was in follow-up. The data requested are further minimised to include only records of participants of studies contributing to the UK LLC and those who have objected to this use of their data are excluded. Information on participants who have ‘opted-out’ will be uploaded to the UK LLC quarterly so that those participants’ data will be excluded from future studies.

UoB are unable to minimise datasets further based on fields or episodes at this time as work continues to be conducted to inform the UK’s Covid-19 research programme which aims to be responsive to health and government policy makers. For example in HES, UoB cannot consider only episodes with a specific diagnosis or procedure at this time, UoB do not yet understand the longer-term health implications from Covid-19 infection or whether there exist patterns and predictors of Covid-19 infection based on previous episodes in HES.

The study have minimised datasets so they only include records of participants of longitudinal population studies contributing to the UK LLC and to NHS Digital data that are directly relevant to the research programme. Whilst UoB cannot provide justification through reference to specific analysis that will be conducted using these data, given their purpose as a resource developed to enable researchers to respond to the emerging priorities as determined by the UK Scientific Advisory Group for Emergencies.

1) The UK LLC is a resource to enable the longitudinal perspective. The value of longitudinal research lies in the collection of a broad range of data, repeated at regular intervals allowing the investigation of interactions between things occurring over the lifetime of individuals. To enable this perspective to be applied to the Covid-19 research questions UoB require data from NHS Digital and other providers that covers the duration of the longitudinal population studies contributing their study data. For example, the National Study of Health and Development is still collecting data from babies born in one week of 1946. Emerging Covid-19 research for example around disease severity could feasibility be informed by patterns in childhood ill health and historic episodes of hospitalisation within these cohorts.

2) As above with the duration of NHS-linked data requested, the breadth of data within the LLC enable rapid response to emerging research priorities. UoB cannot predict what these priorities will be but can provide indicative examples:

The usefulness of a QOF exception code in GDPPR indicating that a patient has refused an annual health check are that this could be indicative of their behaviour around choosing to have vaccination; insights into vaccination and health behaviour have the potential to be of great importance. The risk is there are unforeseen needs which are then hard to fulfil if the data has been minimised.

The UK LLC is designed to support longitudinal research studies and research analysts working within the UK. Longitudinal research is geared to understand the longer-term outcomes of the pandemic and emerging interactions between health and social factors and including considerations of outcomes such as Long COVID whose duration is yet to be determined, and the impact of disruption of routine health service provision on wider mental and physical health.

LEGAL BASIS, ETHICS AND RISK OF POTENTIAL HARM

UoB will be sole Data Controller under this agreement.

UoB is establishing the UK LLC and using linked NHS records to inform the LH&W NCS research programme as a task carried out in the public interest (Article 6(1)(e) given the UK LLC is designed to answer specific high-priority research questions at a time of national crisis. Given that the requested health records include flows of sensitive information, the UK LLC also draws on Article 9(2)(j) in accordance with Article 89(1) where UoB's purpose is to conduct statistical scientific research. The case for this research being in the public interest is established through balancing the strong policy driver of improved understanding of the pandemic with consideration of potential harms to the patients/participants whose records are involved. Potential harms primarily relate to breach of confidentiality and the subsequent misuse of Personal Data or erosion of trust in longitudinal research/data science.

The lawful basis for the release and use of the confidential data being shared under this version of the agreement is Regulation 3(4) of the National Health Service (Control of Patient Information Regulations) 2002 (COPI) to require NHS Digital to share confidential patient information with organisations entitled to process this under COPI for COVID-19 purposes. The only permitted activities under this Data Sharing Agreement (DSA) are for COVID-19 purposes and within bounds of Reg 3(2) COPI. Reg 3 (2) COPI states that: "2) For the purposes of this regulation, “processing” includes any operations, or set of operations set out in regulation 2(2) which are undertaken for the purposes set out in paragraph (1)." The research relates to the monitoring and managing of COVID-19 and would therefore be covered by Reg 3(1)(d) of COPI.

The risk of this harm is minimised as this is:

1. an observational population-level research which will not result in a direct intervention to any participant;

2. the UK LLC uses tried and tested IT infrastructure and governance frameworks specifically designed to minimise risks to privacy during health data science;

3. the UK LLC is a fully de-identified research environment with sufficient controls that the risk of disclosure is not considered reasonably likely;

4. a reasonable expectation for the use of NHS data in this way has been established by the contributing studies, with studies making best endeavours to inform participants ahead of data extraction taking place, as well as continued fair processing (in relation to UK LLC) in line with usual policies and practices. Each stage of this fair processing will provide participants with the opportunity to object;

5. National Opt-Out will also be respected unless overridden by specific consent for the use of linked health records in the studies research programme;

6. all staff and users are vetted and approved professional researchers operating within controlled and auditable conditions.

This research is ethical given the strong scientific rationale, the process is transparent, best endeavours will be made to inform participants of the use of their data within UK LLC, with a right to object, that risks are mitigated and the participants (and wider public) are likely to directly benefit from the research through improved health care and government policy provision in response to the unprecedented Covid-19 crisis. Participants will be informed about how their data is used through ongoing fair processing communications including details of the process for ‘opting-out’.

OPERATIONAL MANAGEMENT

UoB is the study sponsor and Data Controller. UK LLC staff at UoB and UoE have responsibility for day to day running of the UK LLC databank: only UoB staff will have access to the data and TRE. The UK LLC Director, principle and co-investigators and data managers are employed by UoB. The UK LLC Co-Director and UK LLC administrative staff (responsible for administrative operations, communications, and participant/public engagement) are based at UoE). The Longitudinal Health & Wellbeing National Core Study is led from University College London (UCL) and the UK LLC reports into UCL on delivery of UK LLC objectives.

The UK LLC databank is configured to have two distinct classes of operational areas, which have three functions:

1) Data management and processing: All operational control and access for data processing and management of de-identified, including but not limited to data processing and storage, descriptive analysis and report/publication writing, subset, and provision of data to approved users is restricted to UoB staff. UoB staff are the only individuals who can access all data across all studies and sources in their raw and processed forms.

2a) Onward sharing to contributing studies: A secure operating partition will be created for each contributing study. Within these linked LPS - NHS data, restricted to the participants of that study, can be access by approved study staff through the ‘onward sharing’ mechanism. The study staff will use this area to conduct descriptive analysis on their participants (e.g., to provide documentation to describe the data asset, to inform bias assessments) and to assess feasibility of projects (e.g., to determine whether there are sufficient cases to scientifically justify inclusion in a proposed project) to inform potential users and to conduct study-level due diligence.

2b) Onward sharing to research analysts: Secure operating partitions will be created for each UK LLC approved project. The UK LLC approved users for these projects will have access to their folder which will contain a sub-set of data relevant to that study where onward sharing conditions of data providers are met. The UK LLC research users will control of data processing and analysis and leading publications.

The UoB UK LLC staff and UK LLC infrastructure system administrators will have access across all three operational areas.

The University of Swansea (UoS) is a Data Processor who provide data infrastructure and data management services to UoB and will maintain the integrity of the TRE, will assist with the data management processing (Under UoB’s direction) and will conduct the output disclosure assessments (to UoB’s specification). The UoS have sub-contracted the Data Linkage functions of the UKSeRP (including all management of participant identifiers) to the Digital Health and Care Wales (DHCW) (part of Velindre University NHS Trust) who as such, are Data Processors for UoB.

The UK LLC is funded by HM Treasury as part of the National Core Studies for Covid-19 research and is administered by the UK Medical Research Council (MRC). HM Treasury and MRC will not have access to NHS Digital data.

SUB-LICENSING

For the UK LLC to be a useful and sustainable resource for researchers to interrogate, and obtain meaningful data in a timely manner, contractual arrangements will need to facilitate high-volume and rapid turn-around of data requests. This will only be possible through removing the burden of the need for a high number of additional data sharing agreements across multiple institutions which would impede the speed of research. Therefore, a contract structure is needed to control the data flows, to control the purposes and way these data are processed, and to assign roles and responsibilities and to enforce the governance requirements of the individual studies and data owners and the legal basis under which they are permitted to operate.

These agreements are as follows:

- Data Deposit Agreement(s) established between UoB and the studies institution(s), ensuring fair processing will be carried out, and allowing study collected data to be made available for Covid-19 research to approved users within the UK LLC TRE.

- Shared data will be deposited into TRE where it will be processed, curated, and analysed.

- TRE established through research collaboration with University of Swansea (UoS) controlled through a Data Deposit

Agreement between UoB and UoS comprising:

1) a data deposit agreement for Welsh records and other records (SymptomTracker) to be deposited in the UK LLC and used for research

2) a collaborative contract where UoB procure a UKSeRP (UoS as Data Processor) and UoS data science expertise (an academic collaboration which involves Data Processing). Some aspects of the UoS contribution will be fulfilled by sub-contractor Data Health and Care Wales (DHCW) acting as a Trusted Third Party controlled through the Sub-Contract Agreement between UoS and DHCW.

The contract structure also permits the processes by which the participants of these studies can be identified by the NHS and their records selected and provided into the UK LLC for integration with the study data and onward sharing with researchers. The agreements for selecting and depositing these data has two distinct parts:

1) Record Linkage

• UoB will instruct the contributing studies to send a file of participant identifiers to NHS Wales Informatic Service (controlled through the Data Deposit Agreement between UoB and each contributing studies institution(s)).

• UoB will instruct the DHCW (a process controlled through the TRE procurement contract between UoB and UoS, and as described in the Data Deposit Agreement between UoB and contributing studies) to send participant identifiers to NHS Digital for record linkage purposes

• NHS Digital will use these identifiers to link these to their patient register and select the longitudinal study participants data (controlled through the Data Sharing Agreement between UoB and NHS Digital).

2) Data Deposition & Processing

• NHS Digital will then de-identify and deposit the selected data into the UK LCC where UoB will become the data controller of this patient data, for the purpose of processing it, approving onward research use and providing managed access to relevant sub-sets of data for Covid-19 research (controlled through the Data Sharing Agreement between UoB and NHS Digital). The structure will then enable the UK LLC to determine the onward processing of the integrated data for Covid-19 research purposes by reviewing applications from researchers and onwardly sharing sub-sets of relevant data within the TRE to approved users (controlled through the Data Access Agreement between UoB and approved researchers’ institutions). The nature of the sublicensing arrangements above are in line with the NHS Digital standards - https://digital.nhs.uk/services/data-access-request-service-dars/dars-guidance/sub-licencing-and-onward-sharing-of-data. NHS Digital requirement to audit data use is a requirement in the onward sharing contract and remains feasible in the UK LLC framework.

The UK LLC is likely to generate diverse outputs, with these forming a mix of rapid insight reports to policy makers and substantive research reports. The NCS is specifically designed to establish a clear channel between decision makers and researchers and the UK LLC is designed to enable rapid response and up-to-date information needed to support this. The likely outputs include:

* Peer-reviewed scientific publication (and pre-print publications of initial drafts prior to review which are designed enable rapid awareness of the findings), outputs will be ongoing, preprints and publications to date:

Preprints | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

Publications | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London

* Conference presentations and workshop presentations (to academic, health service practitioners and planners, government policy makers, Scientific Advisory Group for Emergencies (SAGE) members). These will be ongoing as part of our communications and engagement strategy, events to date include: Specifically for UK Longitudinal Linkage Collaboration:

Presentation: Welcome Longitudinal Population Studies conference (2021)

Regular presentations to Health Data Research UK - Data & Connectivity National Core Studies Delivery Group (2021-2022)

Presentation: CLOSER Longitudinal Population Studies conference (2022) ‘UK LLC Overview and explaining the innovations made and challenges faced’

Presentation: LHW National Core Study - Joint EHR cohort leads meeting ‘how to access UK LLC’

Workshop 1: to National Core Studies Analysts ‘Overview of UK LLC and how to access UK LLC’

Workshop 2: to CLOSER, National Core Studies and Vanguard Members ‘overview of UK LLC and how to access data

Presentation to UK LLC Communications Network (2022)

Youtube: UK Longitudinal Linkage Collaboration Verbal YouTube Public Involvement Induction (2022)

Presentation: to Elizabeth Blackwell Institute Health Data Science Showcase (2021)

Presentation: ESRC Strategic Advisory Group ‘ Understanding Data for Population research’ (2021)

Presentation: to Longitudinal Area Network and Data-link (LAND) community of practice (2022)

Poster Presentation: to Welcome Trust Longitudinal Studies Conference ‘overview of UK LLC’ (2021)

Presentation and advice: to Welcome Longitudinal Population Study COVID-19 Steering Group ‘provided specialist advice on linkage opportunities’ (2021)

Presentation and advice: to Population Research UK Programme Group ‘provided specialist advice’ (2021)

Advice: to Department of Education Scientific Advisory Group ‘provided specialist advice on record linkages in longitudinal studies (2021)

* Reports to health service practitioners and planners, government policy makers and Committee’s (e.g., National Institute for Health and Care Excellence (NICE), SAGE), it is anticipated that these will be ongoing, briefings to date include:

Briefing Reports | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

* Rapid evidence synthesis reports as requested by health and government policy makers (through the entirety of the 3-year NCS programme);

* Lay summaries provided through websites, Blogs, social media posts, privacy notices and outputs designed to promote the transparent use of data and the wider research process, regularly updated on our data-use register as follows:

Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk)

All outputs from the TRE will be assessed for disclosure risk and will be anonymous population level findings. The framework for this is based on leading 'five safes’ approach and is closely adapted from equivalent processes accredited to standards of the Office for National Statistics and the UK Statistics Authority.

All ‘meta’ products of the UK LLC research (such as code list definitions, syntax, workings for derived variables) will be made available for reuse and interrogation/replication. Access to the UK LLC is ‘open’ in the sense that any legitimate researcher conducting Covid-19 research can apply for access using a transparent and consistent decision-making process. As such, the UK LLC enhanced resource should be considered as an important ‘output’ and a measurable benefit of the UK LLC/LH&W NCS.

The high profile of Covid-19 means that many findings may have media interest. Importantly, the UK LLC is establishing a Patient/Participant Involvement and Engagement strategy within its communication work-package package with a dedicated budget and experienced Communications PPIE Officer. The package will build on existing and established relationships with participants to involve public/participants in the co-design of mechanisms to engage and effectively communicate UK LLC findings more widely with the public. UoB are committed to public involvement in our design and communications. UoB have already included HDRUK PPIE panel members (public representatives) in a consultation exercise and will draw on the extensive network of participants to further explore and co-develop dissemination mechanisms, in developing the end-to-end animation in the first instance. UoB are also developing with the studies opportunities for a subgroup of participants to be involved, in the data access decision making process for example. UoB have ensured that the programme of work is adequately resourced to compensate the public for their time, this has been an integral part of the project planning.

The UK LLC is a novel research infrastructure. The creation of this for Covid-19 research may provide methodological insights and illustrations of a new way of working for linkage in longitudinal research. Insights from this will be provided to the longitudinal community and its funders/key stakeholders (through academic papers, presentations, and contributions to advisory groups).

There are ten projects approved with data provisioned within the TRE to 20 researchers across seven institutions with six more projects in the pipeline. Key questions being addressed are: comparing the burden of long COVID in the community; mental health and wellbeing consequences of contracting COVID19;impacts of the COVID-19 pandemic on access to elective care in the UK. A full list of approved projects can be found here: Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk).

Findings from the LH&W NCS have led to Cabinet office and NHS England 'teach in' sessions, a national GP alert, an NHS enhanced service specification , rapid reports for SAGE, and they have provided key evidence to guide the NICE long COVID guidelines.

• During lockdown the LHW Team identified instances of inappropriate switching of blood thinners, triggering a national alert to GPs

• Policy makers have been provided with information on long COVID burden of disease, risk factors, definition and long term outcomes via cabinet briefing reports and SAGE reports

• Low levels of long COVID GP coding were reported. This led to a NHS enhanced service specification, directed at GPs to drive an increase in long COVID coding

• NICE used long COVID findings in its evidence to review risk factors

• Briefing notes submitted addressing inequalities in healthcare disruption, mental health decline during lockdown and the impact of furlough on health behaviour an mental health reasons for a low antibody response.

• LHW work was amongst the first to provide Chief Medical Officer with a quantification of the first vaccine side effects and identify differences in vaccine update among ethnic groups, and point out that people with learning disabilities should be prioritised.

The team have also presented early results from the first wave of the convalescence study qualitative interviews to NHS England, highlighting that long COVID participants are failing at the first hurdle of accessing healthcare. The team also communicated that there is an increased risk of developing cardiovascular complications up to a year after covid-19 infection, even in the absence of persistent symptom reporting.

Throughout the project we have worked with the public via our advisory group, to shape and communicate our programme. Public contributors have helped to create material aimed at communicating the results to all stakeholders, including infographics, short videos and an animation. We have also created a long COVID forum to host a diverse and inclusive public/patient conversation on defining and researching long COVID.

UoB will continue to have legal basis under Regulation 3 of the COPI Regulations for processing CPI for Covid-19 Purpose. Where University of Bristol (UoB) will use the data for previously approved research related to a Covid-19 Purpose, and this research is still ongoing, UoB are able to continue to meet the requirements in Regs 3(1), 3(3) and 7 of the COPI Regulations. DARS have assessed that UoB do still meet the 3 limbs of the COPI regulation.

The UK Longitudinal Linkage Collaboration (UK LLC) project itself is cross-institutional and has staff based at University of Bristol and the University of Edinburgh. The University of Edinburgh are collaborators however they have no role in determining how the data will be processed, nor will they have any access to the record level data. For clarification regarding the distinct staff groups noted in the agreement:

Study staff = staff from the contributing cohorts;

University of Bristol (UoB) staff = staff of UK Longitudinal Linkage Collaboration (LLC) employed by UoB;

UK LLC staff = staff employed by UK LLC (UoB and University of Edinburgh (UoE) staff).

BACKGROUND

The UK Chief Scientific Advisor has established (October 2020) a programme of National Core Studies (NCS) for SARS-CoV-2 (Covid-19) research as a coordinated, long-term, national research initiative. This will consider Covid-19 in terms of a viral pandemic (including issues of cases, transmission, symptoms, and outcomes) and in terms of the health and social impacts of behavioural restrictions designed to mitigate the harms of the pandemic. The NCS has six different sub-programmes which are addressing major Covid-19 research areas; one of these is the Longitudinal Health and Well-being (LH&W) NCS which is designed to use data from longitudinal studies to address the impact of Covid-19 and of associated viral suppression measures on health and well-being. The LLC is the central hub component of the Longitudinal Health and Well-being NCS.

The UK LLC has been designed to underpin the LH&W NCS, although not exclusively as the NCS are designed to support each other, where for example: the LLC could form the infrastructure for the long-term patient follow-up of consenting trial participants (e.g. participants in the RECOVERY trial) or, where UK longitudinal population studies (LPS) are being used to collect specific new study data which is not available through routine records (e.g. the ALSPAC birth cohort is providing a frame to collect biological samples to inform the Immunology NCS research programme; the assayed results of which will be uploaded into the LLC). For this reason, users of the LLC may come from across the full range of NCS studies and the resource will be accessible to other legitimate UK-based researchers investigating Covid-19 through a sub-licence framework. The NCS is planned to be a two/three-year research programme commencing October 2020.

THE UK LLC & DATA SUBJECTS

UoB (for LLC purposes) will underpin the NCS programme by combining, within a Trusted Research Environment, study data from >15 major inter-disciplinary UK LPS with a combined total of 1 – 2 million UK participants, centrally link to a wide range of Covid-19 relevant and non-health administrative records to utilise alongside the UK Biobank, the Zoe Symptom Tracker cohort and other sources. The integrated data, infrastructure and accompanying governance aspects will collectively be known as the UK LLC.

The de-identified data from contributing studies will be used to establish a UK LLC participant and household register, this will be used within the UK LLC to index and catalogue available data. Researchers will be able request access to the integrated data via the Health Data Research UK Gateway (https://www.healthdatagateway.org/) which through the NCS programme is being established as the de facto application point for UK Data Science projects and as a systematic and coordinated infrastructure for assessing the legitimacy and ‘safe’ credentials of a project. The value of these LPS data assets will be significantly enhanced through linkage to Covid-19 relevant health and other routine records (e.g., employment status, occupation, and benefit receipt records). This will allow the use of objectively assessed diagnoses and other records and the collection of data not amenable to self-report (such as detailed records of care, or health service utilisation). It will also allow researchers to assess and quantify/document bias and error through triangulating data collected across different organisations and settings.

Throughout the development of the UK LLC, UoB have worked collaboratively with UK LPS to co-design and develop the protocol and methodology. One of the strengths of the collaborative approach championed by the UK LLC is that UK LPS are committed and enthusiastic partners in this work: this initiative comes from within the community. The UK LLC will provide infrastructure to support the LPS and researchers working within the UK to investigate high priority COVID-19 research questions. UK LLC/UoB will continue to seek input from the LPS into decision making, guidance in future development and operation. Crucially, the LPS’s all maintain and manage ‘their’ relationships with ‘their’ participants including making critical decisions about data use, and communicating with participants about the existence of, and work of, the UK LLC and providing reassurance that this does not alter the participants existing relationship.

Data for the following studies will flow under COPI:

1. The National Study of Health and Development (NSHD)

2. The Southall And Brent REvisited (SABRE)

3. Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”)

4. Genetic Links to Anxiety and Depression Study (GLAD)

5. English Longitudinal Study of Ageing (ELSA)

6. 1958 National Child Development Study (NCDS)

7. 1970 British Cohort Study (BCS 70);

8. Next Steps

9. The Millenium Cohort Study (MCS)

10. Track-COVID - subset of INTERVAL COMPARE and STRIDES bioresource

11. NIHR (National Institute for Health Research) BioResouce

12. Extended Cohort for E-health, environment, and DNA (EXCEED)

13. Understanding Society

14. Born in Bradford

15. European Prospective Investigation into Cancer (EPIC)

16. The Fenland Study

17. Twins early Development Study (TEDS)

18. INTERVAL

19. COMPARE

20. STRIDES

21. TwinsUK

There is a longer-term objective to include a larger number of studies into the UK LLC to boost statistical power, increase participant heterogeneity and sample size in population and outcome sub-groups and to increase the range of self-reported information. Any additional study would be included as an amendment to this application.

The UoB scientific programme requires the extraction and use of the NHS Digital data compiled with Covid-19 relevant datasets (primary care, secondary care, community, mental health care provision, Covid-19 testing and outcomes data, NHS service use interactions such as NHS 111 records, mortality, disease registry and demographics data) for the purpose of establishing the UK LLC.

The UK LLC has conducted an assessment as to whether the data held and processed within the UK LLC is identifiable/potentially identifiable and therefore Personal Data which is subject to the Data Protection Act 2018 (DPA), UK General Data Protection Regulations (UK GDPR) and Common Law Duty of Confidentiality (Common Law) in accordance with guidance issued by the Information Commissioner’s Office. The assessment determined that the infrastructure, incorporating all data flows, should be considered as containing Personal Data and being subject to the DPA/UK GDPR and Common Law requirements. However, within this the assessment identified that the UK LLC sits at the threshold of being considered as holding pseudonymous data (Personal Data) and de-identified data where the risk of re-identification is not reasonably likely (not Personal Data), and that this differing status manifests when considering the data from the perspective of the different high-level groups of UK LLC data users and their distinct processing activities. It is evident that for the contributing studies, the use of data is either identifiable (for the provision of identifiers for linkage purposes) or pseudonymous (for the provision of data into the UK LLC and the processing of study data by study staff in their demarcated study processing area) and is therefore Personal Data. Yet, for the UK LLC staff and researcher users, the risk of re-identification of data within the UK LLC is not reasonably likely. This status, for these users, has been achieved through the deployment of data level as well as contextual controls which mitigate the risk of identification. The principal amongst these is the UK LLC operates as a ‘Trusted Research Environment’ which ensures that minimised data are used by minimised number of legitimate users within a secure environment for approved and proportionate purposes.

To contribute data into the UK LLC, LPS’s will be required to flow identifiers to the NHS to establish/filter a linkage between each participants' study record and their NHS registry record(s). A reasonable expectation will have been set through establishing fair processing for the use of NHS records via the UK LLCs infrastructure and for the sharing of these for public benefit research. Studies will update their privacy notices and notify participants of their involvement with UK LLC, explaining what this means and provide options to ‘opt-out’, in line with existing policy and practice.

UK LLC TRUSTED RESEARCH ENVIRONMENT (TRE)

The LLC will be built on a UK Secure eResearch Platform (UKSeRP). This is a well-established infrastructure for data science: it has ISO27001 and NHS Data Security Protection toolkit and Office for National Statistics accreditation under the Digital Economy Act 2017. Other instances of this environment are already used to host LPS and NHS Digital data and have been audited and commended as being well suited to the purpose and low risk.

The UK LLC TRE will be owned, operated, and managed by the LLC team (data management based at UoB). Under instruction from UoB, contributing LPS and data owners including NHS Digital will provide data to the UK LLC TRE using the ‘split file’ approach which is used across UKSeRP. This split file approach (details below) ensures the physical and processing operational separation of personal identifiers and participants de-identified study data and health records. This separation means that no party involved in this process can see both identifiers and data. This split file approach is coupled with the technological and socio-governance controls applied at the UK LLC (e.g., training, contracts, policy and procedures, penalties for misuse). The UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. For UK LLC staff and research users the risk of re-identification is controlled to the point that it is not reasonably likely.

UK LLC C-19 RESEARCH PROGRAMME

The UK LLC is specifically commissioned as a database resource to inform the LH&W NCS but is also designed to be a resource for the investigation of other Covid-19 questions as these are identified and prioritised through the Scientific Advisory Group for Emergencies (see the Health Data Research UK and SAGE research prioritisation funnel - current versions available through the HDRUK SAGE report, see - https://www.hdruk.ac.uk/covid-19/).

Given these questions are likely to change over time (e.g., the identification of ‘Long COVID’ as a potential new set of complex outcomes is introducing important new research questions), the research programme the LLC is designed to inform can be defined as the investigation of Covid-19 questions using data from longitudinal studies linked to routine records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing. For the use of linked longitudinal studies and NHS records this definition is refined to make clear the investigations primary outcomes must be focused on generating benefits to the health and social care system.

The investigation of Covid-19 questions using data from longitudinal studies linked to routine NHS records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing: where the primary research outcome is designed to benefit the health and social care system. These linked data will be of unique value in underpinning a programme of research on Covid-19 informed by the data assets within the UK LLC.

This program will seek to understand the patterns and predictors of infection, (including re-infection) and disease outcomes (such as ‘long covid’), and the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. It will also explore population level changes to physical and mental health, including hospital admission and mortality, in association with viral suppression measures and how these relate to changes in health behaviours.

Given the impact of the Covid-19 containment measures, it is important to investigate the role of socio-economic and neighbourhood/environmental factors in determining population level impacts to physical and mental health, to identify both groups at risk, and factors which offer resilience to adverse outcomes. Within this, UK LLC users will analyse changes in health care service use (using self-reported data and NHS health records) to determine if patterns in these have changed during the Covid-19 pandemic. Linked NHS records will also inform consideration of service use in relation to pre-pandemic health status and regular service interactions (e.g., screening, health reviews and routine service take up (such as annual seasonal flu vaccinations)).

This programme, through the LH&W NCS will specifically seek to answer the following questions:

i) What are the mental health consequences of Covid-19 infection, and of viral suppression measures? How do these differ by key demographics, socioeconomic status and by prior mental health status?

ii) What are the risks, determinants, and expressions of the physical health consequences of long-Covid?

iii) What are the disruptions to health care services (primary and secondary), and what are the health consequences of these disruptions? How do they vary by geography, demographics, and socioeconomic status?

iv) What are the risks of re-infection?

v) How do measures of historical health & physiology impact risk of C-19 infection and its consequences?

vi) How do we best identify cases in population studies, triangulating symptoms, antigen, and antibody testing? How do antibody profiles differ by case status (e.g., those with and without symptoms), how do antibody profiles vary over time? What are the health determinants of such variability?

vii) What are the long-term socioeconomic consequences of viral suppression measures? How do these in turn impact on health and on health care utilisation?

These research themes have been scientifically reviewed by the UK Chief Scientific Advisor and the UK Chief Medical Officer and are high priority questions of critical national interest.

OTHER DATA PROCESSING PURPOSES

In addition to the primary research themes, UoB will:

1) use linked NHS data to help inform descriptive and documentary analysis of the UK LLC dataset. This will describe the combined LPS population and make clear the LLC denominator, its characteristics and how this sub-set of the population relates to the wider population (through comparison with aggregated populations statistics). This will inform research users and those (e.g., policy makers) seeking to draw inferences from findings.

2) conduct feasibility assessments to help determine the viability of proposed projects. To do this, UK LLC staff will query the UK LLC integrated dataset (LPS data and linked NHS Records) to determine if UK LLC have sufficient case numbers to be able to inform specific research proposals. This purpose enhances the ethical and governance case for the resource as it means data are not shared where they cannot meaningfully inform the science, it also helps ensure the efficient use of research funding and resources.

DATA MINIMISATION

The datasets and data items requested have been restricted to only those directly relevant to the Covid-19 research programme that this infrastructure is designed to support. UoB fully adopt the standard NHS Digital definition of the Covid-19 relevant dataset.

As described, the UK LLC C-19 research programme is wide ranging and seeks to understand patterns and predictors of infection, disease outcomes, the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. Datasets requested include measure of Covid-19 related outcomes and pre-Covid-19 baseline datasets (allowing the value of these longitudinal data to be maximised).

The linked data requested are minimised to include only data covering the period each contributing study was in follow-up. The data requested are further minimised to include only records of participants of studies contributing to the UK LLC and those who have objected to this use of their data are excluded. Information on participants who have ‘opted-out’ will be uploaded to the UK LLC quarterly so that those participants’ data will be excluded from future studies.

UoB are unable to minimise datasets further based on fields or episodes at this time as work continues to be conducted to inform the UK’s Covid-19 research programme which aims to be responsive to health and government policy makers. For example in HES, UoB cannot consider only episodes with a specific diagnosis or procedure at this time, UoB do not yet understand the longer-term health implications from Covid-19 infection or whether there exist patterns and predictors of Covid-19 infection based on previous episodes in HES.

The study have minimised datasets so they only include records of participants of longitudinal population studies contributing to the UK LLC and to NHS Digital data that are directly relevant to the research programme. Whilst UoB cannot provide justification through reference to specific analysis that will be conducted using these data, given their purpose as a resource developed to enable researchers to respond to the emerging priorities as determined by the UK Scientific Advisory Group for Emergencies.

1) The UK LLC is a resource to enable the longitudinal perspective. The value of longitudinal research lies in the collection of a broad range of data, repeated at regular intervals allowing the investigation of interactions between things occurring over the lifetime of individuals. To enable this perspective to be applied to the Covid-19 research questions UoB require data from NHS Digital and other providers that covers the duration of the longitudinal population studies contributing their study data. For example, the National Study of Health and Development is still collecting data from babies born in one week of 1946. Emerging Covid-19 research for example around disease severity could feasibility be informed by patterns in childhood ill health and historic episodes of hospitalisation within these cohorts.

2) As above with the duration of NHS-linked data requested, the breadth of data within the LLC enable rapid response to emerging research priorities. UoB cannot predict what these priorities will be but can provide indicative examples:

The usefulness of a QOF exception code in GDPPR indicating that a patient has refused an annual health check are that this could be indicative of their behaviour around choosing to have vaccination; insights into vaccination and health behaviour have the potential to be of great importance. The risk is there are unforeseen needs which are then hard to fulfil if the data has been minimised.

The UK LLC is designed to support longitudinal research studies and research analysts working within the UK. Longitudinal research is geared to understand the longer-term outcomes of the pandemic and emerging interactions between health and social factors and including considerations of outcomes such as Long COVID whose duration is yet to be determined, and the impact of disruption of routine health service provision on wider mental and physical health.

LEGAL BASIS, ETHICS AND RISK OF POTENTIAL HARM

UoB will be sole Data Controller under this agreement.

UoB is establishing the UK LLC and using linked NHS records to inform the LH&W NCS research programme as a task carried out in the public interest (Article 6(1)(e) given the UK LLC is designed to answer specific high-priority research questions at a time of national crisis. Given that the requested health records include flows of sensitive information, the UK LLC also draws on Article 9(2)(j) in accordance with Article 89(1) where UoB's purpose is to conduct statistical scientific research. The case for this research being in the public interest is established through balancing the strong policy driver of improved understanding of the pandemic with consideration of potential harms to the patients/participants whose records are involved. Potential harms primarily relate to breach of confidentiality and the subsequent misuse of Personal Data or erosion of trust in longitudinal research/data science.

The lawful basis for the release and use of the confidential data being shared under this version of the agreement is Regulation 3(4) of the National Health Service (Control of Patient Information Regulations) 2002 (COPI) to require NHS Digital to share confidential patient information with organisations entitled to process this under COPI for COVID-19 purposes. The only permitted activities under this Data Sharing Agreement (DSA) are for COVID-19 purposes and within bounds of Reg 3(2) COPI. Reg 3 (2) COPI states that: "2) For the purposes of this regulation, “processing” includes any operations, or set of operations set out in regulation 2(2) which are undertaken for the purposes set out in paragraph (1)." The research relates to the monitoring and managing of COVID-19 and would therefore be covered by Reg 3(1)(d) of COPI.

The risk of this harm is minimised as this is:

1. an observational population-level research which will not result in a direct intervention to any participant;

2. the UK LLC uses tried and tested IT infrastructure and governance frameworks specifically designed to minimise risks to privacy during health data science;

3. the UK LLC is a fully de-identified research environment with sufficient controls that the risk of disclosure is not considered reasonably likely;

4. a reasonable expectation for the use of NHS data in this way has been established by the contributing studies, with studies making best endeavours to inform participants ahead of data extraction taking place, as well as continued fair processing (in relation to UK LLC) in line with usual policies and practices. Each stage of this fair processing will provide participants with the opportunity to object;

5. National Opt-Out will also be respected unless overridden by specific consent for the use of linked health records in the studies research programme;

6. all staff and users are vetted and approved professional researchers operating within controlled and auditable conditions.

This research is ethical given the strong scientific rationale, the process is transparent, best endeavours will be made to inform participants of the use of their data within UK LLC, with a right to object, that risks are mitigated and the participants (and wider public) are likely to directly benefit from the research through improved health care and government policy provision in response to the unprecedented Covid-19 crisis. Participants will be informed about how their data is used through ongoing fair processing communications including details of the process for ‘opting-out’.

OPERATIONAL MANAGEMENT

UoB is the study sponsor and Data Controller. UK LLC staff at UoB and UoE have responsibility for day to day running of the UK LLC databank: only UoB staff will have access to the data and TRE. The UK LLC Director, principle and co-investigators and data managers are employed by UoB. The UK LLC Co-Director and UK LLC administrative staff (responsible for administrative operations, communications, and participant/public engagement) are based at UoE). The Longitudinal Health & Wellbeing National Core Study is led from University College London (UCL) and the UK LLC reports into UCL on delivery of UK LLC objectives.

The UK LLC databank is configured to have two distinct classes of operational areas, which have three functions:

1) Data management and processing: All operational control and access for data processing and management of de-identified, including but not limited to data processing and storage, descriptive analysis and report/publication writing, subset, and provision of data to approved users is restricted to UoB staff. UoB staff are the only individuals who can access all data across all studies and sources in their raw and processed forms.

2a) Onward sharing to contributing studies: A secure operating partition will be created for each contributing study. Within these linked LPS - NHS data, restricted to the participants of that study, can be access by approved study staff through the ‘onward sharing’ mechanism. The study staff will use this area to conduct descriptive analysis on their participants (e.g., to provide documentation to describe the data asset, to inform bias assessments) and to assess feasibility of projects (e.g., to determine whether there are sufficient cases to scientifically justify inclusion in a proposed project) to inform potential users and to conduct study-level due diligence.

2b) Onward sharing to research analysts: Secure operating partitions will be created for each UK LLC approved project. The UK LLC approved users for these projects will have access to their folder which will contain a sub-set of data relevant to that study where onward sharing conditions of data providers are met. The UK LLC research users will control of data processing and analysis and leading publications.

The UoB UK LLC staff and UK LLC infrastructure system administrators will have access across all three operational areas.

The University of Swansea (UoS) is a Data Processor who provide data infrastructure and data management services to UoB and will maintain the integrity of the TRE, will assist with the data management processing (Under UoB’s direction) and will conduct the output disclosure assessments (to UoB’s specification). The UoS have sub-contracted the Data Linkage functions of the UKSeRP (including all management of participant identifiers) to the Digital Health and Care Wales (DHCW) (part of Velindre University NHS Trust) who as such, are Data Processors for UoB.

The UK LLC is funded by HM Treasury as part of the National Core Studies for Covid-19 research and is administered by the UK Medical Research Council (MRC). HM Treasury and MRC will not have access to NHS Digital data.

SUB-LICENSING

For the UK LLC to be a useful and sustainable resource for researchers to interrogate, and obtain meaningful data in a timely manner, contractual arrangements will need to facilitate high-volume and rapid turn-around of data requests. This will only be possible through removing the burden of the need for a high number of additional data sharing agreements across multiple institutions which would impede the speed of research. Therefore, a contract structure is needed to control the data flows, to control the purposes and way these data are processed, and to assign roles and responsibilities and to enforce the governance requirements of the individual studies and data owners and the legal basis under which they are permitted to operate.

These agreements are as follows:

- Data Deposit Agreement(s) established between UoB and the studies institution(s), ensuring fair processing will be carried out, and allowing study collected data to be made available for Covid-19 research to approved users within the UK LLC TRE.

- Shared data will be deposited into TRE where it will be processed, curated, and analysed.

- TRE established through research collaboration with University of Swansea (UoS) controlled through a Data Deposit

Agreement between UoB and UoS comprising:

1) a data deposit agreement for Welsh records and other records (SymptomTracker) to be deposited in the UK LLC and used for research

2) a collaborative contract where UoB procure a UKSeRP (UoS as Data Processor) and UoS data science expertise (an academic collaboration which involves Data Processing). Some aspects of the UoS contribution will be fulfilled by sub-contractor Data Health and Care Wales (DHCW) acting as a Trusted Third Party controlled through the Sub-Contract Agreement between UoS and DHCW.

The contract structure also permits the processes by which the participants of these studies can be identified by the NHS and their records selected and provided into the UK LLC for integration with the study data and onward sharing with researchers. The agreements for selecting and depositing these data has two distinct parts:

1) Record Linkage

• UoB will instruct the contributing studies to send a file of participant identifiers to NHS Wales Informatic Service (controlled through the Data Deposit Agreement between UoB and each contributing studies institution(s)).

• UoB will instruct the DHCW (a process controlled through the TRE procurement contract between UoB and UoS, and as described in the Data Deposit Agreement between UoB and contributing studies) to send participant identifiers to NHS Digital for record linkage purposes

• NHS Digital will use these identifiers to link these to their patient register and select the longitudinal study participants data (controlled through the Data Sharing Agreement between UoB and NHS Digital).

2) Data Deposition & Processing

• NHS Digital will then de-identify and deposit the selected data into the UK LCC where UoB will become the data controller of this patient data, for the purpose of processing it, approving onward research use and providing managed access to relevant sub-sets of data for Covid-19 research (controlled through the Data Sharing Agreement between UoB and NHS Digital). The structure will then enable the UK LLC to determine the onward processing of the integrated data for Covid-19 research purposes by reviewing applications from researchers and onwardly sharing sub-sets of relevant data within the TRE to approved users (controlled through the Data Access Agreement between UoB and approved researchers’ institutions). The nature of the sublicensing arrangements above are in line with the NHS Digital standards - https://digital.nhs.uk/services/data-access-request-service-dars/dars-guidance/sub-licencing-and-onward-sharing-of-data. NHS Digital requirement to audit data use is a requirement in the onward sharing contract and remains feasible in the UK LLC framework.

The UK LLC is likely to generate diverse outputs, with these forming a mix of rapid insight reports to policy makers and substantive research reports. The NCS is specifically designed to establish a clear channel between decision makers and researchers and the UK LLC is designed to enable rapid response and up-to-date information needed to support this. The likely outputs include:

* Peer-reviewed scientific publication (and pre-print publications of initial drafts prior to review which are designed enable rapid awareness of the findings), outputs will be ongoing, preprints and publications to date:

Preprints | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

Publications | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London

* Conference presentations and workshop presentations (to academic, health service practitioners and planners, government policy makers, Scientific Advisory Group for Emergencies (SAGE) members). These will be ongoing as part of our communications and engagement strategy, events to date include: Specifically for UK Longitudinal Linkage Collaboration:

Presentation: Welcome Longitudinal Population Studies conference (2021)

Regular presentations to Health Data Research UK - Data & Connectivity National Core Studies Delivery Group (2021-2022)

Presentation: CLOSER Longitudinal Population Studies conference (2022) ‘UK LLC Overview and explaining the innovations made and challenges faced’

Presentation: LHW National Core Study - Joint EHR cohort leads meeting ‘how to access UK LLC’

Workshop 1: to National Core Studies Analysts ‘Overview of UK LLC and how to access UK LLC’

Workshop 2: to CLOSER, National Core Studies and Vanguard Members ‘overview of UK LLC and how to access data

Presentation to UK LLC Communications Network (2022)

Youtube: UK Longitudinal Linkage Collaboration Verbal YouTube Public Involvement Induction (2022)

Presentation: to Elizabeth Blackwell Institute Health Data Science Showcase (2021)

Presentation: ESRC Strategic Advisory Group ‘ Understanding Data for Population research’ (2021)

Presentation: to Longitudinal Area Network and Data-link (LAND) community of practice (2022)

Poster Presentation: to Welcome Trust Longitudinal Studies Conference ‘overview of UK LLC’ (2021)

Presentation and advice: to Welcome Longitudinal Population Study COVID-19 Steering Group ‘provided specialist advice on linkage opportunities’ (2021)

Presentation and advice: to Population Research UK Programme Group ‘provided specialist advice’ (2021)

Advice: to Department of Education Scientific Advisory Group ‘provided specialist advice on record linkages in longitudinal studies (2021)

* Reports to health service practitioners and planners, government policy makers and Committee’s (e.g., National Institute for Health and Care Excellence (NICE), SAGE), it is anticipated that these will be ongoing, briefings to date include:

Briefing Reports | COVID-19 Longitudinal Health and Wellbeing National Core Study - UCL – University College London,

* Rapid evidence synthesis reports as requested by health and government policy makers (through the entirety of the 3-year NCS programme);

* Lay summaries provided through websites, Blogs, social media posts, privacy notices and outputs designed to promote the transparent use of data and the wider research process, regularly updated on our data-use register as follows:

Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk)

All outputs from the TRE will be assessed for disclosure risk and will be anonymous population level findings. The framework for this is based on leading 'five safes’ approach and is closely adapted from equivalent processes accredited to standards of the Office for National Statistics and the UK Statistics Authority.

All ‘meta’ products of the UK LLC research (such as code list definitions, syntax, workings for derived variables) will be made available for reuse and interrogation/replication. Access to the UK LLC is ‘open’ in the sense that any legitimate researcher conducting Covid-19 research can apply for access using a transparent and consistent decision-making process. As such, the UK LLC enhanced resource should be considered as an important ‘output’ and a measurable benefit of the UK LLC/LH&W NCS.

The high profile of Covid-19 means that many findings may have media interest. Importantly, the UK LLC is establishing a Patient/Participant Involvement and Engagement strategy within its communication work-package package with a dedicated budget and experienced Communications PPIE Officer. The package will build on existing and established relationships with participants to involve public/participants in the co-design of mechanisms to engage and effectively communicate UK LLC findings more widely with the public. UoB are committed to public involvement in our design and communications. UoB have already included HDRUK PPIE panel members (public representatives) in a consultation exercise and will draw on the extensive network of participants to further explore and co-develop dissemination mechanisms, in developing the end-to-end animation in the first instance. UoB are also developing with the studies opportunities for a subgroup of participants to be involved, in the data access decision making process for example. UoB have ensured that the programme of work is adequately resourced to compensate the public for their time, this has been an integral part of the project planning.

The UK LLC is a novel research infrastructure. The creation of this for Covid-19 research may provide methodological insights and illustrations of a new way of working for linkage in longitudinal research. Insights from this will be provided to the longitudinal community and its funders/key stakeholders (through academic papers, presentations, and contributions to advisory groups).

There are ten projects approved with data provisioned within the TRE to 20 researchers across seven institutions with six more projects in the pipeline. Key questions being addressed are: comparing the burden of long COVID in the community; mental health and wellbeing consequences of contracting COVID19;impacts of the COVID-19 pandemic on access to elective care in the UK. A full list of approved projects can be found here: Data Use Register | UK Longitudinal Linkage Collaboration (ukllc.ac.uk).

Findings from the LH&W NCS have led to Cabinet office and NHS England 'teach in' sessions, a national GP alert, an NHS enhanced service specification , rapid reports for SAGE, and they have provided key evidence to guide the NICE long COVID guidelines.

• During lockdown the LHW Team identified instances of inappropriate switching of blood thinners, triggering a national alert to GPs

• Policy makers have been provided with information on long COVID burden of disease, risk factors, definition and long term outcomes via cabinet briefing reports and SAGE reports

• Low levels of long COVID GP coding were reported. This led to a NHS enhanced service specification, directed at GPs to drive an increase in long COVID coding

• NICE used long COVID findings in its evidence to review risk factors

• Briefing notes submitted addressing inequalities in healthcare disruption, mental health decline during lockdown and the impact of furlough on health behaviour an mental health reasons for a low antibody response.

• LHW work was amongst the first to provide Chief Medical Officer with a quantification of the first vaccine side effects and identify differences in vaccine update among ethnic groups, and point out that people with learning disabilities should be prioritised.

The team have also presented early results from the first wave of the convalescence study qualitative interviews to NHS England, highlighting that long COVID participants are failing at the first hurdle of accessing healthcare. The team also communicated that there is an increased risk of developing cardiovascular complications up to a year after covid-19 infection, even in the absence of persistent symptom reporting.

Throughout the project we have worked with the public via our advisory group, to shape and communicate our programme. Public contributors have helped to create material aimed at communicating the results to all stakeholders, including infographics, short videos and an animation. We have also created a long COVID forum to host a diverse and inclusive public/patient conversation on defining and researching long COVID.

**This is an extension to permit University of Bristol to continue to hold and process already disseminated data whilst University of Bristol address DARS and IGARD feedback. No new data will flow under v1.**

v1 of NIC-420168 is an Amendment to include additional cohort members that were initially recruited under various research studies and move from COPI to consent as the lawful basis to address the common law.

The LLC project itself is cross-institutional and has staff based at University of Bristol and the University of Edinburgh. The University of Edinburgh are collaborators however they have no role in determining how the data will be processed, nor will they have any access to the record level data. For clarification regarding the distinct staff groups noted in the agreement:

Study staff = staff from the contributing cohorts;

University of Bristol (UoB) staff = staff of UK Longitudinal Linkage Collaboration (LLC) employed by UoB;

UK LLC staff = staff employed by UK LLC (UoB and University of Edinburgh (UoE) staff).

BACKGROUND

The UK Chief Scientific Advisor has established (October 2020) a programme of National Core Studies (NCS) for SARS-CoV-2 (Covid-19) research as a coordinated, long-term, national research initiative. This will consider Covid-19 in terms of a viral pandemic (including issues of cases, transmission, symptoms, and outcomes) and in terms of the health and social impacts of behavioural restrictions designed to mitigate the harms of the pandemic. The NCS has six different sub-programmes which are addressing major Covid-19 research areas; one of these is the Longitudinal Health and Well-being (LH&W) NCS which is designed to use data from longitudinal studies to address the impact of Covid-19 and of associated viral suppression measures on health and well-being. The LLC is the central hub component of the Longitudinal Health and Well-being NCS.

The UK LLC has been designed to underpin the LH&W NCS, although not exclusively as the NCS are designed to support each other, where for example: the LLC could form the infrastructure for the long-term patient follow-up of consenting trial participants (e.g. participants in the RECOVERY trial) or, where UK longitudinal population studies (LPS) are being used to collect specific new study data which is not available through routine records (e.g. the ALSPAC birth cohort is providing a frame to collect biological samples to inform the Immunology NCS research programme; the assayed results of which will be uploaded into the LLC). For this reason, users of the LLC may come from across the full range of NCS studies and the resource will be accessible to other legitimate UK-based researchers investigating Covid-19 through a sub-licence framework. The NCS is planned to be a two/three-year research programme commencing October 2020.

THE UK LLC & DATA SUBJECTS

UoB (for LLC purposes) will underpin the NCS programme by combining, within a Trusted Research Environment, study data from >15 major inter-disciplinary UK LPS with a combined total of 1 – 2 million UK participants, centrally link to a wide range of Covid-19 relevant and non-health administrative records to utilise alongside the UK Biobank, the Zoe Symptom Tracker cohort and other sources. The integrated data, infrastructure and accompanying governance aspects will collectively be known as the UK LLC.

The de-identified data from contributing studies will be used to establish a UK LLC participant and household register, this will be used within the UK LLC to index and catalogue available data. Researchers will be able request access to the integrated data via the Health Data Research UK Gateway (https://www.healthdatagateway.org/) which through the NCS programme is being established as the de facto application point for UK Data Science projects and as a systematic and coordinated infrastructure for assessing the legitimacy and ‘safe’ credentials of a project. The value of these LPS data assets will be significantly enhanced through linkage to Covid-19 relevant health and other routine records (e.g., employment status, occupation, and benefit receipt records). This will allow the use of objectively assessed diagnoses and other records and the collection of data not amenable to self-report (such as detailed records of care, or health service utilisation). It will also allow researchers to assess and quantify/document bias and error through triangulating data collected across different organisations and settings.

Throughout the development of the UK LLC, UoB have worked collaboratively with UK LPS to co-design and develop the protocol and methodology. One of the strengths of the collaborative approach championed by the UK LLC is that UK LPS are committed and enthusiastic partners in this work: this initiative comes from within the community. The UK LLC will provide infrastructure to support the LPS and researchers working within the UK to investigate high priority COVID-19 research questions. UK LLC/UoB will continue to seek input from the LPS into decision making, guidance in future development and operation. Crucially, the LPS’s all maintain and manage ‘their’ relationships with ‘their’ participants including making critical decisions about data use, and communicating with participants about the existence of, and work of, the UK LLC and providing reassurance that this does not alter the participants existing relationship.

ROLE OF NHS DIGITAL

COHORTS TO BE INCLUDED

The contributing studies whose data will be linked to NHS Digital data in this application are:

1) The National Study of Health and Development (NSHD); which is owned by University College London (UCL; sole Data Controller). UCL have an existing Data Sharing Framework Contract with NHS Digital and NSHD have already flagged its cohort at NHS Digital (MR1) and has active data sharing agreements for the MR1 study (Ref: DARS-NIC-148100-6RFK9 (application covered by consent) and DARS-NIC-86954-Y0R2N (s251)). NSHD brings value as a nationally representative ageing cohort (age ~74) with lifelong follow-up and relevant Covid-19 data collections.

2) The Southall And Brent REvisited (SABRE); which is owned by University College London (UCL; sole Data Controller). UCL have an existing Data Sharing Framework Contract with NHS Digital. SABRE have not flagged its cohort members at NHS Digital (meaning a new study flagging will be established on behalf of SABRE through this application) and do not have an active NHS Digital data sharing agreement. SABRE brings value as a tri-ethnic cohort with considerable participant heterogeneity. It also characterises a relatively deprived urban area of the UK.

3) Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”); which is owned by University of Bristol (UoB; sole Data Controller). ALSPAC has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (a DARS approval for consenting participants Ref: NIC 13133-B7B3K; and a DARS approval for s251 participants: NIC 152414-W3P6Q). ALSPAC brings value as a multi-generational cohort: our original parents (Generation 0, G0), our index participants born in the early 1990s (Generation 1, G1) and now their children (Generation 2, G2) with relevant Covid-19 data collections.

4) TwinsUK; which is owned by Kings College London (KCL; sole Data Controller). TwinsUK has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (a DARS approval for consenting participants Ref: NHS Digital. DARS-NIC-147955-M8D2Q-v1.3. undergoing amendment for HES, demographics, MRIS, Flagging status. TwinsUK have passed DSPT 19/20 (Details of ISO 27001)). TwinsUK brings value as largest adult twin registry in the UK and the most clinically detailed in the world.

5) Twins early Development Study (TEDS); which is owned by Kings College London (KCL; sole Data Controller). TEDS do not have existing NHS Digital DARS approval. TEDS brings value through collection of genetic and environmental information through Twins early development to adolescence.

6) Genetic Links to Anxiety and Depression Study (GLAD); which is owned by Kings College London (KCL; sole Data Controller). GLAD do not have existing NHS Digital DARS approval. GLAD brings value through coverage of a population over the four nations and collecting information around risk factors for depression and anxiety.

7) English Longitudinal Study of Ageing (ELSA); which is owned by University College London (UCL; sole Data Controller). ELSA has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme across the three data controllers’ organisations: IFS: DARS-NIC-32854-Y8P8B-v2.2, UCL: DARS-NIC-30493-Y0C0K-v1.4, NatCen: DARS-NIC-32854-Y8P8B-v2.1. ELSA brings value as a national ageing cohort with regular follow-up and detailed information on physical and mental health as well as finances and attitudes to ageing.

8) 1958 National Child Development Study (NCDS); which is owned by the University College London Centre for Longitudinal Studies (CLS) (UCL; sole Data Controller). NCDS has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-49297-Q7G1Q). 1958 NDS brings value as a national birth cohort with information relating to lifelong factors influencing health and wellbeing development as well as economic inequalities, and now how these relate to retirement and ageing.

9) 1970 British Cohort Study (BCS 70); which is owned by the University College London Centre for Longitudinal Studies (CLS) (UCL; sole Data Controller). BCS 70 has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-49826-T0J7C). The 1970 British Cohort Study brings value as a birth cohort providing insight, amongst other things into education and social development, economic circumstances in a population of now 50+ year olds.

10) Next Steps; which is owned by the University College London Centre for Longitudinal Studies (CLS) (UCL; sole Data Controller). Next Steps has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS NIC-51342-V1M5W) for the extraction and use of linked NHS records in its research programme. Next Steps brings value with detailed information on academic performance through links to the National Pupil Database records and planned links to higher education and Universities, the cohort comprises people born in England in 1989-1990.

11) The Millenium Cohort Study (MCS); which is owned by the University College London Centre for Longitudinal Studies (CLS) (UCL; sole Data Controller). MCS has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-384504-N2V5B). The Millenium Cohort Study brings value through following the lives of 20 - 22 year olds and their residential parents, providing insight into family life and relationships.

12) INTERVAL; which is owned by the University of Cambridge (UoC; sole Data Controller). INTERVAL has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-156334-711SX). INTERVAL provides value through their representation of blood donors who have frequently given blood to investigate impacts on wellbeing.

13) COMPARE; which is owned by the University of Cambridge (UoC; sole Data Controller). COMPARE has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-156334-711SX). COMPARE provides value through information on blood donors haemoglobin.

14) STRIDES BioResouce; which is owned by the University of Cambridge (UoC; sole Data Controller). STRIDES do not have existing NHS Digital DARS approval. STRIDES provides value as a research blood donor resource and panel of volunteers willing to participate in medical and health related studies.

15) Track-COVID; which is owned by the University of Cambridge (UoC; sole Data Controller). Track-COVID do not have existing NHS Digital DARS approval. Track-COVID provides value with a subset of INTERVAL COMPARE and STRIDES bioresource to determine risk factors for infection.

16) NIHR (National Institute for Health Research) BioResouce; which is owned by the Cambridge University Hospitals NHS Foundation Trust (sole Data Controller). NIHR BioResource provides value through holding unique genetic information including eating and neurological disorders.

17) Extended Cohort for E-health, environment, and DNA (EXCEED); which is owned by the University of Leicester (UoL; sole Data Controller). EXCEED do not have existing NHS Digital DARS approval. EXCEED provides value through collecting information on environment and DNA to provide insight into environmental influences on long-term health.

18) Understanding Society; which is owned by the University of Essex (UoE; sole Data Controller). Understanding Society do not have an existing NHS Digital DARS approval. Understanding Society provides value as a household study providing a generational perspective and changes over the long-term, particularly social, and economic factors.

19) Born in Bradford; which is owned by Bradford Teaching Hospitals NHS Foundation Trust (BTHFT; sole Data Controller). Born in Bradford do not have an existing NHS Digital DARS approval. Born in Bradford provides value through representation of a multi-ethnic population and regional cohort from an area which has high levels of deprivation.

20) The National Study of Health and Development (NSHD); which is owned by University College London (UCL; sole Data Controller). UCL have an existing Data Sharing Framework Contract with NHS Digital and NSHD have already flagged its cohort at NHS Digital (MR1) and has active data sharing agreements for the MR1 study (Ref: DARS-NIC-148100-6RFK9 (application covered by consent) and DARS-NIC-86954-Y0R2N (s251)). NSHD brings value as a nationally representative ageing cohort (age ~74) with lifelong follow-up and relevant Covid-19 data collections.

21) The Southall And Brent REvisited (SABRE); which is owned by University College London (UCL; sole Data Controller). UCL have an existing Data Sharing Framework Contract with NHS Digital. SABRE have not flagged its cohort members at NHS Digital (meaning a new study flagging will be established on behalf of SABRE through this application) and do not have an active NHS Digital data sharing agreement. SABRE brings value as a tri-ethnic cohort with considerable participant heterogeneity. It also characterises a relatively deprived urban area of the UK.

22) Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”); which is owned by University of Bristol (UoB; sole Data Controller). ALSPAC has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (a DARS approval for consenting participants Ref: NIC 13133-B7B3K; and a DARS approval for s251 participants: NIC 152414-W3P6Q). ALSPAC brings value as a multi-generational cohort: our original parents (Generation 0, G0), our index participants born in the early 1990s (Generation 1, G1) and now their children (Generation 2, G2) with relevant Covid-19 data collections.

23) TwinsUK; which is owned by Kings College London (KCL; sole Data Controller). TwinsUK has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (a DARS approval for consenting participants Ref: NHS Digital. DARS-NIC-147955-M8D2Q-v1.3. undergoing amendment for HES, demographics, MRIS, Flagging current status. TwinsUK have passed DSPT 19/20 (Details of ISO 27001)). TwinsUK brings value as largest adult twin registry in the UK and the most clinically detailed in the world. Only data covering participants in England will be included.

There is a longer-term objective to include a larger number of studies into the UK LLC to boost statistical power, increase participant heterogeneity and sample size in population and outcome sub-groups and to increase the range of self-reported information. Any additional study would be included as an amendment to this application. Additional studies not included in this agreement will be added as an amendment to this agreement.

The UoB scientific programme requires the extraction and use of the NHS Digital data compiled with Covid-19 relevant datasets (primary care, secondary care, community, mental health care provision, Covid-19 testing and outcomes data, NHS service use interactions such as NHS 111 records, mortality, disease registry and demographics data) for the purpose of establishing the UK LLC.

The UK LLC has conducted an assessment as to whether the data held and processed within the UK LLC is identifiable/potentially identifiable and therefore Personal Data which is subject to the Data Protection Act 2018 (DPA), UK General Data Protection Regulations (UK GDPR) and Common Law Duty of Confidentiality (Common Law) in accordance with guidance issued by the Information Commissioner’s Office. The assessment determined that the infrastructure, incorporating all data flows, should be considered as containing Personal Data and being subject to the DPA/UK GDPR and Common Law requirements. However, within this the assessment identified that the UK LLC sits at the threshold of being considered as holding pseudonymous data (Personal Data) and de-identified data where the risk of re-identification is not reasonably likely (not Personal Data), and that this differing status manifests when considering the data from the perspective of the different high-level groups of UK LLC data users and their distinct processing activities. It is evident that for the contributing studies, the use of data is either identifiable (for the provision of identifiers for linkage purposes) or pseudonymous (for the provision of data into the UK LLC and the processing of study data by study staff in their demarcated study processing area) and is therefore Personal Data. Yet, for the UK LLC staff and researcher users, the risk of re-identification of data within the UK LLC is not reasonably likely. This status, for these users, has been achieved through the deployment of data level as well as contextual controls which mitigate the risk of identification. The principal amongst these is the UK LLC operates as a ‘Trusted Research Environment’ which ensures that minimised data are used by minimised number of legitimate users within a secure environment for approved and proportionate purposes.

To contribute data into the UK LLC, LPS’s will be required to flow identifiers to the NHS to establish/filter a linkage between each participants' study record and their NHS registry record(s). A reasonable expectation will have been set through establishing fair processing for the use of NHS records via the UK LLCs infrastructure and for the sharing of these for public benefit research. Studies will update their privacy notices and notify participants of their involvement with UK LLC, explaining what this means and provide options to ‘opt-out’, in line with existing policy and practice.

UK LLC TRUSTED RESEARCH ENVIRONMENT (TRE)

The LLC will be built on a UK Secure eResearch Platform (UKSeRP). This is a well-established infrastructure for data science: it has ISO27001 and NHS Data Security Protection toolkit and Office for National Statistics accreditation under the Digital Economy Act 2017. Other instances of this environment are already used to host LPS and NHS Digital data and have been audited and commended as being well suited to the purpose and low risk (e.g. https://digital.nhs.uk/binaries/content/assets/website-assets/services/dars/data-sharing-post-audit-review_university-of-bristol-.pdf).

The UK LLC TRE will be owned, operated, and managed by the LLC team (data management based at UoB). Under instruction from UoB, contributing LPS and data owners including NHS Digital will provide data to the UK LLC TRE using the ‘split file’ approach which is used across UKSeRP. This split file approach (details below) ensures the physical and processing operational separation of personal identifiers and participants de-identified study data and health records. This separation means that no party involved in this process can see both identifiers and data. This split file approach is coupled with the technological and socio-governance controls applied at the UK LLC (e.g., training, contracts, policy and procedures, penalties for misuse). The UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. For UK LLC staff and research users the risk of re-identification is controlled to the point that it is not reasonably likely.

UK LLC C-19 RESEARCH PROGRAMME

The UK LLC is specifically commissioned as a database resource to inform the LH&W NCS but is also designed to be a resource for the investigation of other Covid-19 questions as these are identified and prioritised through the Scientific Advisory Group for Emergencies (see the Health Data Research UK and SAGE research prioritisation funnel - current versions available through the HDRUK SAGE report, see - https://www.hdruk.ac.uk/covid-19/).

Given these questions are likely to change over time (e.g., the identification of ‘Long COVID’ as a potential new set of complex outcomes is introducing important new research questions), the research programme the LLC is designed to inform can be defined as the investigation of Covid-19 questions using data from longitudinal studies linked to routine records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing. For the use of linked longitudinal studies and NHS records this definition is refined to make clear the investigations primary outcomes must be focused on generating benefits to the health and social care system.

The investigation of Covid-19 questions using data from longitudinal studies linked to routine NHS records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing: where the primary research outcome is designed to benefit the health and social care system. These linked data will be of unique value in underpinning a programme of research on Covid-19 informed by the data assets within the UK LLC.

This program will seek to understand the patterns and predictors of infection, (including re-infection) and disease outcomes (such as ‘long covid’), and the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. It will also explore population level changes to physical and mental health, including hospital admission and mortality, in association with viral suppression measures and how these relate to changes in health behaviours.

Given the impact of the Covid-19 containment measures, it is important to investigate the role of socio-economic and neighbourhood/environmental factors in determining population level impacts to physical and mental health, to identify both groups at risk, and factors which offer resilience to adverse outcomes. Within this, UK LLC users will analyse changes in health care service use (using self-reported data and NHS health records) to determine if patterns in these have changed during the Covid-19 pandemic. Linked NHS records will also inform consideration of service use in relation to pre-pandemic health status and regular service interactions (e.g., screening, health reviews and routine service take up (such as annual seasonal flu vaccinations)).

This programme, through the LH&W NCS will specifically seek to answer the following questions:

i) What are the mental health consequences of Covid-19 infection, and of viral suppression measures? How do these differ by key demographics, socioeconomic status and by prior mental health status?

ii) What are the risks, determinants, and expressions of the physical health consequences of long-Covid?

iii) What are the disruptions to health care services (primary and secondary), and what are the health consequences of these disruptions? How do they vary by geography, demographics, and socioeconomic status?

iv) What are the risks of re-infection?

v) How do measures of historical health & physiology impact risk of C-19 infection and its consequences?

vi) How do we best identify cases in population studies, triangulating symptoms, antigen, and antibody testing? How do antibody profiles differ by case status (e.g., those with and without symptoms), how do antibody profiles vary over time? What are the health determinants of such variability?

vii) What are the long-term socioeconomic consequences of viral suppression measures? How do these in turn impact on health and on health care utilisation?

These research themes have been scientifically reviewed by the UK Chief Scientific Advisor and the UK Chief Medical Officer and are high priority questions of critical national interest.

OTHER DATA PROCESSING PURPOSES

In addition to the primary research themes, UoB will:

1) use linked NHS data to help inform descriptive and documentary analysis of the UK LLC dataset. This will describe the combined LPS population and make clear the LLC denominator, its characteristics and how this sub-set of the population relates to the wider population (through comparison with aggregated populations statistics). This will inform research users and those (e.g., policy makers) seeking to draw inferences from findings.

2) conduct feasibility assessments to help determine the viability of proposed projects. To do this, UK LLC staff will query the UK LLC integrated dataset (LPS data and linked NHS Records) to determine if UK LLC have sufficient case numbers to be able to inform specific research proposals. This purpose enhances the ethical and governance case for the resource as it means data are not shared where they cannot meaningfully inform the science, it also helps ensure the efficient use of research funding and resources.

DATA REQUEST & JUSTIFICATION

UoB propose linking Covid-19 relevant data from the studies listed above to the following datasets managed by NHS Digital. The proposed uses of these data include but are not limited to:

1. Demographics Data i.e., area of residence, neighbourhood socio-economic indicators such as Indices of Multiple Deprivation to help characterise the UK LLC participants and to inform sub-group analysis;

2. Civil Registration (Deaths) Data and Cancer Registration Data to determine Covid-19 mortality and to consider the impact of the pandemic on general rates of mortality and cancer diagnosis and outcomes;

3. GPES Data for Pandemic Planning and Research (COVID-19) to define Covid-19 caseness, Covid-19 symptoms and outcomes (including ‘long COVID’, wider health outcomes (including mental health, substance use, addiction), patterns in health service interactions, uptake of Covid-19 vaccine and vaccine behaviours in general, to define pre Covid-19 health status, case ascertainment across non Covid-19 outcomes and multi-morbidities;

4. Hospital Episode Statistics (Accident and Emergency; Critical Care; Admitted Patient Care; Outpatients; Emergency Care Data Set) for reasons stated under GPES.

5. Mental Health Services DataSet and Improving Access to Psychological Therapies DataSet to consider changes in mental health outcomes and to define changes in help seeking behaviours and health care interactions;

6. COVID-19 Testing Data to determine testing, caseness and Covid-19 outcomes;

7. NHS 111 records to define Covid-19 symptoms and caseness, wider health symptoms (including mental health status) and consider changing patterns in help seeking behaviours and health care interactions;

8. Community Services Data Set to consider changing patterns in help seeking behaviours and health care interactions;

9. The Shielded patient list to conduct sub-group analysis and to investigate outcomes and behaviours specific to this group and people co-habiting with this group;

10. Vaccine and adverse reaction datasets (as and when they become available via NHS Digital) to inform uptake of Covid-19 vaccine, patterns in uptake, and vaccine behaviour in sub-groups.

The UK LLC will separately apply to the relevant data controllers to link to other datasets not available through NSH Digital. These other datasets will be integrated with study and NHS Digital data where necessary for specific research projects. This will include NHS records from other NHS authorities (i.e., Scotland, Wales, Northern Ireland) and other health databases (e.g. ICNARC, NICOR, SSNAP, Diabetes registers) necessary to measure long-term adverse health outcomes (not requested from NHS Digital). The LHW NCS also considers the intersections between health and social circumstances relating to COVID-19 and the impact of COVID-19 restrictions on health outcomes. To inform these investigations the UK LLC will apply to access and integrate administrative records (e.g. occupation, employment status, benefits provision, education attainment, attendance and school census data) and environmental data related to the space in which people live (e.g. air pollution, greenspace) and the neighbourhood (population density, service provision, broadband facilities).

Through separate approval mechanisms these will be further linked with data from the Office for National Statistics (e.g. the ONS Covid-19 Infection Survey); records from the Zoe Symptom Tracker application; and geo-spatial data modelling environmental exposure estimates (including but not restricted to air pollutants and noise) and indicators describing the natural and built and social environment (including but not restricted to deprivation indices, urban/rural status, neighbourhood characteristics such as housing density, service provision and facilities, green and blue space). It will not be permitted for users to link any additional public data, but they can seek approval for UK LLC staff to do this on their behalf subject to disclosure and other IG risk assessments.

All data will be de-identified (and the UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. Yet for UK LLC staff and research users the risk of re-identification of data within the UK LLC is not reasonably likely.

It is necessary to request the full NHS Digital Covid-19 dataset in terms of years and geographical area of coverage to inform the most accurate assessments of pre-pandemic health status and comprehensive understanding of behaviours with those during the pandemic (including risk factors for Covid-19 and its outcomes) and to be inclusive of all UK LLC participants resident and/or seeking health care services in England. There is no alternative source of this objectively recorded data.

DATA MINIMISATION

The datasets and data items requested have been restricted to only those directly relevant to the Covid-19 research programme that this infrastructure is designed to support. UoB fully adopt the standard NHS Digital definition of the Covid-19 relevant dataset.

As described, the UK LLC C-19 research programme is wide ranging and seeks to understand patterns and predictors of infection, disease outcomes, the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. Datasets requested include measure of Covid-19 related outcomes and pre-Covid-19 baseline datasets (allowing the value of these longitudinal data to be maximised).

The linked data requested are minimised to include only data covering the period each contributing study was in follow-up. The data requested are further minimised to include only records of participants of studies contributing to the UK LLC and those who have objected to this use of their data are excluded. Information on participants who have ‘opted-out’ will be uploaded to the UK LLC quarterly so that those participants’ data will be excluded from future studies.

UoB are unable to minimise datasets further based on fields or

**This is an extension to permit University of Bristol to continue to hold and process already disseminated data whilst University of Bristol they address DARS and IGARD feedback. No new data will flow under v1.**

The UK LLC is likely to generate diverse outputs, with these forming a mix of rapid insight reports to policy makers and substantive research reports. The NCS is specifically designed to establish a clear channel between decision makers and researchers and the UK LLC is designed to enable rapid response and up-to-date information needed to support this. The likely outputs include:

* Peer-reviewed scientific publication (and pre-print publications of initial drafts prior to review which are designed enable rapid awareness of the findings), first publications are expected during 2021;

* Conference presentations and workshop presentations (to academic, health service practitioners and planners, government policy makers, Scientific Advisory Group for Emergencies (SAGE) members), first outputs expected within the first 6 months;

* Reports to health service practitioners and planners, government policy makers and Committee’s (e.g., National Institute for Health and Care Excellence (NICE), SAGE), these are expected during 2021

* Rapid evidence synthesis reports as requested by health and government policy makers (through the entirety of the 3-year NCS programme);

* Lay summaries provided through websites, Blogs, social media posts, privacy notices and outputs designed to promote the transparent use of data and the wider research process (anticipated within the first 6 months).

All outputs from the TRE will be assessed for disclosure risk and will be anonymous population level findings. The framework for this is based on leading 'five safes’ approach and is closely adapted from equivalent processes accredited to standards of the Office for National Statistics and the UK Statistics Authority.

All ‘meta’ products of the UK LLC research (such as code list definitions, syntax, workings for derived variables) will be made available for reuse and interrogation/replication. Access to the UK LLC is ‘open’ in the sense that any legitimate researcher conducting Covid-19 research can apply for access using a transparent and consistent decision-making process. As such, the UK LLC enhanced resource should be considered as an important ‘output’ and a measurable benefit of the UK LLC/LH&W NCS.

The high profile of Covid-19 means that many findings may have media interest. Importantly, the UK LLC is establishing a Patient/Participant Involvement and Engagement strategy within its communication work-package package with a dedicated budget and experienced Communications PPIE Officer. The package will build on existing and established relationships with participants to involve public/participants in the co-design of mechanisms to engage and effectively communicate UK LLC findings more widely with the public. UoB are committed to public involvement in our design and communications. UoB have already included HDRUK PPIE panel members (public representatives) in a consultation exercise and will draw on the extensive network of participants to further explore and co-develop dissemination mechanisms, in developing the end-to-end animation in the first instance. UoB are also developing with the studies opportunities for a subgroup of participants to be involved, in the data access decision making process for example. UoB have ensured that the programme of work is adequately resourced to compensate the public for their time, this has been an integral part of the project planning.

The UK LLC is a novel research infrastructure. The creation of this for Covid-19 research may provide methodological insights and illustrations of a new way of working for linkage in longitudinal research. Insights from this will be provided to the longitudinal community and its funders/key stakeholders (through academic papers, presentations, and contributions to advisory groups).

For clarification regarding the distinct staff groups noted in the agreement:

Study staff = staff from the contributing cohorts;

University of Bristol (UoB) staff = staff of UK Longitudinal Linkage Collaboration (LLC) employed by UoB;

UK LLC staff = staff employed by UK LLC (UoB and University of Edinburgh (UoE) staff).

The LLC project itself is cross-institutional and has staff based at University of Bristol and the University of Edinburgh. The University of Edinburgh are collaborators however they have no role in determining how the data will be processed nor will they have any access to the record level data.

BACKGROUND

The UK Chief Scientific Advisor has established (October 2020) a programme of National Core Studies (NCS) for SARS-CoV-2 (Covid-19) research as a coordinated, long-term, national research initiative. This will consider Covid-19 in terms of a viral pandemic (including issues of cases, transmission, symptoms, and outcomes) and in terms of the health and social impacts of behavioural restrictions designed to mitigate the harms of the pandemic.

The NCS has six different sub-programmes which are addressing major Covid-19 research areas; one of these is the Longitudinal Health and Well-being (LH&W) NCS which is designed to use data from longitudinal studies to address the impact of Covid-19 and of associated viral suppression measures on health and well-being. The LLC is the central hub component of the Longitudinal Health and Well-being NCS.

The UK LLC has been designed to underpin the LH&W NCS, although not exclusively as the NCS are designed to support each other, where for example: the LLC could form the infrastructure for the long-term patient follow-up of consenting trial participants (e.g. participants in the RECOVERY trial) or, where UK longitudinal population studies (LPS) are being used to collect specific new study data which is not available through routine records (e.g. the ALSPAC birth cohort is providing a frame to collect biological samples to inform the Immunology NCS research programme; the assayed results of which will be uploaded into the LLC). For this reason, users of the LLC may come from across the full range of NCS studies and the resource will be accessible to other legitimate UK-based researchers investigating Covid-19 through a sub-licence framework. The NCS is planned to be a two/three-year research programme commencing October 2020.

THE UK LLC & DATA SUBJECTS

UoB (for LLC purposes) will underpin the NCS programme by combining, within a Trusted Research Environment, study data from >15 major inter-disciplinary UK LPS with a combined total of 1 – 2 million UK participants, centrally link to a wide range of Covid-19 relevant and non-health administrative records to utilise alongside the UK Biobank, the Zoe Symptom Tracker cohort and other sources. The integrated data, infrastructure and accompanying governance aspects will collectively be known as the UK LLC.

The de-identified data from contributing studies will be used to establish a UK LLC participant and household register, this will be used within the UK LLC to index and catalogue available data. Researchers will be able request access to the integrated data via the Health Data Research UK Gateway (https://www.healthdatagateway.org/) which through the NCS programme is being established as the de facto application point for UK Data Science projects and as a systematic and coordinated infrastructure for assessing the legitimacy and ‘safe’ credentials of a project. The value of these LPS data assets will be significantly enhanced through linkage to Covid-19 relevant health and other routine records (e.g., employment status, occupation, and benefit receipt records). This will allow the use of objectively assessed diagnoses and other records and the collection of data not amenable to self-report (such as detailed records of care, or health service utilisation). It will also allow researchers to assess and quantify/document bias and error through triangulating data collected across different organisations and settings.

Throughout the development of the UK LLC, UoB have worked collaboratively with UK LPS to co-design and develop the protocol and methodology. One of the strengths of the collaborative approach championed by the UK LLC is that UK LPS are committed and enthusiastic partners in this work: this initiative comes from within the community. The UK LLC will provide infrastructure to support the LPS and researchers working within the UK to investigate high priority COVID-19 research questions. UK LLC/UoB will continue to seek input from the LPS into decision making, guidance in future development and operation. Crucially, the LPS’s all maintain and manage ‘their’ relationships with ‘their’ participants including making critical decisions about data use, and communicating with participants about the existence of, and work of, the UK LLC and providing reassurance that this does not alter the participants existing relationship.

ROLE OF NHS DIGITAL

COHORTS TO BE INCLUDED

The contributing studies whose data will be linked to NHS Digital data in this application are:

1) The National Study of Health and Development (NSHD); which is owned by University College London (UCL; sole Data Controller). UCL have an existing Data Sharing Framework Contract with NHS Digital and NSHD have already flagged its cohort at NHS Digital (MR1) and has active data sharing agreements for the MR1 study (Ref: DARS-NIC-148100-6RFK9 (application covered by consent) and DARS-NIC-86954-Y0R2N (s251)). NSHD brings value as a nationally representative ageing cohort (age ~74) with lifelong follow-up and relevant Covid-19 data collections.

2) The Southall And Brent REvisited (SABRE); which is owned by University College London (UCL; sole Data Controller). UCL have an existing Data Sharing Framework Contract with NHS Digital. SABRE have not flagged its cohort members at NHS Digital (meaning a new study flagging will be established on behalf of SABRE through this application) and do not have an active NHS Digital data sharing agreement. SABRE brings value as a tri-ethnic cohort with considerable participant heterogeneity. It also characterises a relatively deprived urban area of the UK.

3) Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”); which is owned by University of Bristol (UoB; sole Data Controller). ALSPAC has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (a DARS approval for consenting participants Ref: NIC 13133-B7B3K; and a DARS approval for s251 participants: NIC 152414-W3P6Q). ALSPAC brings value as a multi-generational cohort: our original parents (Generation 0, G0), our index participants born in the early 1990s (Generation 1, G1) and now their children (Generation 2, G2) with relevant Covid-19 data collections.

4) TwinsUK; which is owned by Kings College London (KCL; sole Data Controller). TwinsUK has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (a DARS approval for consenting participants Ref: NHS Digital. DARS-NIC-147955-M8D2Q-v1.3. undergoing amendment for HES, demographics, MRIS, Flagging status. TwinsUK have passed DSPT 19/20 (Details of ISO 27001)). TwinsUK brings value as largest adult twin registry in the UK and the most clinically detailed in the world.

5) Twins early Development Study (TEDS); which is owned by Kings College London (KCL; sole Data Controller). TEDS do not have existing NHS Digital DARS approval. TEDS brings value through collection of genetic and environmental information through Twins early development to adolescence.

6) Genetic Links to Anxiety and Depression Study (GLAD); which is owned by Kings College London (KCL; sole Data Controller). GLAD do not have existing NHS Digital DARS approval. GLAD brings value through coverage of a population over the four nations and collecting information around risk factors for depression and anxiety.

7) English Longitudinal Study of Ageing (ELSA); which is owned by University College London (UCL; sole Data Controller). ELSA has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme across the three data controllers’ organisations: IFS: DARS-NIC-32854-Y8P8B-v2.2, UCL: DARS-NIC-30493-Y0C0K-v1.4, NatCen: DARS-NIC-32854-Y8P8B-v2.1. ELSA brings value as a national ageing cohort with regular follow-up and detailed information on physical and mental health as well as finances and attitudes to ageing.

8) 1958 National Child Development Study (NCDS); which is owned by the University College London Centre for Longitudinal Studies (CLS) (UCL; sole Data Controller). NCDS has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-49297-Q7G1Q). 1958 NDS brings value as a national birth cohort with information relating to lifelong factors influencing health and wellbeing development as well as economic inequalities, and now how these relate to retirement and ageing.

9) 1970 British Cohort Study (BCS 70); which is owned by the University College London Centre for Longitudinal Studies (CLS) (UCL; sole Data Controller). BCS 70 has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-49826-T0J7C). The 1970 British Cohort Study brings value as a birth cohort providing insight, amongst other things into education and social development, economic circumstances in a population of now 50+ year olds.

10) Next Steps; which is owned by the University College London Centre for Longitudinal Studies (CLS) (UCL; sole Data Controller). Next Steps has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS NIC-51342-V1M5W) for the extraction and use of linked NHS records in its research programme. Next Steps brings value with detailed information on academic performance through links to the National Pupil Database records and planned links to higher education and Universities, the cohort comprises people born in England in 1989-1990.

11) The Millenium Cohort Study (MCS); which is owned by the University College London Centre for Longitudinal Studies (CLS) (UCL; sole Data Controller). MCS has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-384504-N2V5B). The Millenium Cohort Study brings value through following the lives of 20 - 22 year olds and their residential parents, providing insight into family life and relationships.

12) INTERVAL; which is owned by the University of Cambridge (UoC; sole Data Controller). INTERVAL has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-156334-711SX). INTERVAL provides value through their representation of blood donors who have frequently given blood to investigate impacts on wellbeing.

13) COMPARE; which is owned by the University of Cambridge (UoC; sole Data Controller). COMPARE has an existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (DARS-NIC-156334-711SX). COMPARE provides value through information on blood donors haemoglobin.

14) STRIDES BioResouce; which is owned by the University of Cambridge (UoC; sole Data Controller). STRIDES do not have existing NHS Digital DARS approval. STRIDES provides value as a research blood donor resource and panel of volunteers willing to participate in medical and health related studies.

15) Track-COVID; which is owned by the University of Cambridge (UoC; sole Data Controller). Track-COVID do not have existing NHS Digital DARS approval. Track-COVID provides value with a subset of INTERVAL COMPARE and STRIDES bioresource to determine risk factors for infection.

16) NIHR (National Institute for Health Research) BioResouce; which is owned by the Cambridge University Hospitals NHS Foundation Trust (sole Data Controller). NIHR BioResource provides value through holding unique genetic information including eating and neurological disorders.

17) Extended Cohort for E-health, environment, and DNA (EXCEED); which is owned by the University of Leicester (UoL; sole Data Controller). EXCEED do not have existing NHS Digital DARS approval. EXCEED provides value through collecting information on environment and DNA to provide insight into environmental influences on long-term health.

18) Understanding Society; which is owned by the University of Essex (UoE; sole Data Controller). Understanding Society do not have an existing NHS Digital DARS approval. Understanding Society provides value as a household study providing a generational perspective and changes over the long-term, particularly social, and economic factors.

19) Born in Bradford; which is owned by Bradford Teaching Hospitals NHS Foundation Trust (BTHFT; sole Data Controller). Born in Bradford do not have an existing NHS Digital DARS approval. Born in Bradford provides value through representation of a multi-ethnic population and regional cohort from an area which has high levels of deprivation.

20) The National Study of Health and Development (NSHD); which is owned by University College London (UCL; sole Data Controller). UCL have an existing Data Sharing Framework Contract with NHS Digital and NSHD have already flagged its cohort at NHS Digital (MR1) and has active data sharing agreements for the MR1 study (Ref: DARS-NIC-148100-6RFK9 (application covered by consent) and DARS-NIC-86954-Y0R2N (s251)). NSHD brings value as a nationally representative ageing cohort (age ~74) with lifelong follow-up and relevant Covid-19 data collections.

21) The Southall And Brent REvisited (SABRE); which is owned by University College London (UCL; sole Data Controller). UCL have an existing Data Sharing Framework Contract with NHS Digital. SABRE have not flagged its cohort members at NHS Digital (meaning a new study flagging will be established on behalf of SABRE through this application) and do not have an active NHS Digital data sharing agreement. SABRE brings value as a tri-ethnic cohort with considerable participant heterogeneity. It also characterises a relatively deprived urban area of the UK.

22) Avon Longitudinal Study of Parents and Children (ALSPAC aka “Children of the 90s”); which is owned by University of Bristol (UoB; sole Data Controller). ALSPAC has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (a DARS approval for consenting participants Ref: NIC 13133-B7B3K; and a DARS approval for s251 participants: NIC 152414-W3P6Q). ALSPAC brings value as a multi-generational cohort: our original parents (Generation 0, G0), our index participants born in the early 1990s (Generation 1, G1) and now their children (Generation 2, G2) with relevant Covid-19 data collections.

23) TwinsUK; which is owned by Kings College London (KCL; sole Data Controller). TwinsUK has existing NHS Digital DARS approval for the extraction and use of linked NHS records in its research programme (a DARS approval for consenting participants Ref: NHS Digital. DARS-NIC-147955-M8D2Q-v1.3. undergoing amendment for HES, demographics, MRIS, Flagging current status. TwinsUK have passed DSPT 19/20 (Details of ISO 27001)). TwinsUK brings value as largest adult twin registry in the UK and the most clinically detailed in the world. Only data covering participants in England will be included.

There is a longer-term objective to include a larger number of studies into the UK LLC to boost statistical power, increase participant heterogeneity and sample size in population and outcome sub-groups and to increase the range of self-reported information. Any additional study would be included as an amendment to this application. Additional studies not included in this agreement will be added as an amendment to this agreement.

The UoB scientific programme requires the extraction and use of the NHS Digital data compiled with Covid-19 relevant datasets (primary care, secondary care, community, mental health care provision, Covid-19 testing and outcomes data, NHS service use interactions such as NHS 111 records, mortality, disease registry and demographics data) for the purpose of establishing the UK LLC.

The UK LLC has conducted an assessment as to whether the data held and processed within the UK LLC is identifiable/potentially identifiable and therefore Personal Data which is subject to the Data Protection Act 2018 (DPA), UK General Data Protection Regulations (UK GDPR) and Common Law Duty of Confidentiality (Common Law) in accordance with guidance issued by the Information Commissioner’s Office. The assessment determined that the infrastructure, incorporating all data flows, should be considered as containing Personal Data and being subject to the DPA/UK GDPR and Common Law requirements. However, within this the assessment identified that the UK LLC sits at the threshold of being considered as holding pseudonymous data (Personal Data) and de-identified data where the risk of re-identification is not reasonably likely (not Personal Data), and that this differing status manifests when considering the data from the perspective of the different high-level groups of UK LLC data users and their distinct processing activities. It is evident that for the contributing studies, the use of data is either identifiable (for the provision of identifiers for linkage purposes) or pseudonymous (for the provision of data into the UK LLC and the processing of study data by study staff in their demarcated study processing area) and is therefore Personal Data. Yet, for the UK LLC staff and researcher users, the risk of re-identification of data within the UK LLC is not reasonably likely. This status, for these users, has been achieved through the deployment of data level as well as contextual controls which mitigate the risk of identification. The principal amongst these is the UK LLC operates as a ‘Trusted Research Environment’ which ensures that minimised data are used by minimised number of legitimate users within a secure environment for approved and proportionate purposes.

To contribute data into the UK LLC, LPS’s will be required to flow identifiers to the NHS to establish/filter a linkage between each participants' study record and their NHS registry record(s). A reasonable expectation will have been set through establishing fair processing for the use of NHS records via the UK LLCs infrastructure and for the sharing of these for public benefit research. Studies will update their privacy notices and notify participants of their involvement with UK LLC, explaining what this means and provide options to ‘opt-out’, in line with existing policy and practice.

UK LLC TRUSTED RESEARCH ENVIRONMENT (TRE)

The LLC will be built on a UK Secure eResearch Platform (UKSeRP). This is a well-established infrastructure for data science: it has ISO27001 and NHS Data Security Protection toolkit and Office for National Statistics accreditation under the Digital Economy Act 2017. Other instances of this environment are already used to host LPS and NHS Digital data and have been audited and commended as being well suited to the purpose and low risk (e.g. https://digital.nhs.uk/binaries/content/assets/website-assets/services/dars/data-sharing-post-audit-review_university-of-bristol-.pdf).

The UK LLC TRE will be owned, operated, and managed by the LLC team (data management based at UoB). Under instruction from UoB, contributing LPS and data owners including NHS Digital will provide data to the UK LLC TRE using the ‘split file’ approach which is used across UKSeRP. This split file approach (details below) ensures the physical and processing operational separation of personal identifiers and participants de-identified study data and health records. This separation means that no party involved in this process can see both identifiers and data. This split file approach is coupled with the technological and socio-governance controls applied at the UK LLC (e.g., training, contracts, policy and procedures, penalties for misuse). The UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. For UK LLC staff and research users the risk of re-identification is controlled to the point that it is not reasonably likely.

UK LLC C-19 RESEARCH PROGRAMME

The UK LLC is specifically commissioned as a database resource to inform the LH&W NCS but is also designed to be a resource for the investigation of other Covid-19 questions as these are identified and prioritised through the Scientific Advisory Group for Emergencies (see the Health Data Research UK and SAGE research prioritisation funnel - current versions available through the HDRUK SAGE report, see - https://www.hdruk.ac.uk/covid-19/).

Given these questions are likely to change over time (e.g., the identification of ‘Long COVID’ as a potential new set of complex outcomes is introducing important new research questions), the research programme the LLC is designed to inform can be defined as the investigation of Covid-19 questions using data from longitudinal studies linked to routine records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing. For the use of linked longitudinal studies and NHS records this definition is refined to make clear the investigations primary outcomes must be focused on generating benefits to the health and social care system.

The investigation of Covid-19 questions using data from longitudinal studies linked to routine NHS records to address the impact of Covid-19 and of associated viral suppression measures on health and wellbeing: where the primary research outcome is designed to benefit the health and social care system. These linked data will be of unique value in underpinning a programme of research on Covid-19 informed by the data assets within the UK LLC.

This program will seek to understand the patterns and predictors of infection, (including re-infection) and disease outcomes (such as ‘long covid’), and the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. It will also explore population level changes to physical and mental health, including hospital admission and mortality, in association with viral suppression measures and how these relate to changes in health behaviours.

Given the impact of the Covid-19 containment measures, it is important to investigate the role of socio-economic and neighbourhood/environmental factors in determining population level impacts to physical and mental health, to identify both groups at risk, and factors which offer resilience to adverse outcomes. Within this, UK LLC users will analyse changes in health care service use (using self-reported data and NHS health records) to determine if patterns in these have changed during the Covid-19 pandemic. Linked NHS records will also inform consideration of service use in relation to pre-pandemic health status and regular service interactions (e.g., screening, health reviews and routine service take up (such as annual seasonal flu vaccinations)).

This programme, through the LH&W NCS will specifically seek to answer the following questions:

i) What are the mental health consequences of Covid-19 infection, and of viral suppression measures? How do these differ by key demographics, socioeconomic status and by prior mental health status?

ii) What are the risks, determinants, and expressions of the physical health consequences of long-Covid?

iii) What are the disruptions to health care services (primary and secondary), and what are the health consequences of these disruptions? How do they vary by geography, demographics, and socioeconomic status?

iv) What are the risks of re-infection?

v) How do measures of historical health & physiology impact risk of C-19 infection and its consequences?

vi) How do we best identify cases in population studies, triangulating symptoms, antigen, and antibody testing? How do antibody profiles differ by case status (e.g., those with and without symptoms), how do antibody profiles vary over time? What are the health determinants of such variability?

vii) What are the long-term socioeconomic consequences of viral suppression measures? How do these in turn impact on health and on health care utilisation?

These research themes have been scientifically reviewed by the UK Chief Scientific Advisor and the UK Chief Medical Officer and are high priority questions of critical national interest.

OTHER DATA PROCESSING PURPOSES

In addition to the primary research themes, UoB will:

1) use linked NHS data to help inform descriptive and documentary analysis of the UK LLC dataset. This will describe the combined LPS population and make clear the LLC denominator, its characteristics and how this sub-set of the population relates to the wider population (through comparison with aggregated populations statistics). This will inform research users and those (e.g., policy makers) seeking to draw inferences from findings.

2) conduct feasibility assessments to help determine the viability of proposed projects. To do this, UK LLC staff will query the UK LLC integrated dataset (LPS data and linked NHS Records) to determine if UK LLC have sufficient case numbers to be able to inform specific research proposals. This purpose enhances the ethical and governance case for the resource as it means data are not shared where they cannot meaningfully inform the science, it also helps ensure the efficient use of research funding and resources.

DATA REQUEST & JUSTIFICATION

UoB propose linking Covid-19 relevant data from the studies listed above to the following datasets managed by NHS Digital. The proposed uses of these data include but are not limited to:

1. Demographics Data i.e., area of residence, neighbourhood socio-economic indicators such as Indices of Multiple Deprivation to help characterise the UK LLC participants and to inform sub-group analysis;

2. Civil Registration (Deaths) Data and Cancer Registration Data to determine Covid-19 mortality and to consider the impact of the pandemic on general rates of mortality and cancer diagnosis and outcomes;

3. GPES Data for Pandemic Planning and Research (COVID-19) to define Covid-19 caseness, Covid-19 symptoms and outcomes (including ‘long COVID’, wider health outcomes (including mental health, substance use, addiction), patterns in health service interactions, uptake of Covid-19 vaccine and vaccine behaviours in general, to define pre Covid-19 health status, case ascertainment across non Covid-19 outcomes and multi-morbidities;

4. Hospital Episode Statistics (Accident and Emergency; Critical Care; Admitted Patient Care; Outpatients; Emergency Care Data Set) for reasons stated under GPES.

5. Mental Health Services DataSet and Improving Access to Psychological Therapies DataSet to consider changes in mental health outcomes and to define changes in help seeking behaviours and health care interactions;

6. COVID-19 Testing Data to determine testing, caseness and Covid-19 outcomes;

7. NHS 111 records to define Covid-19 symptoms and caseness, wider health symptoms (including mental health status) and consider changing patterns in help seeking behaviours and health care interactions;

8. Community Services Data Set to consider changing patterns in help seeking behaviours and health care interactions;

9. The Shielded patient list to conduct sub-group analysis and to investigate outcomes and behaviours specific to this group and people co-habiting with this group;

10. Vaccine and adverse reaction datasets (as and when they become available via NHS Digital) to inform uptake of Covid-19 vaccine, patterns in uptake, and vaccine behaviour in sub-groups.

The UK LLC will separately apply to the relevant data controllers to link to other datasets not available through NSH Digital. These other datasets will be integrated with study and NHS Digital data where necessary for specific research projects. This will include NHS records from other NHS authorities (i.e., Scotland, Wales, Northern Ireland) and other health databases (e.g. ICNARC, NICOR, SSNAP, Diabetes registers) necessary to measure long-term adverse health outcomes (not requested from NHS Digital). The LHW NCS also considers the intersections between health and social circumstances relating to COVID-19 and the impact of COVID-19 restrictions on health outcomes. To inform these investigations the UK LLC will apply to access and integrate administrative records (e.g. occupation, employment status, benefits provision, education attainment, attendance and school census data) and environmental data related to the space in which people live (e.g. air pollution, greenspace) and the neighbourhood (population density, service provision, broadband facilities).

Through separate approval mechanisms these will be further linked with data from the Office for National Statistics (e.g. the ONS Covid-19 Infection Survey); records from the Zoe Symptom Tracker application; and geo-spatial data modelling environmental exposure estimates (including but not restricted to air pollutants and noise) and indicators describing the natural and built and social environment (including but not restricted to deprivation indices, urban/rural status, neighbourhood characteristics such as housing density, service provision and facilities, green and blue space). It will not be permitted for users to link any additional public data, but they can seek approval for UK LLC staff to do this on their behalf subject to disclosure and other IG risk assessments.

All data will be de-identified (and the UK LLC is classified as containing pseudonymous data (Personal Data under UK GDPR). This is the case for studies who contribute pseudonymous data into the UK LLC and access data for their participants within the UK LLC. Yet for UK LLC staff and research users the risk of re-identification of data within the UK LLC is not reasonably likely.

It is necessary to request the full NHS Digital Covid-19 dataset in terms of years and geographical area of coverage to inform the most accurate assessments of pre-pandemic health status and comprehensive understanding of behaviours with those during the pandemic (including risk factors for Covid-19 and its outcomes) and to be inclusive of all UK LLC participants resident and/or seeking health care services in England. There is no alternative source of this objectively recorded data.

DATA MINIMISATION

The datasets and data items requested have been restricted to only those directly relevant to the Covid-19 research programme that this infrastructure is designed to support. UoB fully adopt the standard NHS Digital definition of the Covid-19 relevant dataset.

As described, the UK LLC C-19 research programme is wide ranging and seeks to understand patterns and predictors of infection, disease outcomes, the role of antecedent and current health behaviours, health status, medication use, sociodemographic status, built and natural environmental factors, in impacting these outcomes. Datasets requested include measure of Covid-19 related outcomes and pre-Covid-19 baseline datasets (allowing the value of these longitudinal data to be maximised).

The linked data requested are minimised to include only data covering the period each contributing study was in follow-up. The data requested are further minimised to include only records of participants of studies contributing to the UK LLC and those who have objected to this use of their data are excluded. Information on participants who have ‘opted-out’ will be uploaded to the UK LLC quarterly so that those participants’ data will be excluded from future studies.

UoB are unable to minimise datasets further based on fields or episodes at this time as work continues to be conducted to inform the UK’s Covid-19 research programme which aims to be responsive to health and government policy makers. For example in HES, UoB cannot consider only episodes with a specific diagnosis or procedure at this time, UoB do not yet understand the longer-term health implications from Covid-19 infection or whether there exist patterns and predictors of

The UK LLC is likely to generate diverse outputs, with these forming a mix of rapid insight reports to policy makers and substantive research reports. The NCS is specifically designed to establish a clear channel between decision makers and researchers and the UK LLC is designed to enable rapid response and up-to-date information needed to support this. The likely outputs include:

* Peer-reviewed scientific publication (and pre-print publications of initial drafts prior to review which are designed enable rapid awareness of the findings), first publications are expected during 2021;

* Conference presentations and workshop presentations (to academic, health service practitioners and planners, government policy makers, Scientific Advisory Group for Emergencies (SAGE) members), first outputs expected within the first 6 months;

* Reports to health service practitioners and planners, government policy makers and Committee’s (e.g., National Institute for Health and Care Excellence (NICE), SAGE), these are expected during 2021

* Rapid evidence synthesis reports as requested by health and government policy makers (through the entirety of the 3-year NCS programme);

* Lay summaries provided through websites, Blogs, social media posts, privacy notices and outputs designed to promote the transparent use of data and the wider research process (anticipated within the first 6 months).

All outputs from the TRE will be assessed for disclosure risk and will be anonymous population level findings. The framework for this is based on leading 'five safes’ approach and is closely adapted from equivalent processes accredited to standards of the Office for National Statistics and the UK Statistics Authority.

All ‘meta’ products of the UK LLC research (such as code list definitions, syntax, workings for derived variables) will be made available for reuse and interrogation/replication. Access to the UK LLC is ‘open’ in the sense that any legitimate researcher conducting Covid-19 research can apply for access using a transparent and consistent decision-making process. As such, the UK LLC enhanced resource should be considered as an important ‘output’ and a measurable benefit of the UK LLC/LH&W NCS.

The high profile of Covid-19 means that many findings may have media interest. Importantly, the UK LLC is establishing a Patient/Participant Involvement and Engagement strategy within its communication work-package package with a dedicated budget and experienced Communications PPIE Officer. The package will build on existing and established relationships with participants to involve public/participants in the co-design of mechanisms to engage and effectively communicate UK LLC findings more widely with the public. UoB are committed to public involvement in our design and communications. UoB have already included HDRUK PPIE panel members (public representatives) in a consultation exercise and will draw on the extensive network of participants to further explore and co-develop dissemination mechanisms, in developing the end-to-end animation in the first instance. UoB are also developing with the studies opportunities for a subgroup of participants to be involved, in the data access decision making process for example. UoB have ensured that the programme of work is adequately resourced to compensate the public for their time, this has been an integral part of the project planning.

The UK LLC is a novel research infrastructure. The creation of this for Covid-19 research may provide methodological insights and illustrations of a new way of working for linkage in longitudinal research. Insights from this will be provided to the longitudinal community and its funders/key stakeholders (through academic papers, presentations, and contributions to advisory groups).

Yielded Benefits is not a requirement for new applications.