Imperial College London - REACT Data and Connectivity National Core Studies

The Real-Time Assessment of COVID Transmission (REACT) Study was established in May 2020 and provides monthly estimates of the prevalence of SARS-CoV-2 virus and bi-monthly estimates of the prevalence of antibodies to SARS-CoV-2 virus in the general population of England. Under this iteration of the Agreement a further 638,299 participants have been added to the cohort with the overall cohort totalling 2,774,600.

This overarching purpose of this application is to enhance the existing REACT study research cohorts by linking the study data from participants (who have provided consent) to their health records held by the NHS. This enhanced dataset will then be used to advance understanding of the risks of infection and reinfection with COVID and people’s future health following a COVID infection.

Under version 1 of this Agreement the following amendments have been requested:

• The DSA has been updated to reflect that the duration of the project has been extended from 12 months to 3 years

• The study is now requesting to receive historic data dating back from 2014/15, previously the study only required historic data dating back to one year before entry to the cohort. Data is being requested from before entry to the cohort (one year, from 1st March 2015). The rationale for “looking back” over the 5 years before the pandemic is to establish a “baseline” health status for individuals before people were infected with SARS-CoV-2 and before disruptions to healthcare may have introduced bias into data capture.

• The study cohort is being expanded to include adult participants recruited between June 2021 and February 2022 (Rounds 11-19 of REACT I).

• The Agreement has been updated to reflect the request that Students and Honorary Contract holders be permitted to process NHS England data on behalf of ICL.

• The Agreement has been updated to reflect the request to receive HES Critical Care data for the entire cohort.

Department for Health and Social Care (DHSC) and Imperial College London are joint data controllers of the REACT study. IPSOS MORI is a data processor in the REACT study but will not participate in data processing for the data requested in this application- as such IPSOS are not listed as a data processor. This application will allow completion of the research study which is funded by HDR UK Data and Connectivity funding scheme for National Core Studies. Imperial College London will be the only data processor on this research study.

Aim and purpose:

The aim of this application is to achieve linkage between the participants in the REACT study to information held by the NHS about their health, utilisation of healthcare, COVID antigen and antibody testing and COVID vaccination status. To achieve this, Imperial College will provide NHS Numbers and agreed demographic identifiers plus a study identification number (for participants who have provided consent for data linkage) to NHS England to provide pseudonymised data with Study ID rather than patient identifiers in the dataset. This work has an overall aim of measuring the prevalence of COVID-19 and to understanding its impact on health outcomes (for example influenza, heart and lung disease, depression or anxiety, vaccination).

The REACT study has two arms. REACT-1 provides near-real time estimates of the prevalence of SARS-CoV-2 virus in the general population of England (aged 5 and over) using a self- or parent-administered throat and nose swab tested by reverse transcription polymerase chain reaction (RT-PCR). Around 150,000 people have been recruited into the REACT-1 cohort each month since May 2020 and the most recent findings are from Round 19 in March 2022. REACT-2 provides estimates of the prevalence of immune globulin (IgG) antibodies against SARS-CoV-2 spike protein using a self-administered lateral flow immunoassay (LFIA). Around 150,00 people are recruited during a two-week window every 2 months since June 2020 and the most recent findings from the Round 6 in May 2021. To date, there are over 2.7 million people in the REACT cohort and over 90% have provided consent for linkage of their study data to records held by NHS England.

This enriched cohort will be used for research to deliver urgent information for healthcare planning on individual-risk of COVID-19 and medium-term health outcomes of COVID-19 across the spectrum from asymptomatic to severe disease. The overarching research question is “What are the biological, social and environmental drivers of medium-term health outcomes following infection with SARS-CoV-2?”. This agreement is for the follow up data extract after an initial extract in December 2021. This will update the study cohort with new participants since June 2021; add an additional dataset from HES Critical Care and allow the Study Team to update preliminary analysis with findings from a larger dataset with a longer follow-up period. This study was initially funded by ONS HDR UK Data and Connectivity National Core Study until 31st August 2021. Current work is being funded by a range of sources including infrastructure support to the Department of Epidemiology and Biostatistics at Imperial College London from the NIHR Imperial Biomedical Research Centre; HDR UK London, Public Health Theme- with the exception of ICL, none of these organisations process the data, or play a role in determining the purpose and means of processing.

The IPSOS MORI/Imperial REACT I Antigen study latest Data Sharing Agreement is DARS-NIC-393650-B7J6 and the IPSOS MORI/Imperial REACT II Antibody Study latest Data Sharing Agreement is DARS-NIC-389914-N9R8. These studies are under the data controllership of the Department of Health and Social Care and Imperial College London with IPSOS MORI as a data processor.

This study is relying on Article 6(1)(e) processing is necessary for the performance of a task in the public interest, the data controller Imperial College London are a public authority carrying out the research in the public interest. Article 9(2)(j) (processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes) the data are required for research purposes in the public interest, the research on the biological, social and environmental drivers of medium-term health outcomes following infection with SARS-CoV-2 is of public health importance. The study is performing research on a disease of high public health importance and the data request is to support a high-profile National Core Study.

In accordance with GDPR Article 89(1) processing is subject to appropriate safeguards. These include:

ii. The data recipient’s technical and organisational measures to safeguard the data have been assessed and meet NHS England’s acceptance criteria

iii. The requested data has been assessed as proportionate to the aim pursued

iv. Controls, data retention and processing activities have been assessed to ensure respect to the essence of the right to data protection

v. Measures to protect the rights and freedoms of data subjects have been assessed including transparency (fair processing) publishing subject’s rights to withdraw consent and/or have their data erased or rectified, etc.

Data requested:

Data is requested at the individual level from the following datasets:

1. COVID-19 Hospitalization in England Surveillance System: this dataset has national coverage of clinical and administrative data collected from people during a hospital admission with COVID-19. Its purpose is for surveillance of secondary care cases of COVID-19. It was established on 15th March 2020 adapted from the system in pace for Influenza surveillance and provides a core dataset that enhances Hospital Episode Statistics through the inclusion of detailed, coded clinical information relevant to COVID-19 disease risk and outcomes. It will be used to provide information about the clinical status of people within the cohort who are treated in hospital for COVID-19.

2. COVID-19 Second Generation Surveillance System: this dataset is the national reporting system for routine laboratory testing for SARS-CoV-2, it builds on the existing SGSS for laboratories. Its purpose is for surveillance of confirmed cases of SARS-CoV-2 and testing began on 24th January 2020. It contains positive test results from Pillar 1 and Pillar 2 antigen testing. It will be used to define the COVID exposure status of participants in the cohort by providing dates when people had a positive test.

3. GPES Data for Pandemic Planning and Research (COVID-19): this dataset has been established to support research and planning to inform the UK pandemic response. All GP practices in England are required to submit a defined, coded dataset every two weeks and the national data-opt-out may not apply. Collection commenced in May 2020. It will be used to provide information about the clinical status of people within the cohort following a diagnosis of COVID. People who do not get diagnosed with COVID will also be followed up to allow comparison between health and healthcare in those who do and do not have episodes of infection.

4. Covid-19 UK Nonhospital Antibody Testing Results (Pillar 3): This dataset is national (UK-wide) surveillance of serology testing, both positive and negative test results. It draws in data from multiple sources. It will be used to define the COVID exposure status of participants in the cohort by providing information about whether people have antibodies or not. It will be specifically interpreted in the context of (6).

5. Covid-19 UK Nonhospital Antigen Testing Results (pillar 2): this dataset is surveillance of the national swab testing for the general population (excluding patients in hospital; health and social care workers; test processed by PHE or hospital laboratories) undertaken in venues such as at home, drive through or mobile test centres. It will be used in combination with (2) to define the COVID exposure status of participants in the cohort by providing dates when people had a positive test.

6. COVID-19 Vaccination Status Data: this dataset has recently been established and contains records of COVID vaccination events in any setting. It will be used to identify the dates when participants in the cohort received doses of a COVID vaccine. People will be followed from this point onwards and their post-vaccine COVID exposure status as obtained from (2); (4) and (5) will be characterised and explored.

7. COVID Vaccination adverse reactions: this dataset is related to dataset (6) and provides information about adverse reactions that occur immediately following the administration of a COVID vaccine.

8. Hospital Episode Statistics Outpatients, Admitted Patient Care and Critical Care: these datasets contain a complete record of all episodes of hospital healthcare in English NHS Hospitals. They will be used to provide information about the clinical status and healthcare utilisation of people within the cohort following a diagnosis of COVID. People who do not get diagnosed with COVID will also be followed up to allow comparison between health and healthcare in those who do and do not have episodes of infection.

9. Civil Registration – Deaths: contains a complete record of all deaths in the UK with information about date, place and cause of death. This dataset will be used to identify death as a health outcome and as the date of leaving the cohort (for time-to-event analyses).

These further data will be linked to the initial data extract and the data from the REACT study and used to construct an extended longitudinal record for the participating members of the REACT cohort that described pre-pandemic health status (since March 2015); COVID-19 testing and results; healthcare utilisation (quantity and reason) since COVID exposure status was determined. It is necessary to request linkage for all consenting participants in the REACT study as this is a representative population-based cohort and it will allow robust comparisons across the longitudinal health profiles of participants with all possible histories of COVID exposure, including across the spectrum from asymptomatic to severe disease. Data is being requested from before entry to the cohort (one year, from 1st March 2015). The rationale for “looking back” over the 5 years before the pandemic is to establish a “baseline” health status for individuals before people were infected with SARS-CoV-2 and before disruptions to healthcare may have introduced bias into data capture. This will allow comparison of health outcomes to be adjusted for pre-existing health and healthcare utilisation. Participants will be followed up into the future for as long as the data allows, the permissions in the overarching REACT Research Ethics Committee (REC) approvals allow for 20 years of follow up. The study protocol permits the study to obtain follow up data until December 2022 and analyse this data until the end of the study , further applications will be submitted should the study choose to extend this follow-up period (with suitable funding in place). It is anticipated that extensions to these data sharing agreements will be requested to permit longer duration of follow up (up to 20 years) subject to appropriate research needs as more is learnt about the long term natural history of COVID-19. Any further data requests would be subject to an amendment to this agreement being approved.

It is necessary to request individual level data as the data will be linked to individual participants in the REACT cohort and analysis will be conducted taking into account variables including age, sex and location of residence (postcode in cohort data). The data that will be in the extract can be pseudonymised through a study ID that can be linked to NHS number in the original cohort and will be provided by Imperial College. Follow up is currently a maximum of 12 months given the timing of the origin of COVID-19 and will extend as time since infection/vaccination increases as the study follows the health outcomes post-infection/vaccination (subject to further DARS applications/extensions and research need). Data is requested on all health outcomes as this research aims to advance understanding of future health status of people characterised by their exposure to COVID-19. As more information becomes available about COVID-19 natural history, this request will be reviewed. The study has selected the fields relevant to this research and minimised the number of identifiable/high-risk fields requested.

Data minimisation has been considered and applied using the following approach (example for HES CC):

1. Datasets: All datasets are relevant to address the study research questions. It is necessary to request individual level data as the analysis will be conducted taking into account variables including age, sex and location of residence (postcode in cohort data). The data that will be in the extract can be pseudonymised through a study ID that can be linked to NHS number in the original cohort.

2. Years: The study requires data from before entry to the cohort (five years) to establish a baseline health status. Follow up is currently a maximum of 2.5 years given the timing of the origin of COVID-19 and will extend as time since infection/vaccination increases as researchers follow the health outcomes post infection/ vaccination.

3. Filtering: This data will be linked to an established cohort (REACT) that is nationally representative of the population of England and contains adults over the age of 18. The health outcomes and healthcare utilisation during and post-covid may vary by age group, ethnicity, geography and are currently not characterised. Therefore, it is necessary to study information about the range of diagnoses and procedures.

4. Episodes: The relationship between COVID and future health has not yet been characterised therefore this epidemiological study requires information of the range of reasons for admission as an inpatient. It has been demonstrated that people have been infected with COVID after admission to hospital therefore it is important to have data on hospital admissions not due to COVID.

5. Fields: The study has selected the fields relevant to this research and no identifiable/high-risk fields have been chosen.

6. Cohorts / Linkages: Researchers are studying the longitudinal health of a cohort of people tested for COVID. It is necessary to follow the health outcomes (all cause) into the future. As more information becomes available about the health outcomes of COVID and the timings of these outcomes it may be possible to revise the data request.

REACT has a strong programme of PPIE led by the Patient Experience Research Centre (PERC) at Imperial College London. The public voice has been included at all stages of the research programme. The REACT public advisory group (PAG) was formed on 28 May 2020, comprised of 11 members of the public from different parts of England and covering different ages, genders, ethnicities and lived experiences [https://www.imperial.ac.uk/medicine/research-and-impact/groups/react-study/react_pag/]. The aim of the REACT PAG was to represent the voice of the general public and the participants who have taken part in the REACT programme to ensure their needs, experiences and feedback inform the design, management and communication of the research and its sub-studies. The PAG meet (initially monthly, then as needed) to share and discuss study updates, challenges, plans and recent findings, new research ideas and funding proposals. The PAG provided strategic input and public perspectives to inform all stages of the research. This proposal was first discussed with the PAG on 16th March 2021. On starting the Health Data Research UK (HDR UK) Data & Connectivity (D&C) REACT Data Linkage project, PERC together with the original REACT PAG (comprised of 7 members) identified that the existing group was missing perspectives from people who had experience of ongoing ill health after having COVID-19 (now commonly known as Long Covid), either for themselves or a close family member; and that there was insufficient gender, patient/carer or young person representation. These voices are crucial to the HDR UK project where we are trying to understand the healthcare use and natural history of people who develop COVID-19 and may experience ongoing ill health as a result. Therefore, PERC recruited 4 more members of the public with lived experience of COVID-19 (and Long Covid) to act REACT Public Advisors on the Data Linkage project and join the REACT PAG if they wish. A further outcome of the initial meeting was to review the consenting process to ensure it's as clear and understandable as possible. A member of the PAG sits on the REACT Data Access Committee. As the REACT Study expanded, the COVID-19 Community Network was formed, open to anyone interested in keeping up to date with REACT.

The processing of this data during the conduct of the proposed research will produce preliminary study findings that will be shared internally with the REACT study team (Imperial College London) and DHSC/UKHSA (Joint Data Controller of REACT study) to achieve timely impact from study insights and findings.

Outputs from the Study aim to include:

• Reports, and academic papers published in leading biomedical and scientific journals

• Presentations to Government, the scientific community and the public

Specific outputs which are to be delivered through this study within a year of receipt of the NHS England data include:

• Natural history of SARS-CoV-2 infection, including asymptomatic infection, including hospitalisations and deaths

• Determining the symptoms, prevalence and effects of Long COVID and effectiveness of vaccination and other interventions

• Identifying the inequalities in long-term outcomes following SARS-CoV-2 infection.

The study has a Public and Patient panel comprised of 4 members of the existing REACT Patient Advisory Group and 4 new members. The Panel will be involved in the production of outputs and the dissemination strategy to ensure that materials are fit for purpose and meet the needs of participants and the wider public. The dissemination strategy aims to link into the existing programme of activity for the wider REACT programme.

All outputs have appropriate statistical disclosure control applied and no identifying information will leave the secure research environment. All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

The REACT Investigators intend to also apply for future funding to maintain and extend the research undertaken on this enhanced research cohort. As part of this activity, a Data Access Committee (consisting of Imperial College London and UK Health Security Agency) is in development to oversee future research direction. The terms of reference for the REACT Data Access Committee will be made publicly available shortly. Membership includes representatives from ICL and UKHSA.

Outputs are yet to be generated because the full linked data from the cohort is not yet available for analysis. The key study output is expected to be produced by September of 2023.

The REal-time Assessment of Community Transmission (REACT) programme was funded by the UK Department of Health and Social Care and has been informing the UK response to the COVID-19 pandemic since its inception in 2020. The programme provided independent, timely and authoritative analyses of the SARS-CoV-2 epidemic in England tracking (REACT-1) the spread of the virus and (REACT-2) development of potentially protective antibodies in real-time.

Measurable Benefits:

• POLICY - the REACT findings were a key source of situational awareness that fed into the UK response to the COVID-19 pandemic, including timing and implementation of non-pharmaceutical interventions (e.g. lockdowns) and vaccination strategy. Key findings were reported directly to the Government Chief Scientist and Chief Medical Officer and data from the study were included in the 10 Downing Street briefings. The weekly reports to government had a direct impact on policy and provided reliable and timely estimates of prevalence, R (reproduction number), identification of variants (genome sequencing) and estimates of vaccine effectiveness

• PUBLIC AWARENESS - the REACT Study has a strong programme of public involvement and engagement. The public were involved in co-designing the study materials to maximise public response and understanding. Study outputs were placed in the public domain in real-time which provided members of the public with key descriptions of the epidemic as it was evolving. The study investigators engaged in public dissemination through written and spoken media to inform the public of the findings as they emerged

• RESEARCH CAPABILITY - the REACT cohort has been made possible because it was widely supported by the residents of England who participated in the study. It is the largest consented cohort of individuals in the UK (~2.7 million people) and a major output of the programme is the cohort "data asset" which we are working to enhance and enrich through this research programme. These data are expected to be used to address key questions, including about the long-term health impacts of COVID-19 (Long COVID)

• Academic Advancement - PhD students and Early Career Researchers are working on the REACT/NHS England data and this programme is supporting their academic development. This DSA specifically includes permissions for Masters and PhD students to analyse the cohort linked NHS England data and this will support further research capacity development.

• Presentations - The work has been widely presented at international and national meetings, both to policy making audiences, the wider academic community and the public.

Outputs from the REACT programme which have contributed to the yielded benefits of this study:

• Riley, S., et al. Science, 2021 May 28; 372(6545):990-995. https://doi.org/10.1126/science.abf0874: first to detect the rise in infections in London, Essex and Kent at the start of the second wave in August-September 2020 (Alpha variant)

• Elliott, P., et al. Science, 2021 Nov 2; 374(6574). https://doi.org/10.1126/science.abl9551: estimated vaccine effectiveness against infection (18 to 64 years) of 49% from June to July 2021.

• Chadeau-Hyam, M., et al. Lancet Resp Med, 2022; 10(4):355-366. https://doi.org/10.1016/s2213-2600(21)00542-7: identified an exponential increase in infections among (largely unvaccinated) children aged 5-17 years in September 2021.

• Elliott, P., et al. Science, 2022 Feb 8; 375(6587):1406-1411. https://doi.org/10.1126/science.abn8347: identified the rapid rise of infections in December 2021 as Omicron (BA.1) replaced Delta.

• Elliott, P., et al. Science, 2022 24 May; 376(660). https://doi.org/10.1126/science.abq4411: identified the rapid replacement of Omicron BA.1 by BA.2 in February and March 2022.

The Real-Time Assessment of COVID Transmission (REACT) Study was established in May 2020 and provides monthly estimates of the prevalence of SARS-CoV-2 virus and bi-monthly estimates of the prevalence of antibodies to SARS-CoV-2 virus in the general population of England. Under this iteration of the Agreement a further 638,299 participants have been added to the cohort with the overall cohort totalling 2,774,600.

This overarching purpose of this application is to enhance the existing REACT study research cohorts by linking the study data from participants (who have provided consent) to their health records held by the NHS. This enhanced dataset will then be used to advance understanding of the risks of infection and reinfection with COVID and people’s future health following a COVID infection.

Under version 1 of this Agreement the following amendments have been requested:

• The DSA has been updated to reflect that the duration of the project has been extended from 12 months to 3 years

• The study is now requesting to receive historic data dating back from 2014/15, previously the study only required historic data dating back to one year before entry to the cohort. Data is being requested from before entry to the cohort (one year, from 1st March 2015). The rationale for “looking back” over the 5 years before the pandemic is to establish a “baseline” health status for individuals before people were infected with SARS-CoV-2 and before disruptions to healthcare may have introduced bias into data capture.

• The study cohort is being expanded to include adult participants recruited between June 2021 and February 2022 (Rounds 11-19 of REACT I).

• The Agreement has been updated to reflect the request that Students and Honorary Contract holders be permitted to process NHS England data on behalf of ICL.

• The Agreement has been updated to reflect the request to receive HES Critical Care data for the entire cohort.

Department for Health and Social Care (DHSC) and Imperial College London are joint data controllers of the REACT study. IPSOS MORI is a data processor in the REACT study but will not participate in data processing for the data requested in this application- as such IPSOS are not listed as a data processor. This application will allow completion of the research study which is funded by HDR UK Data and Connectivity funding scheme for National Core Studies. Imperial College London will be the only data processor on this research study.

Aim and purpose:

The aim of this application is to achieve linkage between the participants in the REACT study to information held by the NHS about their health, utilisation of healthcare, COVID antigen and antibody testing and COVID vaccination status. To achieve this, Imperial College will provide NHS Numbers and agreed demographic identifiers plus a study identification number (for participants who have provided consent for data linkage) to NHS England to provide pseudonymised data with Study ID rather than patient identifiers in the dataset. This work has an overall aim of measuring the prevalence of COVID-19 and to understanding its impact on health outcomes (for example influenza, heart and lung disease, depression or anxiety, vaccination).

The REACT study has two arms. REACT-1 provides near-real time estimates of the prevalence of SARS-CoV-2 virus in the general population of England (aged 5 and over) using a self- or parent-administered throat and nose swab tested by reverse transcription polymerase chain reaction (RT-PCR). Around 150,000 people have been recruited into the REACT-1 cohort each month since May 2020 and the most recent findings are from Round 19 in March 2022. REACT-2 provides estimates of the prevalence of immune globulin (IgG) antibodies against SARS-CoV-2 spike protein using a self-administered lateral flow immunoassay (LFIA). Around 150,00 people are recruited during a two-week window every 2 months since June 2020 and the most recent findings from the Round 6 in May 2021. To date, there are over 2.7 million people in the REACT cohort and over 90% have provided consent for linkage of their study data to records held by NHS England.

This enriched cohort will be used for research to deliver urgent information for healthcare planning on individual-risk of COVID-19 and medium-term health outcomes of COVID-19 across the spectrum from asymptomatic to severe disease. The overarching research question is “What are the biological, social and environmental drivers of medium-term health outcomes following infection with SARS-CoV-2?”. This agreement is for the follow up data extract after an initial extract in December 2021. This will update the study cohort with new participants since June 2021; add an additional dataset from HES Critical Care and allow the Study Team to update preliminary analysis with findings from a larger dataset with a longer follow-up period. This study was initially funded by ONS HDR UK Data and Connectivity National Core Study until 31st August 2021. Current work is being funded by a range of sources including infrastructure support to the Department of Epidemiology and Biostatistics at Imperial College London from the NIHR Imperial Biomedical Research Centre; HDR UK London, Public Health Theme- with the exception of ICL, none of these organisations process the data, or play a role in determining the purpose and means of processing.

The IPSOS MORI/Imperial REACT I Antigen study latest Data Sharing Agreement is DARS-NIC-393650-B7J6 and the IPSOS MORI/Imperial REACT II Antibody Study latest Data Sharing Agreement is DARS-NIC-389914-N9R8. These studies are under the data controllership of the Department of Health and Social Care and Imperial College London with IPSOS MORI as a data processor.

This study is relying on Article 6(1)(e) processing is necessary for the performance of a task in the public interest, the data controller Imperial College London are a public authority carrying out the research in the public interest. Article 9(2)(j) (processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes) the data are required for research purposes in the public interest, the research on the biological, social and environmental drivers of medium-term health outcomes following infection with SARS-CoV-2 is of public health importance. The study is performing research on a disease of high public health importance and the data request is to support a high-profile National Core Study.

In accordance with GDPR Article 89(1) processing is subject to appropriate safeguards. These include:

ii. The data recipient’s technical and organisational measures to safeguard the data have been assessed and meet NHS Digital’s acceptance criteria

iii. The requested data has been assessed as proportionate to the aim pursued

iv. Controls, data retention and processing activities have been assessed to ensure respect to the essence of the right to data protection

v. Measures to protect the rights and freedoms of data subjects have been assessed including transparency (fair processing) publishing subject’s rights to withdraw consent and/or have their data erased or rectified, etc.

Data requested:

Data is requested at the individual level from the following datasets:

1. COVID-19 Hospitalization in England Surveillance System: this dataset has national coverage of clinical and administrative data collected from people during a hospital admission with COVID-19. Its purpose is for surveillance of secondary care cases of COVID-19. It was established on 15th March 2020 adapted from the system in pace for Influenza surveillance and provides a core dataset that enhances Hospital Episode Statistics through the inclusion of detailed, coded clinical information relevant to COVID-19 disease risk and outcomes. It will be used to provide information about the clinical status of people within the cohort who are treated in hospital for COVID-19.

2. COVID-19 Second Generation Surveillance System: this dataset is the national reporting system for routine laboratory testing for SARS-CoV-2, it builds on the existing SGSS for laboratories. Its purpose is for surveillance of confirmed cases of SARS-CoV-2 and testing began on 24th January 2020. It contains positive test results from Pillar 1 and Pillar 2 antigen testing. It will be used to define the COVID exposure status of participants in the cohort by providing dates when people had a positive test.

3. GPES Data for Pandemic Planning and Research (COVID-19): this dataset has been established to support research and planning to inform the UK pandemic response. All GP practices in England are required to submit a defined, coded dataset every two weeks and the national data-opt-out may not apply. Collection commenced in May 2020. It will be used to provide information about the clinical status of people within the cohort following a diagnosis of COVID. People who do not get diagnosed with COVID will also be followed up to allow comparison between health and healthcare in those who do and do not have episodes of infection.

4. Covid-19 UK Nonhospital Antibody Testing Results (Pillar 3): This dataset is national (UK-wide) surveillance of serology testing, both positive and negative test results. It draws in data from multiple sources. It will be used to define the COVID exposure status of participants in the cohort by providing information about whether people have antibodies or not. It will be specifically interpreted in the context of (6).

5. Covid-19 UK Nonhospital Antigen Testing Results (pillar 2): this dataset is surveillance of the national swab testing for the general population (excluding patients in hospital; health and social care workers; test processed by PHE or hospital laboratories) undertaken in venues such as at home, drive through or mobile test centres. It will be used in combination with (2) to define the COVID exposure status of participants in the cohort by providing dates when people had a positive test.

6. COVID-19 Vaccination Status Data: this dataset has recently been established and contains records of COVID vaccination events in any setting. It will be used to identify the dates when participants in the cohort received doses of a COVID vaccine. People will be followed from this point onwards and their post-vaccine COVID exposure status as obtained from (2); (4) and (5) will be characterised and explored.

7. COVID Vaccination adverse reactions: this dataset is related to dataset (6) and provides information about adverse reactions that occur immediately following the administration of a COVID vaccine.

8. Hospital Episode Statistics Outpatients, Admitted Patient Care and Critical Care: these datasets contain a complete record of all episodes of hospital healthcare in English NHS Hospitals. They will be used to provide information about the clinical status and healthcare utilisation of people within the cohort following a diagnosis of COVID. People who do not get diagnosed with COVID will also be followed up to allow comparison between health and healthcare in those who do and do not have episodes of infection.

9. Civil Registration – Deaths: contains a complete record of all deaths in the UK with information about date, place and cause of death. This dataset will be used to identify death as a health outcome and as the date of leaving the cohort (for time-to-event analyses).

These further data will be linked to the initial data extract and the data from the REACT study and used to construct an extended longitudinal record for the participating members of the REACT cohort that described pre-pandemic health status (since March 2015); COVID-19 testing and results; healthcare utilisation (quantity and reason) since COVID exposure status was determined. It is necessary to request linkage for all consenting participants in the REACT study as this is a representative population-based cohort and it will allow robust comparisons across the longitudinal health profiles of participants with all possible histories of COVID exposure, including across the spectrum from asymptomatic to severe disease. Data is being requested from before entry to the cohort (one year, from 1st March 2015). The rationale for “looking back” over the 5 years before the pandemic is to establish a “baseline” health status for individuals before people were infected with SARS-CoV-2 and before disruptions to healthcare may have introduced bias into data capture. This will allow comparison of health outcomes to be adjusted for pre-existing health and healthcare utilisation. Participants will be followed up into the future for as long as the data allows, the permissions in the overarching REACT Research Ethics Committee (REC) approvals allow for 20 years of follow up. The study protocol permits the study to obtain follow up data until December 2022 and analyse this data until the end of the study , further applications will be submitted should the study choose to extend this follow-up period (with suitable funding in place). It is anticipated that extensions to these data sharing agreements will be requested to permit longer duration of follow up (up to 20 years) subject to appropriate research needs as more is learnt about the long term natural history of COVID-19. Any further data requests would be subject to an amendment to this agreement being approved.

It is necessary to request individual level data as the data will be linked to individual participants in the REACT cohort and analysis will be conducted taking into account variables including age, sex and location of residence (postcode in cohort data). The data that will be in the extract can be pseudonymised through a study ID that can be linked to NHS number in the original cohort and will be provided by Imperial College. Follow up is currently a maximum of 12 months given the timing of the origin of COVID-19 and will extend as time since infection/vaccination increases as the study follows the health outcomes post-infection/vaccination (subject to further DARS applications/extensions and research need). Data is requested on all health outcomes as this research aims to advance understanding of future health status of people characterised by their exposure to COVID-19. As more information becomes available about COVID-19 natural history, this request will be reviewed. The study has selected the fields relevant to this research and minimised the number of identifiable/high-risk fields requested.

Data minimisation has been considered and applied using the following approach (example for HES CC):

1. Datasets: All datasets are relevant to address the study research questions. It is necessary to request individual level data as the analysis will be conducted taking into account variables including age, sex and location of residence (postcode in cohort data). The data that will be in the extract can be pseudonymised through a study ID that can be linked to NHS number in the original cohort.

2. Years: The study requires data from before entry to the cohort (five years) to establish a baseline health status. Follow up is currently a maximum of 2.5 years given the timing of the origin of COVID-19 and will extend as time since infection/vaccination increases as researchers follow the health outcomes post infection/ vaccination.

3. Filtering: This data will be linked to an established cohort (REACT) that is nationally representative of the population of England and contains adults over the age of 18. The health outcomes and healthcare utilisation during and post-covid may vary by age group, ethnicity, geography and are currently not characterised. Therefore, it is necessary to study information about the range of diagnoses and procedures.

4. Episodes: The relationship between COVID and future health has not yet been characterised therefore this epidemiological study requires information of the range of reasons for admission as an inpatient. It has been demonstrated that people have been infected with COVID after admission to hospital therefore it is important to have data on hospital admissions not due to COVID.

5. Fields: The study has selected the fields relevant to this research and no identifiable/high-risk fields have been chosen.

6. Cohorts / Linkages: Researchers are studying the longitudinal health of a cohort of people tested for COVID. It is necessary to follow the health outcomes (all cause) into the future. As more information becomes available about the health outcomes of COVID and the timings of these outcomes it may be possible to revise the data request.

REACT has a strong programme of PPIE led by the Patient Experience Research Centre (PERC) at Imperial College London. The public voice has been included at all stages of the research programme. The REACT public advisory group (PAG) was formed on 28 May 2020, comprised of 11 members of the public from different parts of England and covering different ages, genders, ethnicities and lived experiences [https://www.imperial.ac.uk/medicine/research-and-impact/groups/react-study/react_pag/]. The aim of the REACT PAG was to represent the voice of the general public and the participants who have taken part in the REACT programme to ensure their needs, experiences and feedback inform the design, management and communication of the research and its sub-studies. The PAG meet (initially monthly, then as needed) to share and discuss study updates, challenges, plans and recent findings, new research ideas and funding proposals. The PAG provided strategic input and public perspectives to inform all stages of the research. This proposal was first discussed with the PAG on 16th March 2021. On starting the Health Data Research UK (HDR UK) Data & Connectivity (D&C) REACT Data Linkage project, PERC together with the original REACT PAG (comprised of 7 members) identified that the existing group was missing perspectives from people who had experience of ongoing ill health after having COVID-19 (now commonly known as Long Covid), either for themselves or a close family member; and that there was insufficient gender, patient/carer or young person representation. These voices are crucial to the HDR UK project where we are trying to understand the healthcare use and natural history of people who develop COVID-19 and may experience ongoing ill health as a result. Therefore, PERC recruited 4 more members of the public with lived experience of COVID-19 (and Long Covid) to act REACT Public Advisors on the Data Linkage project and join the REACT PAG if they wish. A further outcome of the initial meeting was to review the consenting process to ensure it's as clear and understandable as possible. A member of the PAG sits on the REACT Data Access Committee. As the REACT Study expanded, the COVID-19 Community Network was formed, open to anyone interested in keeping up to date with REACT.

The processing of this data during the conduct of the proposed research will produce preliminary study findings that will be shared internally with the REACT study team (Imperial College London) and DHSC/UKHSA (Joint Data Controller of REACT study) to achieve timely impact from study insights and findings.

Outputs from the Study aim to include:

• Reports, and academic papers published in leading biomedical and scientific journals

• Presentations to Government, the scientific community and the public

Specific outputs which are to be delivered through this study within a year of receipt of the NHS-Digital data include:

• Natural history of SARS-CoV-2 infection, including asymptomatic infection, including hospitalisations and deaths

• Determining the symptoms, prevalence and effects of Long COVID and effectiveness of vaccination and other interventions

• Identifying the inequalities in long-term outcomes following SARS-CoV-2 infection.

The study has a Public and Patient panel comprised of 4 members of the existing REACT Patient Advisory Group and 4 new members. The Panel will be involved in the production of outputs and the dissemination strategy to ensure that materials are fit for purpose and meet the needs of participants and the wider public. The dissemination strategy aims to link into the existing programme of activity for the wider REACT programme.

All outputs have appropriate statistical disclosure control applied and no identifying information will leave the secure research environment. All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

The REACT Investigators intend to also apply for future funding to maintain and extend the research undertaken on this enhanced research cohort. As part of this activity, a Data Access Committee (consisting of Imperial College London and UK Health Security Agency) is in development to oversee future research direction. The terms of reference for the REACT Data Access Committee will be made publicly available shortly. Membership includes representatives from ICL and UKHSA.

Outputs are yet to be generated because the full linked data from the cohort is not yet available for analysis. The key study output is expected to be produced by September of 2023.

The REal-time Assessment of Community Transmission (REACT) programme was funded by the UK Department of Health and Social Care and has been informing the UK response to the COVID-19 pandemic since its inception in 2020. The programme provided independent, timely and authoritative analyses of the SARS-CoV-2 epidemic in England tracking (REACT-1) the spread of the virus and (REACT-2) development of potentially protective antibodies in real-time.

Measurable Benefits:

• POLICY - the REACT findings were a key source of situational awareness that fed into the UK response to the COVID-19 pandemic, including timing and implementation of non-pharmaceutical interventions (e.g. lockdowns) and vaccination strategy. Key findings were reported directly to the Government Chief Scientist and Chief Medical Officer and data from the study were included in the 10 Downing Street briefings. The weekly reports to government had a direct impact on policy and provided reliable and timely estimates of prevalence, R (reproduction number), identification of variants (genome sequencing) and estimates of vaccine effectiveness

• PUBLIC AWARENESS - the REACT Study has a strong programme of public involvement and engagement. The public were involved in co-designing the study materials to maximise public response and understanding. Study outputs were placed in the public domain in real-time which provided members of the public with key descriptions of the epidemic as it was evolving. The study investigators engaged in public dissemination through written and spoken media to inform the public of the findings as they emerged

• RESEARCH CAPABILITY - the REACT cohort has been made possible because it was widely supported by the residents of England who participated in the study. It is the largest consented cohort of individuals in the UK (~2.7 million people) and a major output of the programme is the cohort "data asset" which we are working to enhance and enrich through this research programme. These data are expected to be used to address key questions, including about the long-term health impacts of COVID-19 (Long COVID)

• Academic Advancement - PhD students and Early Career Researchers are working on the REACT/NHS Digital data and this programme is supporting their academic development. This DSA specifically includes permissions for Masters and PhD students to analyse the cohort linked NHS Digital data and this will support further research capacity development.

• Presentations - The work has been widely presented at international and national meetings, both to policy making audiences, the wider academic community and the public.

Outputs from the REACT programme which have contributed to the yielded benefits of this study:

• Riley, S., et al. Science, 2021 May 28; 372(6545):990-995. https://doi.org/10.1126/science.abf0874: first to detect the rise in infections in London, Essex and Kent at the start of the second wave in August-September 2020 (Alpha variant)

• Elliott, P., et al. Science, 2021 Nov 2; 374(6574). https://doi.org/10.1126/science.abl9551: estimated vaccine effectiveness against infection (18 to 64 years) of 49% from June to July 2021.

• Chadeau-Hyam, M., et al. Lancet Resp Med, 2022; 10(4):355-366. https://doi.org/10.1016/s2213-2600(21)00542-7: identified an exponential increase in infections among (largely unvaccinated) children aged 5-17 years in September 2021.

• Elliott, P., et al. Science, 2022 Feb 8; 375(6587):1406-1411. https://doi.org/10.1126/science.abn8347: identified the rapid rise of infections in December 2021 as Omicron (BA.1) replaced Delta.

• Elliott, P., et al. Science, 2022 24 May; 376(660). https://doi.org/10.1126/science.abq4411: identified the rapid replacement of Omicron BA.1 by BA.2 in February and March 2022.

The Real-Time Assessment of COVID Transmission (REACT) Study was established in May 2020 and provides monthly estimates of the prevalence of SARS-CoV-2 virus and bi-monthly estimates of the prevalence of antibodies to SARS-CoV-2 virus in the general population of England. To date, there are over 1.5 million people (including 30,000 adults who have tested positive for SARS-CoV-2 antigen or antibodies) in the REACT cohort and over 90% have provided consent for linkage of their study data to records held by NHS Digital.

This application is to enhance the existing REACT study research cohorts by linking the study data from adult participants (who have provided consent) to their health records held by the NHS. This enhanced dataset will then be used to advance understanding of the risks of infection and reinfection with COVID and people’s future health following a COVID infection. The outputs from this study will be delivered by June 2021 and directly feed into the UK government’s pandemic response through the partnership between Imperial College London and the Department of Health and Social Care.

Specifically, this DARS application is to request linkage between participants and the data held about them in the following data assets:

- COVID-19 Hospitalization in England Surveillance System

- COVID-19 Second Generation Surveillance System

- GPES Data for Pandemic Planning and Research (COVID-19)

- Covid-19 UK Nonhospital Antibody Testing Results (Pillar 3)

- Covid-19 UK Nonhospital Antigen Testing Results (pillar 2)

- COVID Vaccine status

- COVID Vaccine adverse reactions

- Hospital Episode Statistics Outpatients

- Hospital Episode Statistics Admitted Patient Care

- Civil Registration – Deaths

Department for Health and Social Care and Imperial College London are joint data controllers of the REACT study. IPSOS MORI is a data processor in the REACT study but will not participate in data processing for the data requested in this application. This application will allow completion of the research study which is funded by HDR UK Data and Connectivity funding scheme for National Core Studies. Imperial College London will be the only data processor on this research study.

Aim and purpose:

The aim of this application is to achieve linkage between the adult participants in the REACT study to information held by the NHS about their health, utilisation of healthcare, COVID antigen and antibody testing and COVID vaccination status. To achieve this, Imperial College will provide NHS Numbers and agreed demographic identifiers plus a study identification number (for participants who have provided consent for data linkage) to NHS Digital to provide pseudonymised data with Study ID rather than patient identifiers in the dataset.

The REACT study has two arms. REACT-1 provides near-real time estimates of the prevalence of SARS-CoV-2 virus in the general population of England (aged 5 and over) using a self- or parent-administered throat and nose swab tested by RT-PCR. Around 150,000 people have been recruited into the REACT-1 cohort each month since May 2020 and the most recent findings are from Round 9 in February 2021. REACT-2 provides estimates of the prevalence of IgG antibodies against SARS-CoV-2 spike protein using a self-administered lateral flow immunoassay (LFIA). Around 150,00 people are recruited during a two-week window every 2 months since June 2020 and the most recent findings from the Round in January 2021. To date, there are over 1.5 million people (including 30,000 adults who have tested positive for SARS-CoV-2 antigen or antibodies) in the REACT cohort and over 90% have provided consent for linkage of their study data to records held by NHS Digital.

This enriched cohort will be used for research to deliver urgent information for healthcare planning on individual-risk of COVID-19 and medium-term health outcomes of COVID-19 across the spectrum from asymptomatic to severe disease. The overarching research question is “What are the biological, social and environmental drivers of medium-term health outcomes following infection with SARS-CoV-2?”. This agreement is for an initial data extract at the start of the study with a second updated extract in May 2021 to update the study findings with a larger dataset with a longer follow-up period. This study is funded by ONS HDR UK Data and Connectivity National Core Study until 30th June 2021.

The IPSOS MORI/Imperial REACT I Antigen study latest DARS agreement is # DARS-NIC-393650-B7J6F-v2.3 and the IPSOS MORI/Imperial REACT II Antibody Study latest agreement is # DARS-NIC-389914-N9R8R-v3.2. These studies are under the data controllership of the Department of Health and Social Care and Imperial College London with IPSOS MORI as a data processor.

This study is relying on Article 6(1)(e) processing is necessary for the performance of a task in the public interest, the data controller Imperial College London are a public authority carrying out the research in the public interest. Article 9(2)(j) (processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes) the data are required for research purposes in the public interest, the research on the biological, social and environmental drivers of medium-term health outcomes following infection with SARS-CoV-2 is of public health importance. The study is performing research on a disease of high public health importance and the data request is to support a high profile National Core Study.

In this application, data is requested for individuals who have participated in the REACT study since May 2020 and provided consent for linkage of their data held by the REACT study with information the NHS holds about them.

Data requested:

Data is requested at the individual level from the following datasets:

1. COVID-19 Hospitalization in England Surveillance System: this dataset has national coverage of clinical and administrative data collected from people during a hospital admissions with COVID-19. Its purpose is for surveillance of secondary care cases of COVID-19. It was established on 15th March 2020 adapted from the system in pace for Influenza surveillance and provides a core dataset that enhances Hospital Episode Statistics through the inclusion of detailed, coded clinical information relevant to COVID-19 disease risk and outcomes. It will be used to provide information about the clinical status of people within the cohort who are treated in hospital for COVID-19.

2. COVID-19 Second Generation Surveillance System: this dataset is the national reporting system for routine laboratory testing for SARS-CoV-2, it builds on the existing SGSS for laboratories. Its purpose is for surveillance of confirmed cases of SARS-CoV-2 and testing began on 24th January 2020. It contains positive test results from Pillar 1 and Pillar 2 antigen testing. It will be used to define the COVID exposure status of participants in the cohort by providing dates when people had a positive test.

3. GPES Data for Pandemic Planning and Research (COVID-19): this dataset has been established to support research and planning to inform the UK pandemic response. All GP practices in England are required to submit a defined, coded dataset every two weeks and the national data-opt-out may not apply. Collection commenced in May 2020. It will be used to provide information about the clinical status of people within the cohort following a diagnosis of COVID. People who do not get diagnosed with COVID will also be followed up to allow comparison between health and healthcare in those who do and do not have episodes of infection.

4. Covid-19 UK Nonhospital Antibody Testing Results (Pillar 3): This dataset is national (UK-wide) surveillance of serology testing, both positive and negative test results. It draws in data from multiple sources. It will be used to define the COVID exposure status of participants in the cohort by providing information about whether people have antibodies or not. It will be specifically interpreted in the context of (6).

5. Covid-19 UK Nonhospital Antigen Testing Results (pillar 2): this dataset is surveillance of the national swab testing for the general population (excluding patients in hospital; health and social care workers; test processed by PHE or hospital laboratories) undertaken in venues such as at home, drive through or mobile test centres. It will be used in combination with (2) to define the COVID exposure status of participants in the cohort by providing dates when people had a positive test.

6. COVID-19 Vaccination Status Data: this dataset has recently been established and contains records of COVID vaccination events in any setting. It will be used to identify the dates when participants in the cohort received doses of a COVID vaccine. People will be followed from this point onwards and their post-vaccine COVID exposure status as obtained from (2); (4) and (5) will be characterised and explored.

7. COVID Vaccination adverse reactions: this dataset is related to dataset (6) and provides information about adverse reactions that occur immediately following the administration of a COVID vaccine.

8. Hospital Episode Statistics Outpatients and Admitted Patient Care: these datasets contain a complete record of all episodes of hospital healthcare in English NHS Hospitals. They will be used to provide information about the clinical status and healthcare utilisation of people within the cohort following a diagnosis of COVID. People who do not get diagnosed with COVID will also be followed up to allow comparison between health and healthcare in those who do and do not have episodes of infection.

9. Civil Registration – Deaths: this dataset is obtained via the Office for National Statistics and contains a complete record of all deaths in the UK with information about date, place and cause of death. This dataset will be used to identify death as a health outcome and as the date of leaving the cohort (for time-to-event analyses).

These data will be linked to the data from the REACT study and used to construct a longitudinal record for the participating members of the REACT cohort that described pre-pandemic health status; COVID-19 testing and results; healthcare utilisation (quantity and reason) since COVID exposure status was determined. It is necessary to request linkage for all consenting participants in the REACT study as this is a representative population-based cohort and it will allow robust comparisons across the longitudinal health profiles of participants with all possible histories of COVID exposure, including across the spectrum from asymptomatic to severe disease. Data is being requested from before entry to the cohort (one year, from 1st March 2019). The rationale for “looking back” over the 12 months before the pandemic is to establish a “baseline” health status for individuals before people were infected with SARS-CoV-2 and before disruptions to healthcare may have introduced bias into data capture. This will allow comparison of health outcomes to be adjusted for pre-exisiting health and healthcare utilisation. Participants will be followed up into the future for as long as the data allows. This initial research is funded to June 2021 and it is anticipated that extensions to these data will be requested to permit longer duration of follow up (up to 20 years) subject to appropriate research needs as more is learnt about the long term natural history of COVID-19. Any further data requests would be subject to an amendment to this agreement being approved.

It is necessary to request individual level data as the data will be linked to individual participants in the REACT cohort and analysis will be conducted taking into account variables including age, sex and location of residence (postcode in cohort data). The data that will be in the extract can be pseudonymised through a study ID that can be linked to NHS number in the original cohort and will be provided by Imperial College. Follow up is currently a maximum of 12 months given the timing of the origin of COVID-19 and will extend as time since infection/vaccination increases as we follow the health outcomes post-infection/vaccination (subject to further DARS applications/extensions and research need). Data is requested on all health outcomes as this research aims to advance understanding of future health status of people characterised by their exposure to COVID-19. As more information becomes available about COVID-19 natural history, this request will be reviewed. We have selected the fields relevant to this research and minimised the number of identifiable/high-risk fields requested.

The processing of this data during the conduct of the proposed research will produce preliminary study findings that will be shared internally with the REACT study team and DHSC (Joint Data Controller of REACT study) to achieve timely impact from study insights and findings.

Final outputs of this processing will include working papers, reports, blogs, social media (including institutional twitter accounts e.g. @Imperial_PERC and @ImperialSPH), educational seminars. pre-print journal-style articles and peer-reviewed publications with press-releases, as appropriate (likely to appear in similar journals to those current REACT research is published in i.e. British Medical Journal, Nature Communications etc.). This study is required to produce outputs (including reports and peer reviewed publications) by July 2021.

The study has a Public and Patient panel comprised of 4 members of the existing REACT Patient Advisory Group and 4 new members. The Panel will be involved in the production of outputs and the dissemination strategy to ensure that materials are fit for purpose and meet the needs of participants and the wider public. The dissemination strategy will link into the existing programme of activity for the wider REACT programme.

All outputs will have appropriate statistical disclosure control applied and no identifying information will leave the secure research environment. All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

The REACT Investigators will also apply for future funding to maintain and extend the research undertaken on this enhanced research cohort. As part of this activity, a Data Access Committee is in development to oversee future research direction.

Yielded Benefits is not a requirement for new applications.