KHP Data Analytics and Modelling COVID-19 Research data warehouse

With the spread of novel COVID-19 in the UK since January 2020 there is a rapid need to undertake research to inform the clinical response. Guy’s and St Thomas’ (GSTT) and King’s College Hospital NHS Foundation Trusts (KCH) have been initially at the epicentre of the UK pandemic, and it is anticipated that there will be significant ongoing activities within these trusts as subsequent waves develop. In line with HDRUK championing the use of health data to support research and innovation that addresses challenges presented by the COVID-19 pandemic, GSTT and KCH have set up a KHP COVID Data Analytics & Modelling Group and have established a Research Data Warehouse. Both trusts are part of King’s Health Partners (KHP). Linking the Research data warehouse with nationally collected data sets will allow the development of risk calculators for primary care that would help with investigating best treatment plans for individuals throughout their entire experience with Covid-19.

GSTT and KCH are the data controllers for this project.

The purpose of this programme is to utilise the electronic health records (EHRs) to identify patients tested for COVID-19, Influenza and for respiratory pathogens routinely tested for in the KHP laboratories for those attending with respiratory illness and describe them phenotypically - allowing tracking and comparison of disease progression, care, initial and long term outcomes. Therefore - a wide range of research questions can be asked and answered using this data.

Routine data to be collected includes clinical, laboratory and demographic data from all patients tested for COVID-19, Influenza and other respiratory pathogens (e.g Parainfluenza, Legionella). All data as specified for these patients from 01/01/2016 onwards will be available.. All projects requesting access to the data within the data warehouse will be COVID-19 only focused only

The finalised dataset available in the datawarehouse would include, but will not be limited to:

• Patient history and co-morbidities

• Observations

• Pathology test results

• Demographics

• Prescription data

• Imaging data

Research questions will be discussed and agreed by a core group of clinicians and academics including the Chief Clinical Information Officer and R&D Director. Each question will have clinical input from appropriate persons e.g. ICU consultant. Given the exceptional circumstances of this public health crisis, an expedited review process for all data access requests will be used. Data minimisation principles will be followed for each release of data.

Data will only be permitted to be analysed by a member of Kings Health Partners (KHP). Both trusts mentioned in this agreement are part of King’s Health Partners (KHP), as well as the University.

This programme will utilise clinical and academic expertise across Kings Health Partners (KHP).

KHP brings together a world leading research led university Kings College London, and three successful NHS Foundation Trusts: Guy’s and St Thomas’, King’s College Hospital and South London and Maudsley.

KHP is one of only eight Academic Health Sciences Centres in England designated by the Department of Health. These include:

• Bristol Health Partners

• Cambridge University Health Partners

• Imperial College

• Manchester

• Newcastle Health Innovation Partners

• Oxford Academic Health Partners

• UCL Partners

Only data from KCH and GSTT will be involved in this project, as South London and Maudsley is a Mental health service provider.

The three founding trusts serve a local population which is among the most ethnically, socially and economically diverse in the world. Together they provide a full range of medical and healthcare services, from acute and specialist medical care, local and highly specialised mental healthcare, and services that promote physical and psychological wellbeing.

Linkage with the Hospital Episode Statistics Critical Care data would allow for tracking repeat admissions into GSTT, post and prior COVID-19 diagnosis, Influenza and other respiratory pathogens while providing information on admissions to critical care units besides GSTT.

All linkage will enable better understanding on long term healthcare, utilisation, and outcomes and will allow for a better understanding and prediction of future healthcare requirements.

Studying longer term outcomes will allow an assessment of new policy changes to healthcare or the implementation of a new therapy. There is a risk of Covid-19 being with us for an extended period of time, and investigating the relationship between treatments in primary and secondary care will be crucial in handling any future waves of the epidemic, and is necessary for reasons of public interest in the area of public health. None of the research uses identifiable data.

All data is pseudonymised at source, and data linkage will occur within the GSTT fire wall.

Data requested from NHS Digital will include the HES Critical Care data set, to track long term outcomes of patients tested/treated for Covid-19, flu and other respiratory pathogens and subsequent Critical Care admissions and outcomes. The full HES Critical Care data set, for patients tested for Covid-19, Flu and other respiratory pathogens is requested. Data minimisation standards will be adhered to, as only data from January 2020 onwards would be requested, and only for the patients specified. As Covid-19 is likely to be with us for some time, and there is uncertainty around the long term effects of this disease, the patient level data request would be for most data up until the most recent poitn available - looking for Critical Care treatment In the future.

The use of the data described received ethical approval in July 2020 and meets all the required ethical and legal requirements. This data is secondary use, and processing outside of GDPR requirements would be unlawful. The risks around confidentiality are low, Data processing and storage of the source data will be within the trust firewall meeting all security requirements of GSTT. Data released for analysis will be pseudo-anonymised. All steps for data minimisation, secure transfer, storage and access will be adhered to.

Data acquisition would not involve the capture of any new data. Instead it would involve cataloguing EHR data and standardising it in such a way that it has the potential to be aggregated.

The finalised dataset made available to researchers at GSTT/KCH/KCL would include the following hospital data, linked to the critical care data provided by NHS Digital :

• Patient history and co-morbidities

• Observations (Including blood panels and vital signs)

• Pathology test results

• Demographics (including gender, age, index of multiple deprivation)

• Prescription data

• Imaging data (including x-ray and CT scan data)

• Primary care data from Lambeth data net (including data on long term conditions, medications, vaccination history and demographics)

GSTT has obtained generic ethical approval for research projects using the stored data, under conditions agreed with the Research Ethics Committee (REC), without requirement for KHP researchers to apply individually to the REC for approval, for COVID-19 projects only.

An example of such projects would be:

- Impact of ethnicity on outcome of severe COVID-19 infection. Data from an ethnically diverse UK tertiary centre, IMD effects of outcomes of Covid -19.

Association between Renin-angiotensin-aldosterone system (RASS) inhibitors and clinical outcomes in patients with COVID-19: a cohort study in London).

All data requests will require an application, that will be discussed and agreed by a core group of clinicians and academics including the Chief Clinical Information Officer and Research & Development Director.

Each question will have clinical input from appropriate persons e.g. ICU consultant, and each request must be approved by this group before any data is released.

Once an application has been approved, selected data points will be made available for researchers in a secure, restricted access KCL supported platform in which to undertake the analysis.

Data minimisation principles will be followed for each release of data. Data will only be permitted to be analysed by a member of KHP. No data will be shared with any other organisations not mentioned in this agreement. Only data points specifically required for each project will be requested/released. Identifiable data will not be available to researchers.

A log of all questions will be maintained and will be reported annually to the REC and to the host Research & Development department via the annual progress report.

All data transfers for research/investigative purposes will be logged and linked to the research questions. Data transfers will be covered by data sharing agreements confirming the terms and conditions of transfer where applicable.

Access to data will include researcher data retention and limit use to a particular study or set of studies.

All research/studies will be COVID19 related only.

All data are copies of data held in clinical systems, there is no obligation to retain these data. In line with NHS guidance if the project were to end, Data would be securely deleted.

Data controllers: Guys and ST Thomas NHS Foundation Trust, Kings College Hospital NHS Foundation Trust

Data Processors: Certain members of KHP including Guys and St Thomas NHS Foundation Trust, Kings College Hospital NHS Foundation Trust, and KCL (King's College London)

The legal basis relied on for processing under the General Data Protection Regulation (GDPR) is Article 6 (1)(E) - (processing is necessary for the performance of a task in the public interest or in the exercise of official authority vested in the controller) and Article 9 (2)(I) ("Processing is necessary for reasons of public interest in the area of public health, such as protecting against serious cross-border threats to health or ensuring high standards of quality and safety of health care and of medicinal products or medical devices, on the basis of Union or Member State law which provides for suitable and specific measures to safeguard the rights and freedoms of the data subject, in particular professional secrecy";) as the legal basis to process special category data (health data). Both data controllers for this project are considered public bodies as they are health care trusts - therefore can rely on Article 6 (1)(E). The projects this data will be used for are solely related to COVID-19, a current worldwide pandemic, therefore meeting the criteria as necessary for the interest of public health.

All data flows will be at a patient level.

NHS Number, Local Identifier (Study ID), Gender, Date of Birth and postcode will be transferred to NHS Digital from GSTT. The cohort will consists of people who were admitted to GSTT and KCH since January 2020 and were tested for CV19 and other respiratory pathogens as described above.

NHS Digital will return a pseudonymised data set back to GSTT - including a study ID (all other identifiers removed) and the HES Criticial Care dataset from 19/20 onwards.

It is anticipated this process will be of an adhoc nature, and routine flows will not be required from NHS Digital (ie - data is not required on a repeat routine quarterly basis)

Once HES data is linked to the GSTT Data, using the local identifier (Study ID), all data minimisation standards, and pseudonymisation and transformation will take within the GSTT fire wall, before data is realised for analysis.

Data is pseudonymised and linked by GSTT within the GSTT firewall. Statistical analysis will take place within secure KCL research environments. All data that is made available for analysis will be separated from source data, and re linkage will not be possible.

There will be no requirement/attempt to re-identify individuals.

All data requests will require an application, that will be discussed and agreed by a core group of clinicians and academics including the Chief Clinical Information Officer and R&D Director. Each question will have clinical input from appropriate persons e.g. ICU consultant, and each request must be approved by this group before any data is released.

Given the exceptional circumstances of this public health crisis, an expedited review process for all data access requests will be used

Data requests will only be granted for applications related to CV19 research.

Once an application has been approved, selected data points will be made available for researchers in a secure, restricted access KCL supported platform in which to undertake the analysis.

All data transfers for research /investigative purposes will be logged and linked to the research questions. Data transfers will be covered by data sharing agreements confirming the terms and conditions of transfer where applicable.

All data processing is only carried out by substantive employees of the data processor, GSTT, KCH or KCL or data controller, GSTT or KCH. All staff have been appropriately trained in data protection and confidentiality.

All data will be accessed using remote desktop access onto the secure research environments within KCL. Access to data within GSTT will also be through secure remote desktop connections into the GSTT/KCH fire walls.

A log of all research questions and projects will be maintained and will be reported annually to the REC and to the host Research + Development department via the annual progress report.

The following may be produced because of this data processing:

a. Reports as requested

b. Submissions to peer reviewed journals

c. Presentations

d. Conferences as requested and as it is feasible.

e. Development and implementation of prediction tools and algorithms

Examples of such would be:

- Submissions around looking at symptom onset date before hospital presentation, outcome impact

- Algorithms around Prediction which patients have COVID19 based on presentation/’at the door’ symptoms and preliminary tests results, to enable accurate traiging of patients.

- Aggregate level data (with small number suppression applied in line with the HES analysis guide) would be contained in these outputs only.

Data will always be released at a high enough level of aggregation to prevent others being able to ‘recognise' a particular individual.

Each research project will aim to have their own website with contact details and key results. Research Snapshots will also be published. These are summaries of projects suitable for patients and lay audiences. The Research Database team will maintain a publicly accessible register of research projects using data from the database on the GSTT Biomedical Research Centre website. For transparency and interaction, the Health Reasearch Authority are publicly publishing a list of all COVID-19 NHS REC approved studies which is already live here: https://www.hra.nhs.uk/covid-19-research/approved-covid-19-research.

With the spread of novel COVID-19 in the UK since January 2020 there is a rapid need to undertake research to inform the clinical response. Guy’s and St Thomas’ (GSTT) and King’s College Hospital NHS Foundation Trusts (KCH) are at the epicentre of the UK pandemic having seen over 4000 patients at the time of writing. This is in line with HDRUK who are championing the use of health data to support research and innovation that addresses challenges presented by the COVID-19 pandemic.

All linkage will enable understanding of long term healthcare utilisation and outcomes and will allow for a better understanding and prediction of future healthcare requirements. Studying longer term outcomes will allow an assessment of new policy changes to healthcare or the implementation of a new therapy. There is a risk of Covid-19 being with us for an extended period, and investigating the relationship between treatments in primary and secondary care will be crucial in handling any future waves of the pandemic and is necessary for reasons of public interest in the area of public health.

Out puts generated using this data will allow for more informed information around treatments, risk factors and severity of disease. Outputs of these messages to be communicated with the scientific and medical community as well as the public and patients.

Benefits to medical community and patients of this information would be to allow more well informed decisions with regards to Covid-19. For example, one project will be looking at the impact of symptom on set date and hospital presentation and comparing with future presentations. This particular project (amongts othrs) has the potential to inform medical staff at the front door, and of patients and public of when is a more demonstrably advantageous time to present given severity and length of symptoms.

GSTT have previously engaged with the GSTT Biomedical Research Centre Public and Patient Involvement Advisory Group and have sought their opinions on the sharing, collection and use of data in research. A FAQ (Frequencly Asekd Questions) session on data concerns of the patients has been completed and is available on the GSTT BRC website. Please see here:

https://www.guysandstthomas.nhs.uk/research/patients/use-of-data.aspx

Yielded Benefits is not a requirement for new applications.