COVID-19 vaccination and disease and the risks of events and complications including major venous and arterial vascular events

The main objective for processing data for this data access request is to support vaccine safety and vaccine effectiveness research.

BACKGROUND

There is controversy and public concern about the extent of significant adverse events after COVID-19 vaccination. Overwhelmingly, vaccination is important to the general health of the population and to ensure as a society we can recover from the effects of the pandemic. However, there remains a critical need to understand which groups might be more at risk of vaccine-associated adverse events, and which vaccines might be most likely to be associated with these adverse events. Reports in particular of adverse events relating to blood clotting, and an increase in the rare events of blood clotting in the veins in unusual places or low levels of platelet counts.

This agreement is led by University Hospitals Birmingham NHS Foundation Trust (UHB) via PopulatION systEms for acutE caRe (PIONEER) Data Hub and is part of a larger project led by Health Data Research UK (HDR UK). The larger National Core Study has commissioned work from PIONEER, but also from: Barts Health NHS Trust/Imperial College London - they are the Data Controllers for patients in their London locality; Combined Intelligence for Population Health Action (CIPHA) – they are the Data Controller for patients in their North West locality. The British Heart Foundation (BHF) Data Science Centre are providing analytical support to the study. The additional Data Controllers for the larger National Core Study are mentioned and explained in this section for completeness in this agreement. However, only UHB via PIONEER will be Data Controller for the data applied for in this agreement. Furthermore, no data will be linked across the regions participating in the larger National Core Study, but aggregate summaries (with small numbers suppressed) will be produced across the regions to provide a national report.

UHB is already the data controller of the PIONEER data and will also be the data controller of the data following the linkage of Vaccination Status and Vaccination Adverse Reaction Data for the UHB cohort of patients (see also further below). The concept here is that the dataset, which UHB already holds for their patients for research purposes under the PIONEER health research database (which has obtained favourable ethical opinion, as well as Confidentiality Advisory Group (CAG) approval), is expanded to include Vaccination Status and Vaccine Adverse Reaction Status which is hoped to greatly benefit further Covid research.

The data obtained under this agreement is hoped to provide information about adverse events for the Birmingham population, who have attended UHB Emergency Department or been admitted to one of the UHB hospital sites. Staff employed at UHB will work on this data, and identified analysts from other organisations may work on an anonymised version of the data to supplement expertise via a sublicence application through PIONEER. Any external parties would follow the CAG and Research Ethics Committee (REC) approved PIONEER processes as defined in the protocol on data access. All requests for licensed data will be reviewed against the principles of the ‘Five Safes’ by the Data Trust Committee (members of the public), the PIONEER Management team and the PIONEER Data Controller (UHB). The Five Safes framework is a set of principles to provide safe research access to data. The framework has become best practice in data protection whilst fulfilling the demands of open science and transparency. Safe data (data is treated to protect any confidentiality concerns), Safe projects (research projects are approved for the public good), Safe people (researchers are trained and authorised to use data safely, Safe settings (a secure environment prevents unauthorised use) and Safe outputs (screened and approved outputs that are non-disclosive).

Subject to approval, the vaccine data will be linked with UHB data in the PIONEER database. The sub-licence will be applicable, due to the addition of the NHS Digital Vaccine data, and for the avoidance of doubt, only relates to the data being requested in this agreement, all other PIONEER data is a direct licence application via UHB via the existing established processes. The sub-license for the NHS Digital data (the CV19 vaccine data and the PIONEER data) will be as per the NHS Digital terms.

The aim of the project is to explore the added value of enabling regional, acute care admissions and proxy data flows linked to key data sets such as Vaccination Status to enable surveillance and research on the COVID-19 vaccination programme – specifically to evaluate the safety and effectiveness of the vaccination programme and answer key questions of clinical, regulatory and policy importance, including the incidence of rare adverse events such vaccine-induced immune thrombocytopenia and thrombosis (VITT). Current national data feeds (e.g Hospital Episode Statistics (HES) are limited by being dated, incomplete or lacking sufficient diagnostic coding granularity. Recent analyses of national data to determine the association of COVID-19 vaccines with VITT events have confirmed limitations in these analyses due to time lag and lack of linked laboratory data flows. The critical requirements for a continuous data feed, therefore, are as follows:

* Multi-source linked patient-level data from acute care hospital admissions and vaccinations

* Haematology lab patient-level data feeds (for platelet counts +/- other haematological indices)

* Clear definition of the population covered for the project, so the project can calculate the incidence of adverse effects per head of population.

* Diagnostic coding as close to real time as possible and resolution beyond ICD-10 4 digits (e.g to enable different types of vein thromboses to be identified)

Vaccine data will be requested using NHS number, Date of Birth and postcode, with a study ID. NHS number, Date of Birth and postcode will not be used when data is returned to UHB PIONEER. Only the study ID will be returned with the NHS Digital Vaccine status/adverse reaction data to perform linkage with the PIONEER database.

This research project hope to first address the following research questions:

1) What is the incidence of potential vaccine related complications using data ascertained within hospital compared with the incidence from hospital discharge data only?

2) What is the agreement in ascertainment of vaccine-related complications from within hospital data compared with hospital discharge?

3) What is the average difference in time from vaccination to complication using data from Emergency Departments (ED) or the initial objective was to assess the feasibility of curating a minimum dataset to demonstrate thrombotic complications from Covid vaccination. If successful, the approach may have value in addressing other important policy-related research questions relevant to vaccine safety and effectiveness.

Depending on the answers to the above questions further research questions relating to the safety of vaccines may be worked on.

To date, the project has resulted in a curated data set (acute admissions data coded pre-discharge and laboratory data)– included in PIONEER data Hub (UHB NHS Trust). This agreement is requesting linkage of Vaccination Status and Vaccination Adverse Reaction Data for a cohort of patients provided by UHB. The cohort for linking will be any patients who have had an acute admission, including Emergency Department (ED) attendance to UHB since 8th December 2020 with either a diagnosis of COVID-19 or a symptom relating to a vaccine adverse event. The indicative cohort size is approximately 27,200. The project and planned analyses may a) establish proof of concept, b) support immediate data and connectivity needs of vaccine safety research, and c) explore the added value of enabling such rapid regional data flows with increased granularity (e.g. enhanced ICD coding to distinguish between different types of venous thrombosis or availability of laboratory data for sensitive identification of thrombocytopenia).

The approach taken and phenotypic coding (which are measurable, observable characteristics in admission, laboratory and vaccination coded data to identify thrombotic adverse events) proposed for the analyses were reviewed and validated by an expert group of clinical haematologists (including members of the British Society for Haematology Expert Haematology Panel) and clinical informatics leads at a workshop in early October 2021

In order to meet these requirements, Patient and Public involvement and engagement (PPIE) work was undertaken to obtain feedback and opinion on the project and provide PPIE input to the planned research. The feedback received from the patients was positive and supportive, and a group of the PPIE agreed to be involved in shaping this research and helping to disseminate the results. The team acknowledge that care needs to be taken in how the results are fed back to the public and hopes that the engagement with the PPIE team will help to mitigate some of this risk.

Several themes emerged from the workshop discussions, including:

Public benefit: The perceived benefit to the public meant that attendees were supportive of this project being scaled up across the UK. While it was identified that there are risks to this type of work, there was a strong consensus that the benefits outweigh the risks. This can be assisted by ensuring that appropriate safeguards are in place.

The role of the patient/public: There was discussion around how much responsibility is with patients and the public to be actively seeking to find out about how their data is being used, or if researchers need to be providing that information more freely. It was agreed that involving patients and the public in the development of communications and engagement plans is vital to adequately ensure the public is informed. This helps to build trust and transparency.

Additional data linkages: There were no concerns around additional data linkages from national data sets (such as vaccination status data) being made to routine health data support this project and other urgent pandemic research. There was also broad support for data to be joined up nationally if that was of best interest to research and outcomes. As well, uses of the data beyond COVID-19 were supported, and it was felt that the learnings from COVID-19 should be used in other areas moving forward.

This workshop has demonstrated that there is public support for this work, and this is hoped to support the proposal for a near-real time hospital admissions feed to be scaled up across the UK. One key next step identified from the workshop is to develop a communications plan, with input from patients and the public, to ensure that this work is being communicated transparently to the public. This work is being led centrally by HDR UK. PIONEER/UHB are working on their own supporting PPIE comms and engagement to supplement this and work is continuing. PIONEER/UHB also plan to hold follow-up workshop/engagement after results have been produced. Patient and Public input has been embedded into the structure and culture of PIONEER/UHB. Patient and Public involvement and engagement is not just to disseminate, PIONEER ensure this is throughout the lifecycle.

The outputs (aggregated (with small number suppressed) summaries and reports) from this project is hoped to feed into a proposal for national scale up of enabling such regional data flows, to support the vaccine safety research and potentially other use cases. The project is funded by the Data and Connectivity National Core Study and will run for an approximate duration of 9 months. Data will be required to be retained for 1 year to address any queries or challenges post publication of results, if an extension to this period is required an application will be made to NHS Digital. The data from NHS Digital is vital to enable linkage to Vaccination data to understand the association between post vaccination adverse events and vaccination type. No elements of the work are occurring outside the UK.

LEGAL BASIS

GDPR 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” As a teaching hospital, it facilitates research, which brings this processing in line with our performance of a task in the public interest. Further, as a public benefit organisation, the organisation is unable to rely on an alternative legal basis such as ‘legitimate interests'.

Additionally, as health data is a special category of personal data, UHB will be using GDPR Article 9(2) (j) – “processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject-“as the data are required for research purposes in the public interest (meeting the conditions in the DPA 2018 Schedule 1 Part 1 (4) - which GDPR Recital 52(2) determines is an appropriate derogation from the prohibition on processing special categories of personal data) and data processing is subject to appropriate safeguards in accordance with GDPR Article 89(1)). UHB are the sole data controller for PIONEER Data hub, and also the NHS Digital Vaccine Data. UHB are also the only organisation processing the vaccine data unless there is an approved sub-licence. The funders, HDR UK, only act in an advisory role and UHB make the final decisions on the study. The other organisations (CIPHA, Barts, Imperial) are conducting a study on their own data so only act in advisory roles for this project.

The combined data within PIONEER will be different to that held within NHS Digital and will provide a value-added service. The COVID vaccination status and adverse reaction data will be linked to UHB granular electronic health records, including serial physiology and laboratory results which are not available via NHS Digital.

1) UHB NHS Foundation Trust will provide a cohort to NHS Digital – to include the following data items (identifiable data):

a) Study ID

b) NHS number

c) Date of Birth

d) Postcode

2) NHS Digital will then link cohort data to Vaccination Status and Vaccination Adverse Reaction data as a data extract– monthly refreshes of data will be requested. Data will only be requested for patients in the UHB Cohort with acute admissions. The pseudonymised data plus the Study ID only (not the identifiable NHS number, Date of Birth and Postcode data) will then be returned by NHS Digital to UHB Birmingham via Secure Electronic File Transfer (SEFT). The cohort (as stated in Section 5a) will be any patients who have had an acute admission, including an Emergency Department (ED) attendance to UHB since 8th December 2020 with either a diagnosis of COVID-19 or a symptom relating to a vaccine adverse event. The exact number has not been calculated but an estimate can be derived from the number of COVID-19 positive patients who attended or were admitted to UHB since March 2020 through to date – this is approximately 27,200. This study starts in September 2020 so the number will be less than the total reported here.

3) Substantive staff at UHB will then link the data from NHS Digital to UHB internal data. No patients will be re-identified by UHB staff during this linking process. The NHS Digital Vaccine and UHB PIONEER data will be kept in distinct secure folders, to protect NHS Digital data from reidentification. A research dataset using this linked cohort will be created; this dataset will be pseudonymised as per PIONEER protocols. All employees at UHB must undergo annual Information Governance training, which includes passing the subsequent test.

The initial data cohort will be processed and stored on a secure Structured Query Language (SQL) server which sits within the Trust's IT framework and is not available for querying by any unauthorised external parties or tools. Passwords are strictly controlled by UHB IT services, with access permissions for each user administered on the authorisation through an identified senior member of staff at UHB. The physical servers are located on site at UHB in a locked storage room where entry in and out of the rooms is logged and controlled by senior managers.

4) Access to the linked cohort data set for analysis will be via the secure PIONEER Trusted Research Environment (TRE). Data will be stored on a secure, UHB-controlled Microsoft Azure cloud platform in accordance with the 14 UK Cyber Cloud Principles. Applications for access to the curated, linked dataset approved researchers will be via the PIONEER Data Trust Committee - UHB, as Data Controller for PIONEER, has final approval of what data can be shared, and oversees the legal contractual process to ensure this happens safely.

5) Data made available in the PIONEER TRE for secure, approved researcher analysis will be de-identified, with all outputs checked for risk of disclosure or reidentification, to ensure that statistical results of data analysis are non-disclosive.

The legal basis for the data being sent to (or reused by) NHS Digital is National Health Service Act 2006 - s251 - 'Control of patient information'.

UHB use datacentre-as-a-service (co-location) from Crown Hosting, a HM Government-approved datacentre facility, supplied by ARK Data Centres. The server and storage equipment is Trust owned and operated via a privately dedicated link to UHB. IT Services span the local network to the servers and storage hosted at ARK Data Centres, accessing data using the same protocols that UHB do with servers and storage at their other location. Therefore, ARK Data Centres do not access data held under this agreement as they only supply the building. Therefore, any access to the data held under this agreement would be considered a breach of the agreement. This includes granting of access to the database[s] containing the data.

Small number suppression and data disclosure: In order to protect patient confidentiality, when presenting results calculated from patient record level data, outputs will contain only aggregate level data with small numbers suppressed to stop people from identifying themselves and others, and to ensure confidentiality is maintained.

Plan for National Project Outputs

Subject to approval, analyses produced by this data processing will be written up and published. Planned examples are :

• reports to funders (HDRUK, UKRI), and Governance Boards of the National Core Studies (NCS)

• report to NHS England, other key policy makers

• Academic publication, dissemination via presentation at conferences, NCS meetings

This project is part of the Data and Connectivity (D&C) NCS, one of six National Core Studies that were set up by the Government Office to tackle a number of areas where the UK needed to increase its research scale or infrastructure to respond to key near-term strategic, policy and operational questions regarding COVID-19. The Data and Connectivity NCS is led by Health Data Research UK (HDR UK) in partnership with Office for National Statistics (ONS), and connects UK health data to support and accelerate research on COVID-19, enabling streamlined data access and analysis for COVID-19. Data and Connectivity supports the other NCS established by the UK Government.

Level of data contained in outputs listed above will be aggregate data (with small numbers suppressed) only compliant with all disclosure rules.

The dissemination plan includes:

• Report describing methods, results and next steps (including use cases of this work beyond COVID) and to support future pandemic preparedness) to HDRUK, UK Research and Innovation (UKRI), Data and Connectivity NCS Governance, NHSE, key policymakers in December 2022. This work will form a key part of recommendations to NHS England, NHS Digital and devolved nation equivalents on the use of regional, acute hospital admissions data feeds to support urgent research, surveillance and clinical audit.

• Draft for academic publication December 2022, with the aim for publication Spring 2023 – any publication will be open access. It is anticipated the outputs will be submitted to the British Medical Journal, the Journal for Healthcare Quality, Health Services Research Journal for publication, subject to acceptance.

• Series of presentations at dissemination events, and conferences across 2023. The expected conferences for dissemination are Digital Healthcare Conferences London, Wired Health and IPDLN Conference (Data linkage research: Informing policy and practice).

• Series of dissemination channels employed to share findings – publicly accessible D&C web pages (with lay reports produced for the public), D&C NCS newsletters, Gov.UK National Core webpage, briefing documents, stakeholder mailings lists.

• Series of presentations at dissemination events to PPIE groups, to ensure they see both the outputs and the benefits of the project.

The project is working with organisations based in four regional integrated health care data systems (ICS) across England who are working to curate regional rapid near-real time hospital admissions data feeds. These are PIONEER Data Hub (Birmingham and West Midlands ICS), CIPHA (Cheshire and Mersey ICS), Barts Health NHS Trust (North-East London ICS) and Imperial College healthcare (North-West London ICS). The findings and learnings will inform a proposal to scale this across the UK, developing regional capabilities that can provide intelligence at a national scale to support research, including future pandemic preparedness and urgent non COVID use cases such as pharmacovigilance and early identification of admissions pressures. Each of these sites are acting as a pilot for their region, depending on the results there is a plan to extend this work (ethics will be addressed by each region and area). This is outside the scope of this agreement, as this only refers to the work PIONEER is undertaking. The other regions are mentioned for completeness.

UHB PIONEER Projects Outputs

UHB and the wider project group will collaborate on several research papers including the aggregated, small number suppressed (in line with the PIONEER and HES analysis guide) results of the data analyses with a view to submitting to peer-reviewed journals. Some of the results will be presented at local meetings.

Local meetings include regular seminar series organised and run by the partner organisations. These include:

• Monthly seminar series organised by Institute of Applied Health Research, University of Birmingham. These seminars are available to staff in the University of Birmingham and advertised through University of Birmingham communications channels.

• Seminar series organised by University Hospitals Birmingham NHS Foundation Trust. These health seminars are open to members of the University Hospitals Birmingham NHS Foundation Trust (patients, carers and staff) and are free to attend.

• Seminars at the Institute of Translational Medicine. Seminars@ITM are cross-discipline seminars designed to showcase the research conducted across the Birmingham Health Partners campus (University Hospitals Birmingham NHS Foundation Trust, University of Birmingham, Birmingham Women’s and Children’s NHS Foundation Trust)

The study team have involved members of the public and patients. Their insights confirmed that this study’s aims were relevant to patients and the study methods were acceptable from an ethical perspective.

The Study team will work with their public contributors to co-produce materials that are in formats that are accessible. All outputs of data will be aggregated with small number suppression applied as per the PIONEER protocol and HES analysis guide. Once the study team have analysed the data, they hope to work with their contributors to identify relevant local and national organisations (including charities) to share their findings with.

The PIONEER team intend to update their website (https://www.pioneerdatahub.co.uk/) and undertake a series of blogs, and communication events and will produce a poster to ensure patients and the public are updated on the expansion of their Research Database and the addition of the vaccine data. The study team hope to complete the first set of analysis by December 2022. Further analyses would then be undertaken during the remaining period of the DSA. Outputs will also be shared with policy makers and other relevant national bodies such as NHS-England.

As detailed in Section 5a, it is hoped that the benefits of this work will directly support urgent vaccine safety research, adding to the existing analysis from national linked data sets, to provide additional background rate information on potential VITT events. This may also be important as the Covid-19 vaccine booster programme is rolled out. The outputs of this project may help provide policymakers and NHS managers with more information about the potential increase in patients presenting with adverse vaccine events. As well as highlighting the potential burden in terms of hospital admissions, it may identify whether there is a difference in the presentation of the disease (from non-vaccine related) through analysis of laboratory results and observations (such as heart rate, temperature, blood pressure etc.). The knowledge gained from this research may also be able to stratify which patient groups would benefit from which formulation of vaccine, therefore reducing patient harm. The addition of lab results may be able to identify patients who may be at risk of developing an adverse reaction.

The magnitude of impact for this project is based upon the patient population, for UHB this is 2.2 million and the total population of the study is approximately 9.6 million, so this will provide statistical power and confidence in the findings for National scale-up. The 27,000 (indicative cohort is the numerator and 2.2 million is the denominator for UHB work). The 9.6 million is total population when all the areas in the pilot are combined, so PIONEER are trying to demonstrate this covers a base population for any of the 2.2million patients vaccinated in Birmingham but also the 9.6 million population of the whole core study.

The project may also provide proof of concept and valuable insights into the added value of supplementing currently available national datasets with regional, more rapid and granular data sets. The VITT example will act as the initial use case, however potential applications to other projects requiring timeline data flow – e.g. support for future pandemic/emergencies and rapid identification of patients with emergency conditions for trials.

As stated in Section 5a, this project is part of the National Core Studies supported by the UK Government Chief Scientific Advisor. A letter supporting this work from the UK Government Chief Scientific Advisor states these studies will be crucial parts of the UK response to the next phase of the COVID-19 pandemic. These are truly a four-nation endeavour which ensures the programmes of work are scoped, shaped and delivered in a UK-wide manner. The vision is for data across the UK, including the insights from this study, to feed into National research. The outputs generated may feed into policymaking and operational decisions in all relevant parts of the government across the UK.

The initial outputs may also provide an evidence base to support potential scale up across the UK, and will be presented to senior NHS stakeholders to explore the contribution of mature acute care NHS Trusts to the curated linked dataset, and enablement of secure access to such regional data via the UK wide Trusted Research Environment (TRE) network

This project is exploring a proof of concept – do regional, more rapid acute admissions data flows, linked to e.g. primary care data, vaccination status – enhance currently available national data feeds available (e.g. HES) to provide more detailed and near-real time information on hospital admissions across the UK. The sub-licences are projects looking at acute care presentations where vaccine status/outcome is required. Data that is currently available are only accessible post discharge, some with up to a 6-week time delay, and much of the available data is incomplete. A rapid near-real time hospital admissions data feed could provide vital information for policymakers and researchers, such as the identification of adverse effects from new drugs, like the COVID-19 vaccine. The linked NHS Digital/UHB research database may serve numerous research projects with acute care focus where vaccine data is required. All requests as described for sub-licence are reviewed by our public panel and UHB as data controller before a licence/sub licence is issued. Rapid admissions data could also help to identify COVID-19 cases in hospital by variant type or vaccination status which could help the UK’s policy response to the COVID-19 pandemic.

The benefits discussed (above) could be measured by the impact on policies and publications to contribute to an informed body of knowledge. The actions that lead to the benefits could be carried out by the combined project group, which includes the data controller, funder and third party researchers (subject to access approval via the sub-licence process).

The benefits are intended to be continual, to assist policy makers and researchers with the first outputs planned in Spring 2023 or as soon as possible afterwards.

Yielded Benefits is not a requirement for new applications.