Good TREs work
Health Data Research Uk projects
- R14.2 - CVD-COVID-UK. Cardiovascular disease and COVID-19: using UK-wide linked routine healthcare data to address the impact of cardiovascular disease on COVID-19 and the impact of COVID-19 on cardiovascular diseases. (partially via "system access")
135 data files in total were disseminated unsafely (information about files used safely is missing for TRE/"system access" projects).
R14.2 - CVD-COVID-UK. Cardiovascular disease and COVID-19: using UK-wide linked routine healthcare data to address the impact of cardiovascular disease on COVID-19 and the impact of COVID-19 on cardiovascular diseases. — DARS-NIC-381078-Y9C5K
Opt outs honoured: No - data flow is not identifiable, Anonymised - ICO Code Compliant, Identifiable (Does not include the flow of confidential data)
Legal basis: Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii), Health and Social Care Act 2012 - s261 - 'Other dissemination of information', Health and Social Care Act 2012 s261(1) and s261(2)(b)(ii),
Purposes: No (Academic)
Sensitive: Non Sensitive, and Sensitive, and Non-Sensitive
When:DSA runs 2020-07-01 — 2023-06-30 2020.07 — 2021.05.
Access method: System Access
(System access exclusively means data was not disseminated, but was accessed under supervision on NHS Digital's systems)
Data-controller type: UNIVERSITY COLLEGE LONDON (UCL), UNIVERSITY OF BRISTOL, UNIVERSITY OF CAMBRIDGE, UNIVERSITY OF LEICESTER, SWANSEA UNIVERSITY, UNIVERSITY COLLEGE LONDON (UCL), UNIVERSITY OF BRISTOL, UNIVERSITY OF CAMBRIDGE, UNIVERSITY OF LEICESTER, UNIVERSITY OF LIVERPOOL, NATIONAL INSTITUTE FOR HEALTH AND CARE EXCELLENCE (NICE), SWANSEA UNIVERSITY, UNIVERSITY COLLEGE LONDON (UCL), UNIVERSITY OF BRISTOL, UNIVERSITY OF CAMBRIDGE, UNIVERSITY OF LEICESTER, UNIVERSITY OF LIVERPOOL, NATIONAL INSTITUTE FOR HEALTH AND CARE EXCELLENCE (NICE), SWANSEA UNIVERSITY, UNIVERSITY COLLEGE LONDON (UCL), UNIVERSITY OF BRISTOL, UNIVERSITY OF CAMBRIDGE, UNIVERSITY OF LEICESTER, UNIVERSITY OF LIVERPOOL, UNIVERSITY OF OXFORD, KING'S COLLEGE LONDON, NATIONAL INSTITUTE FOR HEALTH AND CARE EXCELLENCE (NICE), SWANSEA UNIVERSITY, UNIVERSITY COLLEGE LONDON (UCL), UNIVERSITY OF BRISTOL, UNIVERSITY OF CAMBRIDGE, UNIVERSITY OF LEICESTER, UNIVERSITY OF LIVERPOOL, UNIVERSITY OF OXFORD, IMPERIAL COLLEGE LONDON, KING'S COLLEGE LONDON, SWANSEA UNIVERSITY, UNIVERSITY COLLEGE LONDON (UCL), UNIVERSITY OF BRISTOL, UNIVERSITY OF CAMBRIDGE, UNIVERSITY OF GLASGOW, UNIVERSITY OF LEICESTER, UNIVERSITY OF OXFORD, IMPERIAL COLLEGE LONDON, KING'S COLLEGE LONDON, SWANSEA UNIVERSITY, THE UNIVERSITY OF MANCHESTER, UNIVERSITY COLLEGE LONDON (UCL), UNIVERSITY OF BRISTOL, UNIVERSITY OF CAMBRIDGE, UNIVERSITY OF GLASGOW, UNIVERSITY OF LEICESTER, UNIVERSITY OF OXFORD
Sublicensing allowed: No
- Civil Registration - Deaths
- COVID-19 Second Generation Surveillance System (Beta version)
- Hospital Episode Statistics Accident and Emergency
- Hospital Episode Statistics Admitted Patient Care
- Hospital Episode Statistics Critical Care
- Hospital Episode Statistics Outpatients
- TRE Secondary Care Data
- COVID-19 Hospitalization in England Surveillance System
- GPES Data for Pandemic Planning and Research (COVID-19)
- Medicines dispensed in Primary Care (NHSBSA data)
- COVID-19 Second Generation Surveillance System
- Covid-19 UK Non-hospital Antibody Testing Results (Pillar 3)
- Covid-19 UK Non-hospital Antigen Testing Results (pillar 2)
- COVID-19 Vaccination Adverse Reactions
- COVID-19 Vaccination Status
- Electronic Prescribing and Medicines Administration (EPMA) data in Secondary Care for COVID-19
- Sentinel Stroke National Audit Programme (SSNAP)
- Civil Registration (Deaths) - Secondary Care Cut
- Improving Access to Psychological Therapies Data Set
- NICOR Heart Failure V5.0
- NICOR Myocardial Ischaemia National Audit Project v10.3.2
- NICOR National Audit of Percutaneous Coronary Interventions (NAPCI) V5.6.6
- Uncurated Low Latency Hospital Data Sets - Admitted Patient Care
- Uncurated Low Latency Hospital Data Sets - Critical Care
- Uncurated Low Latency Hospital Data Sets - Outpatient
- COVID-19 ICNARC Case Mix Programme for Adult Critical Care
The COVID-19 pandemic is a major global challenge, whose impacts on the population’s health, healthcare systems and services and the wider economy will be apparent for many years. Patients with pre-existing cardiovascular disease have a disproportionately elevated risk of symptomatic infection and mortality. The reasons are not fully understood but could be due to cardiovascular conditions per se (heart attack, heart failure, stroke etc.); their risk factors (e.g., age, obesity, raised blood pressure or cholesterol, diet and lifestyle) or medications (e.g. drugs to reduce blood pressure); or combinations of these. Understanding which patients with cardiovascular diseases are affected and why will be a major step towards developing strategies to reduce their risk.
Just as important as the effects of cardiovascular disease, and its risk factors and medications, on COVID-19 disease, are the direct and indirect impacts of COVID-19 on cardiovascular disease. The direct impacts include acute life-threatening complications, such as heart attacks, strokes and clots in the legs and lungs. In addition, since COVID-19 increases both inflammation and the risk of blood clots, there may be an increased risk of heart attack, stroke and other cardiovascular events in the medium and long term. However, the nature and extent of these direct effects is far from fully understood.
Crucially, there are also indirect impacts on the presentation, diagnosis, management and prognosis of cardiovascular diseases resulting from the response by governments and health services to the COVID-19 pandemic. For example, in the UK and elsewhere, the numbers of people attending hospital with acute myocardial infarction and stroke have declined dramatically, while patients who do present to hospital are frequently arriving too late to receive acute treatments proven to improve outcomes (e.g., clot removal or clot busting treatments) or after preventable complications have developed. Anecdotal reports from clinicians suggest that referrals to specialist outpatient services of patients with milder symptoms (e.g., mini strokes and angina), who would benefit from treatment to reduce the risk of more serious vascular events, have also plummeted. A deeper understanding of the nature and extent of these unintended consequences, including the range of conditions affected, variation by patient characteristics (such as age, sex, ethnicity, and deprivation) and geography (both within and between countries), effects on different in- and out-patient services and treatments, and changes over time in response to mitigating actions (e.g. regional and national government advice), is urgently needed to inform government and NHS policy.
The British Heart Foundation (BHF) Data Science Centre (which is embedded within Health Data Research UK, HDR UK) is working in partnership with NHS Digital to establish a Cardiovascular Disease Trusted Research Environment (CVD TRE) [service] for England to enable analyses of linked, nationally collated healthcare datasets. This project is entitled CVD COVID UK. It will:
• Enable timely research on the effects/impacts of cardiovascular disease on COVID-19, and the direct and indirect impacts of COVID-19 on cardiovascular diseases
• Coordinate similar approaches across the four nations of the UK
• Future proof an enduring CVD TRE service post-Covid19
Coordinated by the BHF Data Science Centre, the CVD COVID UK research group which will be made up by proposes to interrogate nationally collated, population level, linked healthcare datasets across the UK population to address the following questions:
1) What are the effects of cardiovascular diseases, their risk factors and medications on susceptibility to and outcomes from COVID-19 disease?
2) What is the direct impact of SARS-CoV-2 infection on acute cardiovascular complications and on medium and longer term cardiovascular risk?
3) What is the indirect impact of the COVID-19 pandemic and the government and NHS response to it on the presentation, diagnosis, management and outcomes of cardiovascular diseases?
The work would be organised into work packages (WPs) as follows:
WP 1: Coordination (led by BHF Data Science Centre):
Identify relevant datasets and required dataset linkages across the UK. Coordinate applications for relevant approvals and access. Coordinate specialist inputs from cardiologists, stroke physicians, vascular surgeons and other relevant clinical groups. Coordinate reporting to SAGE and equivalent bodies in the devolved nations via established HDR UK processes.
WP 2: Analyses:
Refine questions with appropriate clinical specialist input, draw up analysis plans for different datasets (individually and linked), assess data completeness and quality, conduct analyses, interpret results, iterative reporting and refining of analyses.
• WP 2.1 Indirect impact of COVID-19 on cardiovascular diseases: Time trends in hospital activity (admissions by diagnosis, treatments, procedures) using hospital and disease audit datasets, registered deaths by cause and primary care activity before, during and after COVID-19 pandemic
• WP 2.2 Influence/associations of cardiovascular conditions on COVID-19 outcomes (such as admissions to hospital, admission to ITU, mechanical ventilation and death)
• WP 2.3 Influence/associations of cardiovascular risk factors on COVID-19 outcomes
• WP 2.4 Influence/associations of cardiovascular medications on COVID-19 outcomes
• WP 2.5 Direct impact of COVID-19 disease on cardiovascular disease occurrence, re-occurrence and outcomes in short, medium and long term
WP3: Public, patient and professional involvement and communications:
Work with existing HDR UK and BHF public, patient and professional panels to provide input into refining questions, assessing the impact of the results, and preparing reports for lay audiences. Lead on communications of activity and emerging results through websites, social media and other outlets. Lead on interactions with press and other media.
A consortium of organisations as listed below are the data controllers and substantive employees of these organisations will have access to the record level data within the TRE:
University College London
University of Bristol
University of Cambridge
University of Leicester
A future amendment will be made to add the following as data controllers. These organisations will not have access to the data within the TRE until the agreement has been amended:
Imperial College London
Kings College London
London School of Hygiene and Tropical Medicine
University of Aberdeen
University of Dundee
University of Edinburgh
University of Leeds
University of Liverpool
University of Manchester
University of Oxford
Through addressing questions about the impacts of cardiovascular disease on COVID-19 and the impacts (both direct and indirect) of COVID-19 on cardiovascular diseases, BHF expect the outputs of this work to inform public health policy and clinical care, benefiting:
• patients with a history of cardiovascular disease (stroke, heart attack, peripheral arterial disease etc) who are at increased risk of poor outcomes with COVID-19 as a result of their cardiovascular condition, cardiovascular risk factors or cardiovascular medications;
• patients now and in the future who become unwell with COVID-19 and are at risk of short, medium and long term cardiovascular complications;
• the population as a whole whose cardiovascular health services are being affected by the government and health service response to the COVID-19 epidemic.
Outputs providing these benefits are expected to start to emerge within weeks of data becoming available to the research team. Outputs will continue to be produced and to provide benefits as outlined throughout the three year period of the project through to June 2023. By their nature, those outputs providing information on the longer term impacts and implications of these for clinical care and public health policy will take at least several months to start emerging.
The coordination work for WP1 will provide knowledge about linked health care datasets and routes to their access and so will benefit other UK-wide initiatives that require linkage to routinely collected healthcare data. These include: the ISARIC-CCP (UK-wide study of the clinical characteristics of patients hospitalised with COVID-19) and its CAPACITY-COVID extension; collaborative efforts to address the determinants of COVID-19 susceptibility, severity and outcome through analyses of population-based cohorts with bio-samples linked to routinely collected healthcare datasets; the RECOVERY trial; and COG-UK.
Finally, the research group is aware of initiatives in a wide range of other countries, including Italy, Sweden and Korea, that are starting to derive policy-relevant insights from analyses of routine linked datasets, especially on the indirect impacts of COVID-19 on cardiovascular diseases. Through the researchers’ connections with international consortia and organisations, including the European Society of Cardiology, European Stroke Organisation and others, the research group will engage in the bilateral sharing of emerging results, to learn from others as well as to maximise the international reach and relevance of findings.
The outputs of each piece of work will be reported to SAGE and equivalent bodies in the devolved nations as well as (especially in WP 2.4) to NICE, MHRA and Scottish Medicine Consortium, so helping to drive evidence-based policy decisions for health service providers and clinical professional groups. Outputs will also form the basis of manuscripts for publication in peer-reviewed scientific and medical journals, presentations at national and international scientific and medical professional conferences, and reports aimed at lay audiences, available through websites, in particular those of Health Data Research UK and the British Heart Foundation (it is anticipated that . Outputs will inform the clinical management of patients with different types of cardiovascular disease presenting with COVID-19 disease and cardiovascular prevention strategies for patients with COVID-19 and no prior cardiovascular disease.
All analysis plans, protocols and reports arising from this proposal will be made publicly available via the HDR UK website (linking to additional institutional documentation if appropriate), HDR UK github repository and open access publications. Hence all outputs will be freely available.
All reports for government advisory groups and policy makers, the lay public and academic publications will be written in the name of the collaborative group (CVD-COVID-UK) with all relevant individual contributions (coordination, writing, analysis, interpretation etc.) listed.
Outputs will contain only aggregate data with small numbers suppressed in line with the HES analysis guidance from NHS Digital, Public Health Scotland and the SAIL Databank for Wales.
The project will commence immediately and results of clinical and public health relevance are expected to start emerging within weeks of the start. For analyses of longer term outcomes (especially WP 2.5), BHF expect analyses to continue and for results to emerge over several years.
Work is ongoing across all four nations to identify and assemble the relevant national datasets, enable their linkage, agree mechanisms for regular updates and establish routes for expedited approval and access for approved researchers within trusted research environments in each of the four nations. A paper was presented to the UK government’s Scientific Advisory Group for Emergencies (SAGE) in mid-April (see https://www.hdruk.ac.uk/wp-content/uploads/2020/04/200416-COVID19-Research-Data-Final.pdf) by HDR UK, NHS Digital, the UK Health Data Research Alliance, national data custodians in Scotland, Wales and Northern Ireland, national providers of specialist cardiovascular data and the BHF. It was endorsed by SAGE on 15 April 2020 and provides details of datasets to be accessed across the four nations (primary care, community prescribing/dispensing, hospital admissions, intensive care, death registry, COVID-19 laboratory testing, and cardiovascular audit datasets), the trusted research environments within which analyses will be conducted and routes to approval.
The CVD COVID UK project has obtained research ethics approval, has been awarded COVID-19 national flagship research project status by the NIHR-BHF Cardiovascular Partnership COVID-19 prioritisation process and is in the process of being awarded Urgent Public Health prioritisation status by the NIHR. Applications are underway for researchers working on this project to access relevant linked national healthcare datasets within trusted research environments so that work can commence in June 2020. BHF will continue to work with data custodians across the four nations to enable access to an expanding number of linked datasets as these become available as well as on mechanisms for ongoing updates to these linked datasets.
Given that data custodians in each of the four UK nations will bring together and make available their datasets within separate trusted research environments (TREs), and given the differences between countries in the datasets available, the general approach for each analytic work package will be to develop a master protocol from which country-specific protocols will be developed, aiming for maximum consistency but allowing for country-specific differences in the datasets. Where appropriate, results of country-specific analyses will be combined in meta-analyses.
Analyses based on routinely collected, national healthcare datasets have the advantages of large scale and comprehensive coverage, maximising statistical power as well as inclusiveness/representativeness (e.g. across all age groups, ethnicities, geographies and socioeconomic settings).
Detailed analysis plans are already being drawn up as outlined below:
WP 2.1: An immediate priority for informing government policy across the UK is to assess the indirect impact of COVID19 on cardiovascular diseases. Analysis plans to address trends in acute coronary syndromes before and during the COVID-19 pandemic have been developed for England and will be extended to incorporate a broader set of conditions (stroke, heart failure, others) and data from the other UK nations (Scotland, Wales and Northern Ireland). BHF propose to make results of these analyses of trends, based on hospital admissions, mortality and disease audit databases available in rapidly produced reports for government advisory groups across the UK as soon as they start to become available (aiming to commence in June 2020) with regular updates provided thereafter as the coverage of conditions and geographies expands.
WP 2.2: The influence/associations of pre-existing cardiovascular diseases (ischaemic heart disease, stroke etc) on COVID-19 incidence and outcomes will be studied through linkage of large scale population wide datasets that contain information on previous medical history with COVID-19 test results, hospitalisation, critical care and mortality datasets, with adjustment for multiple confounders (including risk factors and co-morbidities).
WP 2.3: The influence/associations of cardiovascular risk factors on COVID-19 incidence and outcomes will be studied through linkage of large scale population wide datasets that contain information on cardiovascular risk factors such as blood pressure, body mass index and smoking status with COVID-19 tests, hospitalisation, critical care and mortality datasets, with adjustment for multiple co-morbidities. Preliminary work has started in Wales and will be developed and extended across the four nations.
WP 2.4: An immediate priority is to provide information to enable government agencies (e.g., MHRA and NICE) to give evidence-based advice to healthcare professionals and patients on drug regimens and risk of COVID. BHF will study the effects of ACE inhibitors, angiotensin receptor blockers and other cardiovascular medications on COVID-19 outcomes (hospitalisation, admission to ICU, mechanical ventilation and mortality) by using primary care physicians’ preferred antihypertensive drug class as an instrumental variable (IV). This approach will enable BHF to deal with the potential for confounding by measured and unmeasured factors, including the indication for which these drugs are prescribed, health-promoting behaviours associated with receipt of one of these medications, and the effects of drugs such as statins and anticoagulants which are often used in combination. BHF will compare the results of IV analyses with multivariable logistic regression analyses, controlling for measured confounding variables.
WP2.5: Linkage of population routine datasets (demography including mortality, primary care, hospital) and audit datasets will enable comprehensive assessment of the impact of COVID-19 disease on cardiovascular disease occurrence, reoccurrence and outcomes in short, medium and long term. With SARS-CoV2 potentially circulating for at least several years in the population, it will be important to estimate the short-, medium- and long-term effects of infection on incidence of stroke, MI, venous thromboembolic disease and other cardiovascular diseases. There are several potential mechanisms of increased long-term risk, including longer term pro-coagulant and inflammatory effects as well as the recognised increased risk of many chronic health conditions among survivors of intensive care. We will estimate the risk of MI, stroke, heart failure and other cardiovascular conditions associated with COVID-19, using a self-controlled case-series design with shorter and longer periods of risk, along with other epidemiological analysis approaches.
The analysis sub-work packages outlines above require access to linked data from the personal demographic service, primary care, hospital emergency, inpatient and outpatient care, intensive care, registered deaths by cause, cardiovascular disease-specific audits/registries and COVID-19 laboratory testing. All data will be accessed by named, approved researchers (certified to have successfully completed safe researcher training) in the Trusted Research Environment within NHS Digital.
The data within the TRE will be pseudonymised. NHS Digital will strip identifiers from each record, apply a pseudo-ID to each record and perform the data linkage. No identifiable data will be accessible within the TRE. Following similar principles to those adopted by the SAIL Databank for Wales and the Scottish National Data Safe Haven, only summary, aggregate results data will be exported from the TRE by approved researchers, subject to the approval of the NHS Digital team providing the TRE. The objective of this will be to ensure that no output contains information which could be used either on its own or in conjunction with other data to breach an individual's privacy.
In the first instance, the following linked datasets are required and will be available within the TRE:
• HES and SUS
These data are needed to provide information on admissions to hospital and outpatient appointments with diagnoses so that history of cardiovascular disease and future cardiovascular conditions and associated procedures can be ascertained.
Access to information on future cardiac conditions will provide prospective information on cardiovascular conditions which may occur as a direct or indirect consequence of COVID-19 in the short, medium and long term. Cardiovascular conditions in this context covers a broad range of conditions, including heart attack, heart failure, stroke, peripheral arterial disease, deep venous thrombosis and pulmonary embolism.
• Emergency Care Dataset
These data are needed to provide information on attendances for cardiovascular conditions before, during and (in due course) after the COVID-19 emergency.
• SGSS (COVID-19 laboratory test results provided to NHS Digital from PHE)
These data are needed to ascertain all cases with proven SARS-CoV-2 infection. Linkage of these data to data on hospitalisations, intensive care and mortality will be used to indicate the severity of COVID-19 disease.
• Mortality data
These data are needed to provide information on dates and underlying + contributing causes of death as part of the assessment of severity of COVID-19 disease and of the impact on cardiovascular diseases.
The following linked datasets are due to be onboarded and an amendment request will be made when available to access them within the TRE:
• GPES data for pandemic planning and research (provided to NHS Digital from GP system suppliers)
These data are needed to provide information on prior medical history (cardiovascular conditions and co-morbidities), other cardiovascular risk factors (age, sex, ethnicity, socioeconomic status, obesity, high blood pressure, high cholesterol, diabetes etc), and prescribed medications for those who have and have not gone on to develop COVID-19 disease with various different levels of severity. It is also needed to provide prospective information on cardiovascular conditions which may occur as a direct or indirect consequence of COVID-19 in the short, medium and long term.
• CHESS (COVID-19 Hospitalisation in England Surveillance System)
These data are needed to provide information on the severity and treatments of people with COVID-19 disease.
The data will be minimised:
~ to only include those datasets required to address the cardiovascular-related questions included within the agreement
~ only for organisations who are part of the BHF DSC/HDRUK CVD-COVID UK consortium
~ only for CVD-related research purposes, as outlined in the proposal
~ a “per project” basis (by dataset, by year, and by “groups” of fields rather than individual fields)
~ as an interim measure until these data minimisation techniques can be applied, projects will only be included if they are urgent (or do not require data minimisation beyond minimisation at the dataset level)