MR104C - British Women's Heart and Health Study (s251 cohort)

This Data Sharing Agreement permits the retention of the data provided under previous iterations of this Agreement for an interim period. This is a pragmatic approach to provide an active Agreement whilst enabling University College London to complete the necessary actions to enable a subsequent application to extend the Agreement meeting all applicable data sharing standards as published in NHS Digital’s website

(see: https://digital.nhs.uk/services/data-access-request-service-dars/dars-guidance).

The following provides background information on the purpose of the original study:

University College London (UCL) currently holds mortality data, cancer registration data and demographic data for use in the British Women’s Heart & Health Study (BWHHS).

The BWHHS has previously been managed by both the University of Bristol and the London School of Hygiene and Tropical Medicine but since 2015 has been managed by UCL at The Institute of Health Informatics. UCL is now the sole Data Controller and, for the primary purpose of the BWHHS, UCL is the sole processor of the data.

The University of Bristol and the London School of Hygiene and Tropical Medicine have no ongoing role in the study other than individuals from the respective organisations working as study collaborators with UCL in accordance with the arrangements and processes described below.

The BWHHS aims to determine the contribution of both established and new risk factors to the considerable variation in ischaemic heart disease and stroke in Great Britain. It is also concerned with the effects of risk factor changes and their impact on Cardiovascular Disease (CVD) events and on other common causes of morbidity and mortality in British women. The present aim is to continue to collect CVD and other common diseases of incident morbidity and mortality to provide information for the prevention and promotion of a disability-free life in older women.

The BWHHS is a prospective cohort study of cardiovascular disease in women aged over 60 years, in England, Scotland and Wales.

The study was set up in 1999 to complement the British Regional Heart Study (BRHS), to describe and establish risk factors and the differences in their impact in women compared to the men followed up by the BRHS.

The study selected women at random from 24 GP practices, in 23 towns from 1999 to 2000. Of the 7,296 invited, 4,286 (60%) were recruited and attended the baseline examinations and completed questionnaires. Follow up consisted of postal questionnaires and regular reviews of GP medical records. This agreement covers the release of data in respect of 1049 participants of the study where the legal basis to address the Common Law Duty of Confidentiality (CLDC) is under Section 251.

The BWHHS has used the patient tracking service provided by NHS Digital and predecessor organisations to receive notifications of its cohort member’s deaths (date and cause), cancer registrations, exits from the NHS and changes in recorded demographics (such as name, NHS number, etc.). The latest demographics data has been used for administrative purposes to support ongoing contact with participants (e.g. for sending questionnaires) and for requesting primary care data from GPs.

The BWHHS was set up to explore the current patterns of CVD & Cardiovascular Heart Disease (CHD) risk factors (and recent changes in this pattern), and prevention and treatment for CHD in older British women.

The BWHHS involves multiple ongoing analyses and investigations testing different hypotheses within this scope. The research activities are determined by the BWHHS study director on a monthly basis. The BWHHS study director is responsible for determining what analyses will be undertaken and what data will be used for each analysis in support of the objectives. All analyses are undertaken by members of the BWHHS team, all of whom are UCL employees working under supervision of the study director.

The clinical outcomes that are the current focuses of the BWHHS are: myocardial infarction, stroke, angina, heart-failure, diabetes, deep vein thrombosis and pulmonary embolism (DVT/PE), dementia, atrial fibrillation and cancers.

The main focus is the measurement of biological variables (biomarkers) using the biological specimens collected at baseline (in the form of DNA, plasma, and serum samples) and their effect on clinical diseases of relevance to post – menopausal women such as cardio-metabolic disease, dementia and common cancers. The continuing provision of mortality and cancer registration data increase the number of events accrued over time, which will increase the statistical power to BWHHS’s analysis.

Under this Data Sharing Agreement, the BWHHS team is not permitted to use the data to expand the focus of the BWHHS to non-cardiovascular conditions of relevance to post-menopausal women that the BWHHS currently does not collect.

Additionally UCL has made data from the BWHHS available to third parties subject to an approval process outlined below. NHS Digital has assessed what data is shared and determined that it is not sufficiently derived and therefore may not be onwardly shared without NHS Digital’s express permission.

Under this Data Sharing Agreement, UCL is not permitted to share any data supplied by NHS Digital or data derived from that data with any third parties or with UCL employees for any purposes other than the primary objectives of the BWHHS.

UCL has informed NHS Digital that it has previously shared pseudonymised data with the following organisations:

1. University College London – for the purpose of a UCL-LSHTM-Edinburgh-Bristol (UCLEB) Consortium

2. UCL investigators at the Department of Medicine

3. University of Bristol

4. University of Cambridge

5. London School of Hygiene & Tropical Medicine

6. Birkbeck, University of London

7. Universidad de Salamanca

8. University of Alcala

9. Imperial College London

No new data may be shared with these third parties. Under this Data Sharing Agreement, the above organisations are permitted to retain the data while UCL prepares and submits an application to NHS Digital for approval to onwardly share data under appropriate controls. The third parties are not permitted to onwardly share the data and may only use it for the purposes of the projects that were previously authorised by the BWHHS Study Director following the approvals process outlined in the ‘Processing activities’ section below. No changes to the existing purposes may be approved.

Data will only be used for academic research purposes and not commercially.

Findings from data have been used to inform policy and clinical guidelines and have also led to the development of prediction models available as web tools. Work conducted operates in the pre-competitive arena so does not contain patentable or commercially exploitable results.

UCL have identified the appropriate legal basis for processing under General Data Protection Regulation (GDPR). Based on the purpose for processing, the legal basis is 'processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller.' Article 6(1)(e). As the research involves health data, which is included in the definition of special categories of personal data, and requires an additional condition for processing. For health research this will be 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...'.

Under this Agreement, the data may be securely stored but not otherwise processed. No new data will be provided by NHS Digital under this Agreement.

The study data, including data provided by NHS Digital under previous agreements, are currently held by University College London.

No onward sharing of data to any third parties is permitted under this Agreement.

The following provides background on the processing activities undertaken prior to this Agreement:

This agreement permits continued retention of the data only. The agreement does not permit any other processing of the data.

Identifying details for study participants were supplied to NHS Digital’s predecessor organisation. These were matched to the patient entries on the NHS Central Register (since replaced by the Personal Demographics Service) and a ‘flag’ was added to each matched entry to facilitate ongoing event reporting.

NHS Digital (and predecessors) provided routine notifications of deaths, cancer registrations and exits from or re-entries to registration with the NHS along with the latest recorded demographic data for members of the cohort.

To ensure that type 2 patient objections can be appropriately upheld, NHS Digital purged the list of participants it currently holds and the cohort will be re-flagged in two subgroups.

UCL will supply to NHS Digital two separate lists of study participants: one listing the individuals who have given informed consent and one listing the individuals for whom section 251 support permits the processing of their personal data without informed consent. The lists will contain the following identifiers: NHS Number, Date of Birth, Surname, Forename, Postcode , Gender and unique study ID .

NHS Digital will trace the relevant patient entries for each participant and flag them as being part of the relevant subgroup. NHS Digital will return separate reports for each subgroup to UCL. These will list all individuals who were successfully traced and, for those covered by section 251 support, who had not registered a type 2 patient objection.

UCL will review the data it has historically received from NHS Digital (and predecessors) and if any data was provided for individuals who were not included in either of the lists UCL will send to NHS Digital (described above), UCL must securely destroy the data of those individuals.

NHS Digital will then provide routine notifications of cancer registrations, deaths and changes to status of registration with the NHS. These will include personal demographics data.

UCL stores the data on a server in its secure safe haven facility which can be remotely accessed at the UCL Institute of Health Informatics.

Data will only be accessed by individuals within the BWHHS team at the UCL Institute of Health Informatics who have authorisation from the study director to access the data.

Participant identifiers are stored in a table that is kept separate to the research dataset which contains all other collated study data. Access to the identifiers is restricted to a small number of authorised individuals – all of whom are substantive employees of UCL – and are used for administrative purposes only such as contact with participants and requesting primary care data from participants’ GPs. The information has only been managed (received, entered and stored) by four people during the 18 years of the study.

From the data provided by NHS Digital, the UCL staff (described above) produce derivations that are added to the separate research dataset which contains only pseudonymised data from multiple sources including data from the baseline interview, questionnaires, GP record reviews and biological variables measured from biological specimens. This dataset does not contain participants’ names, NHS numbers, Dates of Birth or Dates of Death. Instead, it contains variables such as participants’ age rather than Date of Birth and, rather than Date of Death or Date of Cancer Registration, it contains calculated time to event from the date of recruitment (baseline). The baseline date is not included. Full details of how data provided by NHS Digital is converted into data added to the research dataset are given below. All subsequent analyses use only the data in the research dataset.

Extracts from the research dataset are shared with UCL staff and external collaborators for use in analyses in support of the BWHHS objectives subject to a formal approval process (outlined below). Following an audit by NHS Digital in February 2017 the BWHHS team has developed a standard operating procedure for sharing of data. External researchers wishing to collaborate with the BWHHS team will have access to the standard operating procedure document which will be available on the newly updated website at the Institute of Health Informatics, UCL.

Prospective collaborators contact the BWHHS study coordinator to informally discuss the feasibility of the proposed project, to ensure minimum criteria are met and to discuss any questions regarding the application before passing onto the BWHHS Study Director.

The application requires the submission of the BWHHS Collaborator’s Request Form, completed and signed to confirm that prospective researchers have read the terms and conditions before initial approval by the BWHHS research team and the Study Director. The application requires the following information:

• Name, affiliation and contact details of the principal investigator, who should be the main applicant.

• Project Title.

• Aims – should be focused and as specific as possible.

• Higher degree – For those that do not involve a BWHHS researcher (i.e. UCL employees working with

BWHHS and with approval to access BWHHS data and will be ONS approved researchers and have

attended information governance and data safe haven training.), the main applicant should be the

Principal Investigator or PhD supervisor. These applications will be considered as a regular data sharing

project, no additional supervision should be expected from any BWHHS researcher.

• Details of funding – The applicant should specify whether they are seeking funding to carry out the

project and if the grant application is partially or totally based in the use of BWHHS data.

• Variables – list of exact variable names and state from which data collection. Variables requested must be

consistent –with the project proposal.

• Biological samples required.

• Ethical issues.

• Timing of study.

• Signature and date to confirm that the applicant has read and will comply with the terms and conditions.

All requests are reviewed by the BWHHS study coordinator followed by a review with the BWHHS Director. The Director may deem it necessary to hold discussions with other members of the BWHHS team, before deciding whether to approve the application. Where necessary external peer review will take place.

BWHHS’s internal review process will consider:

• That there is no overlap – internal projects

• Researchers have the necessary skills

• Bona Fide Researchers - BWHHS will ensure that collaborators have conducted high quality, ethical

projects for research purposes using rigorous scientific methods. They must also have a formal

relationship with a bona fide research organisation, which is an established academic institution,

research body or organisation with the capability to lead or participate in high quality, ethical research.

• Investigators are based and data is stored in the European Economic Area

• That request is adequate – data requested must consist of the variables necessary to answer the research

question correctly and to a high quality.

• That requested data is relevant to the proposal

• It is not an excessive request –only data necessary to answer the specific research question for which the

request was made will be provided.

• Data is only intended to be used for the aims highlighted in the request form for a specific research

question. If future research questions arise, the investigators will complete a new request for

collaboration.

• No intention to pass on the data to a third party not named in the request

• That request follows BWHHS’s rules of data pseudonymisation and data release as detailed below.

Once the completed and signed proposal has been reviewed and authorised by the PI, the study co-ordinator for BWHHS will be responsible for compiling the pseudonymised dataset.

Any project involving external collaborators is recommended to actively involve at least one member of the BWHHS team at UCL. Further use of data outside those specifically stated in the application form will require a new application being submitted and approval from the BWHHS team.

Sharing the data with collaborators allows wider research, maximising its value. Data shared would include only necessary variables for the research question from any of the sources including, event data from GP records, participant self-reported data, data from biological samples, data from the baseline medical examination and the derived variables from NHS Digital data. No NHS Digital data is shared with collaborators.

The data shared with collaborators complies with the following standards:

The BWHHS does not share the following variables: individual identifiers, names, addresses, postcode, telephone numbers, NHS numbers, GP details, place of date of birth. BWHHS will also not share data on low frequency events (n<30), from which an individual may be re-identified.

The BWHHS does not share any sensitive data including occupations or information on mental health.

The following restrictions have been applied to the data released to all users.

1. Records are pseudonymised and identified by a unique BWHHS numerical ID.

2. All datasets will be stripped of specific variables that can create a risk of participant identification including:

• Date of Birth

• Date of clinical events – This includes date of recruitment (baseline)

• Any variable with a low prevalence

3. Index of multiple deprivation and super output area variables are only used by members of the BWHHS team.

4. ICD codes are hidden behind generic labels such as MI or stroke. A separate variable will specify whether cause of death is underlying, direct or both. The following describes the data received from NHS Digital and shared and an explanation of the derived variables created before sharing.

• Date of Birth – Shared as Derived Variable By subtracting baseline dates from birth dates and dividing by

365.25 BWHHS get the age at baseline. Age in years at baseline is shared.

• ICD-codes for direct & underlying causes of death - Shared as Derived Variable Using the event specific

ICD10 codes, variables are created for the specified outcome classifying them as an underlying cause or

direct cause. If both are true then classification is for Underlying + Direct. If neither are true the

outcome is not a cause of death. Result on whether the outcome played a role in the cause of death

either as a direct (D) or as an underlying (U) cause or as both (D+U) is shared.

• Date of Death – Shared as Derived Variable by subtracting each death date from baseline dates BWHHS

get the number of days to death after the baseline date. Number of days from baseline to death is

shared.

5. Instead of date of birth, age at baseline is provided in years, but not the specific date it was collected (actual date of birth falls in a range of 40 months).

6. Dates of clinical events are not provided; instead this will be time to event as days of event from or to

baseline.

7. Data will be transferred through the Data Safe Haven file transfer mechanism only.

This data will be stored in an Access database where all other BWHHS data is held. Only authorised researchers will have access to this data for the purpose of analysis for publications. All staff working with this data will undergo compulsory annual training in information governance run by UCL Information services Division.

Prior to transfer of the dataset another BWHHS researcher will check the dataset contains authorised variables only, and that personal information has not been mistakenly included. A record of this check is made on BWHHS’s collaboration record along with the collaborators details, project details, date data set is generated and date dataset is confirmed to be destroyed.

All organisations party to this Agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract - i.e. employees, agents and contractors of the Data Recipient who may have access to that data).

This Agreement permits the secure retention of the data only and no other processing.

No new outputs will be produced under this Data Sharing Agreement.

The BWHHS has led to 218 publications in peer reviewed journals to date addressing the aims outlined above.

The BWHHS typically produces manuscripts. Those working on the study do not typically engage with the end users of the study’s findings. UCL has an infrastructure for communications and determines the dissemination strategy for outputs on a case by case basis depending on factors such as the quality and impact of the manuscript. UCL’s press office may engage with the media and the public. There are examples on the webpage.

The following manuscripts were delivered within the last year:

• "Causal Associations of Adiposity and Body Fat Distribution With Coronary Heart Disease, Stroke

Subtypes, and Type 2 Diabetes Mellitus: A Mendelian Randomization Analysis” was published in

Circulation in 2017.

• “Investigating the importance of the local food environment for fruit and vegetable intake in older men

and women in 20 UK towns: a cross-sectional analysis of two national cohorts using novel methods” was

published in Int J Behav Nutr Phys Act. 2017”

• "Functional Analysis of the Coronary Heart Disease Risk Locus on Chromosome 21q22." was published in

Disease Markers in 2017.

• "Identifying low density lipoprotein cholesterol associated variants in the Annexin A2 (ANXA2) gene."

was published in Atherosclerosis in 2017.

• "Optimising measurement of health-related characteristics of the built environment: Comparing data

collected by foot-based street audits, virtual street audits and routine secondary data sources." was

published in Health Place in 2017.

• A manuscript titled "PCSK9 genetic variants and risk of type 2 diabetes: a mendelian randomisation

study." was published in Lancet Diabetes & Endocrinology in 2017.

• A manuscript titled "HAPRAP: a haplotype-based iterative method for statistical fine mapping using

GWAS summary statistics." was published in Bioinformatics in 2017.

• A manuscript titled "Worldwide trends in blood pressure from 1975 to 2015: a pooled analysis of 1479

population-based measurement studies with 19.1 million participants” was published in Lancet in 2017.

• A manuscript titled "Challenges of monitoring global diabetes prevalence." was published in Lancet

Diabetes & Endocrinology in 2017.

Data from the BWHHS was used for PhD training. In the last year 2 PhD thesis that used BWHHS data were awarded:

“Barcoding’ Cardiovascular risk: Predicting cardiovascular disease in patients with systemic lupus erthematosus (SLE)” was completed in 2017.

“Mitochondrial DNA copy number as a phenotypic trait for human diseases in genetic epidemiological studies” was completed in 2017.

The following outputs will be produced:

Scientific papers are being written to investigate:

1) What components of the physical environment (e.g. access to green space) where participants of the BWHHS live affect lifestyle behaviours (e.g. physical activity) known to affect the risk of cardiovascular disorders. This will improve understanding of how the modification of the physical environment could positively impact the cardiovascular health of older women in the UK. The analysis has been completed and the manuscript is currently being drafted. Once completed, it will be submitted in the coming year to peer-review journals such as International Journal of Epidemiology.

2) What substance in the blood (called metabolites) are responsible for correlation of the high levels of good-cholesterol (called HDL-cholesterol) with cardiovascular health. This work will help to inform development of new medications that by raising the levels of the good-cholesterol may lead to an improve cardiovascular health. The analysis has been completed and the manuscript is currently being drafted. Once completed, it will be submitted in the coming year to peer-review journals such as Circulation.

3) Best ways to identify if someone will develop in the future a heart attack or a stroke. This is through the analysis of thousands of substances in the blood (what scientists call proteomics) and is intended to be more accurate than current ways used in clinical practice by GP’s and specialists. The analysis will commence in November (once new data is received) and it is estimated that the analysis will be complete and the manuscript ready for publication by mid-2018. The manuscript will be submitted to peer-review journals such as Circulation.

4) New drug-targets for prevention of cardiovascular disorders. By combining the information on the genetic information of BWHHS participants with the levels of substances in the blood (called proteomics and metabolomics), it is expected that targets for new medications aimed to prevent the occurrence of heart attacks or stroke will be identified. This strategy is called Mendelian randomization. Future work will use this strategy extensively.

All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

This Agreement permits the secure retention of the data only and no other processing.

The benefits are labelled in a way that match specific outputs to be produced (section 5c, above)

1) The BWHHS has on-going work on neighbourhood deprivation focusing directly on government policy to narrow the gap between the most deprived areas and the rest of the country, with the aim to provide a tool to explore the specific attributes of the built environment that affect health in elderly people (work to be completed by end of 2018).

2) BWHHS’s work on metabolomics and proteomics will help to discover which substances in the blood should be measured (called biomarker of therapeutic efficacy) to better inform the development of new medications. The expectation is that this work will help to discover such blood markers that will help to produce new medications to raise the levels of the good-cholesterol (HDL-cholesterol) and hence to improve cardiovascular health (work to be completed by end of 2018).

3) As new cardiovascular medications, still under-patent (e.g. PCSK9 inhibitors), are being adopted by healthcare systems world-wide, there will be a resurgence on risk-prediction models. BWHHS’s work on metabolomics and proteomics and cardiovascular disease is expected to identify new ways to know who in the future will suffer from a cardiovascular disease, this is called risk prediction (work to be completed by end 2020).

4) BWHHS’s work on Mendelian randomization confirm the potential of a genomic lead strategy to not only identify and validate new drug-target , but also to discover more adequate biomarkers of therapeutic efficacy and overall help to optimise the drug discovery process (work to be completed by end of 2020).

The BWHHS has 218 publications within peer reviewed journals to date addressing the aims outlined previously in this application. Many of the journals for BWHHS manuscripts are considered high ranking journals. This “ranking” of journals is measured by something called the ‘impact factor’ which reflects the frequency with which the average article in a journal has been cited within the year. 31% of the publishing journals for BWHHS manuscripts have been published in journals with a high-impact factor (>8). This means that the output is published in journals that are widely read by doctors, scientist and public health practitioners, ensuring a greater impact of the work.

The BWHHS has also had measurable benefit directly as fourteen of the 218 publications have contributed to the following fifteen clinical care and public health guidelines:

1. National Clinical Guideline Centre NICE clinical guideline CG181 Lipid modification (2014)

2. Diabetes, Pre-Diabetes and Cardiovascular Diseases developed with the EASD ESC Clinical Practice

Guidelines (2013)

3. Dyslipidaemias 2016 (Management of) ESC Clinical Practice Guidelines (2011)

4. Dyslipidaemias 2016 (Management of) ESC Clinical Practice Guidelines (2016)

5. Arterial Hypertension (Management of) ESC Clinical Practice Guidelines (2013)

6. CVD Prevention in Clinical Practice (European Guidelines on) (2016)

7. Factors Influencing the Decline in Stroke Mortality: A Statement from the American Heart

Association/American Stroke Association (2013)

8. Genetics and Genomics for the Prevention and Treatment of Cardiovascular Disease: Update A

Scientific Statement From the American Heart Association (2013)

9. Guidelines for the Prevention of Stroke in Patients With Stroke and Transient Ischemic Attack: A

Guideline for Healthcare Professionals From the American Heart Association/American Stroke

Association (2014)

10. Update on Prevention of Cardiovascular Disease in Adults With Type 2 Diabetes Mellitus in Light of

Recent Evidence: A Scientific Statement From the American Heart Association and the American

Diabetes Association

11. Social Determinants of Risk and Outcomes for Cardiovascular Disease: A Scientific Statement From the

American Heart Association

12. Future Translational Applications From the Contemporary Genomics Era: A Scientific Statement From

the American Heart Association

13. Basic Concepts and Potential Applications of Genetics and Genomics for Cardiovascular and Stroke

Clinicians: A Scientific Statement From the American Heart Association

14. Salt Sensitivity of Blood Pressure: A Scientific Statement From the American Heart Association

15. Preventing and Experiencing Ischemic Heart Disease as a Woman: State of the Science. A Scientific

Statement from the American Heart Association