The Million Women Study

The University of Oxford requires access to NHS England Data for the purpose of the following research programme:

The Million Women Study (MWS)

The following is a summary of the aims of the research programme provided by the University of Oxford:

Participants of the MWS were initially asked to complete a questionnaire between 1996 and 2001 about their health and lifestyle, childbirth history, and contraceptive pill and menopausal hormone therapy (MHT) use. Participants also provided signed consent for follow up through their screening and other medical records. Postal questionnaires have been sent to all study participants around 3, 8, 12, 15 and 20 years after recruitment, to keep track of participants’ health, to ask some new questions, and to keep participants informed of study progress. Selected women (either randomly selected or on the basis of self-reported disease) have also provided answers to online questionnaires, e.g., on diet, and to disease specific questionnaires, e.g., for cervical cancer and breast cancer. Tumour tissue samples are also being collected. In addition, since 2006, some women have been recruited into a related study, the Million Women Study: Disease Susceptibility in Women study, which involves collection of blood and saliva samples for biochemical and genetic analyses and has separate ethical approval. More recently, a small number of women have been asked to take part in another ethically approved MWS related study, Changes to Lifestyle in Retirement.

While the initial stimulus of the MWS was to investigate use of different types of menopausal hormone therapy (MHT), sometimes also referred to as Hormone Replacement Therapy (HRT), and risk of breast cancer, the study was designed from the outset to investigate the role of many other factors (such as smoking and obesity) and other health outcomes. The primary aim of the study is to investigate common, largely modifiable, risk factors for serious and common diseases in women.

The MWS research agenda is peer-reviewed every few years as part of applications for renewed funding. The program of research is determined by the MWS Principal Investigators based at the Cancer Epidemiology Unit (CEU) within Nuffield Department of Population Health (NDPH) at the University of Oxford, with advice from the MWS Advisory Committee and from the MWS Participant Panel.

The research strategy takes into account:

- the importance of age-appropriate research questions, in relation to new evidence and to public health and participant priorities;

- the availability of sufficiently long follow-up and number of cases for less common diseases;

- additional information which becomes available for linkage, and which may allow new analytic approaches;

- in light of the COVID-19 pandemic, risk factors for health, and healthcare systems outcomes, from COVID-19.

Data provided by NHS England enable unbiased and virtually complete follow-up of the cohort for health outcomes and support continued contact with participants. Over time, the study has developed a comprehensive data resource to enable MWS Principal Investigators and their research team to discover which risk factors contribute to the development of particular diseases, such as various types of cancer or cardiovascular disease and which do not. This database is also a rich UK resource for other bona fide investigators worldwide to study women’s health.

As MWS possess a large cohort, it is not feasible for the MWS to follow up participants directly on an individual basis and so the Data provided via NHS England is crucial to the viability of the MWS, as the primary source for follow up of the cohort in England.

There are several objectives for processing, detailed below:

1. Primary objective: Million Women Study research and research into related studies that fit within the MWS programme of work and the purposes described in this DSA. This research will only be conducted by University of Oxford employees and students affiliated with the University of Oxford.

2. Second objective: Collaborative research with researchers outside the MWS study team for purposes within scope of the MWS Programme. All Data processing for these collaborations will occur within the University of Oxford, and only University of Oxford employees and students will have access to the Data. Collaborative research will be in line with the MWS research aims listed in the MWS Protocol and will be subject to internal review and approval by MWS investigators. Collaborative research will require additional approval from relevant ethics committees and/or CAG where appropriate.

Collaborations may include external researchers who can provide an element of expertise not available within the MWS study team. Where external collaborative researchers require access to MWS data containing NHS England Data, a standard Collaborative Agreement between collaborating bona fide institutions will be required, as described in the Third Objective below.

The process for internal data requests is as follows:

• Informal discussion with Principal Investigators to discuss data analysis or project

• Complete Data Access Application Form

• MWS data access administrative checks

• Confirm that applicant and team have completed necessary data security training

• MWS PIs review applications to ensure that they meet the criteria stated in the MWS Protocol and approve if appropriate

• MWS PI’s review the data fields selected to ensure that they include only those that are required to answer the research question in the application

• Final decision communicated

• Details of data sharing recorded on internal spreadsheet

• Pseudonymised dataset released to applicant

3. Third objective: Sharing of linked NHS England Data under sub-licence with external researchers (through Open Access Data Sharing) and collaborators (through formal Collaborations) from bona fide organisations worldwide. The MWS wish to make it possible for other researchers to apply directly to MWS for datasets which include linked Data provided to the study by NHS England, to streamline the application process for all parties and thus to maximise the health research benefit to be obtained through wider use of study data. As the primary source of health outcomes Data, the Data provided by NHS England is an essential and integral part of the study data resource and is required to support full data sharing. The value of the resource is in the combination of routinely collected NHS Data with information obtained directly from participants on medical and lifestyle factors (such as childbearing history, full lifetime history of smoking and alcohol intake, and social factors) which are not collected, or not readily available, through NHS central sources. The combined study datasets therefore provide added value to the NHS England data included. NHS England Data will only be onwardly shared where it has been linked with other data for the purpose of the MWS study, and will only be shared for the purposes described in the signed sublicensing agreement.

(i) For collaborative projects where data sharing is required, e.g., where a collaboration between the MWS and a bona fide institution providing specialist expertise to the MWS research is approved by the MWS Investigators, and analyses need to be performed by a specialist team external to NDPH. In this scenario, for projects approved by the MWS Investigators, pseudonymised NHS England Data are processed within NDPH by the MWS team together with linked MWS data to create pseudonymised, minimised analysis datasets for sharing, and subsequently for use by the external researchers in their institution. Collaborations may include commercial providers of expertise which is not available via academic institutions.

(ii) For approved project applications for Open Access Data from researchers at bona fide institutions, for non-collaborative research, data are processed within NDPH by the MWS team together with linked MWS data to create de-identified minimised analysis datasets for sharing, and subsequently for use by the external researchers in their institution. Derived data may be returned to the MWS by external researchers, to enhance the MWS data. This will be granted under a Data Use Agreement with sub-licence terms and conditions signed by both institutions.

For further details on the sublicensing process, please see heading ‘SUBLICENSING’ later in this section.

The following NHS England Data will be accessed:

> Hospital Episode Statistics Admitted Patient Care (HES APC) and Civil Registration Mortality – necessary to provide the key mechanism for identification of relevant health outcomes. These include cardiovascular disease, metabolic and endocrine disorders, neurodegenerative disease, cancers, stroke, musculo-skeletal conditions, mental health and any other conditions relevant to this population.

> Demographics – necessary to enable the MWS to be able to maintain accurate information about participants’ contact details including postal address and registered GP practice. These details were collected for the whole cohort at the time of recruitment but there will have inevitably been changes since then.

> Mental Health – necessary to enhance analyses for dementia and other mental health conditions, by increasing the completeness of ascertainment of disease. This Data will allow identification of those in the study who have a relevant diagnosis recorded in the MHSDS through use of mental health services, but not in HES APC as they have not had a relevant hospital admission.

>NDRS Cancer Registrations- necessary to enhance the analyses for tumour characteristics, allowing for more informative estimates on specific cancer subtypes and to provide the key mechanism for identification of relevant health outcomes

The level of the Data will be:

Identifiable – necessary because;

> Details of changes to participants’ demographics, postal addresses or registered GP practices are required to ensure MWS have up-to-date contact information required to follow up participants.

> Up to date details of deaths and of identifiers such as name are needed to ensure that, when re-contacting participants, MWS minimise the risk of mail being sent to recently deceased participants or to out of date addresses.

> Ongoing approved linkages to, and current applications for, participant level primary care data (CPRD) require accurate information about registered GP practice to ensure that data about a particular participant are requested from the correct practice. Primary care data adds value to the HES Admitted Patient Care (APC) Data: comparisons allow estimation of the completeness of coverage by HES for disease ascertainment, and for some conditions diagnosed and treated mainly in primary care, are the main source of outcome.

The Data will be minimised as follows:

> Limited to a study cohort of 1.3 million women who consented to participate. The cohort contains a range of women and is not confined to individuals with specific diseases or risk factors. This allows the MWS to support research into the causes of a wide range of diseases affecting women in middle and older age.

> To assess the relationship between risk factors and disease, the MWS research team need to consider any pre-existing conditions and confounding factors. For this reason, the study requires a complete picture of the medical history of the participant going back as far as NHS linked records permit, as well as new diagnoses going forward. To maintain accurate full records for study participants, Data updates may be provided as full record datasets from start of available record to updated end date.

> The MWS requires linkage to the complete records in all required datasets. With incomplete linkage, or with follow-up using only other methods such as participant report, there would be a high risk of the results being biased by selective lack of follow-up e.g., women who have died and many who are ill cannot complete follow-up questionnaires.

> Datasets provided for external use are prepared by a small number of named individuals, all of whom are substantive employees of the University of Oxford. MWS PI’s review the Data fields requested to ensure that they include only those that are required to answer the research question in the application. Identifying details including NHS Number, Names, full Dates of Birth, Address and Postcode are removed and replaced with a unique project specific patient ID. The University of Oxford will endeavour to provide variables to the level of detail that is the minimum required e.g., age at diagnosis may be provided rather than date when this is considered adequate.

> A historic resupply of the Data is also required to check that there have been no changes since the last time the Data was provided. Although the Data should be static, this may not always be the case – especially as there have been changes to the process of providing HES Data over time. The previously provided historic Data will be destroyed once both sets of the data have been compared.

The University of Oxford is the research sponsor and the controller as the organisation responsible for ensuring that the Data will only be processed for the purpose described above.

The lawful basis for processing personal data under the UK GDPR is:

Article 6(1)(e) - 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.

The lawful basis for processing special category data under the UK GDPR is:

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.

This processing is in the public interest because the study is directly related to population health benefits which could offer benefits to the public.

The funding is provided by Cancer Research UK (CRUK). The funding is for the programme of work and is not specifically limited to the study described.

The funder(s) will have no ability to suppress or otherwise limit the publication of findings.

Data will be accessed by:

> Substantive employees of the University of Oxford

> Students affiliated with the University of Oxford (approx. 12 PhD students and 5 MSc students per calendar year). Research students are assigned at least 2 supervisors who will be regularly reviewing their work.

> External researchers and collaborators under approved sublicensing agreements

Participants are kept informed of MWS activities through regular mailings and through their website and publications. During the past 22 years since participants joined the MWS, fewer than 200 of the 1.3 million participants have withdrawn from the study and asked that their information be deleted. Advice and feedback on data use are provided by the MWS Participant Panel and by the study’s Advisory Group, which has independent representation.

SUBLICENSING

Under the “sub-licensing model”, NHS England shares Data with the MWS, who are in turn licensed to share these Data with other worldwide organisations, subject to agreed controls, scoped in this agreement between NHS England and the University of Oxford. Worldwide sublicensing will be restricted to organisations within the European Economic Area (EEA), countries considered by the EU to provide an adequate level of data protection; and the United States of America. In line with this onward sharing model, the data sharing controls in place between NHS England and University of Oxford are replicated between University of Oxford and the other organisations. The University of Oxford is fully accountable for the actions of the parties involved in subsequent data share and use. The Data Use Agreements between the University of Oxford and the applicant’s institution mirrors the Data Sharing Framework Contract in place between NHS England (formerly NHS Digital) and the University of Oxford. It includes information about the research proposal, benefits to health and/or social care, organisational security assurance and terms and conditions regarding onward sharing of Data, responsibilities and processing activities etc. NHS England will retain the ability to directly audit the sub-licensee’s compliance with the outlined and agreed data access arrangements.

Only NHS England Data, which has been linked with other data for the purpose of the MWS study, will be onwardly shared.

Most data sharing requests require individual-level personal data. Shared datasets are de-identified but retain shared-dataset-specific pseudo-ID codes to which specified MWS investigators have the key (to provide revision/additions to shared datasets, and to integrate returned derived variables into the main database for future research use).

Data sharing is strictly managed according to the MWS Data Access Policy, in compliance with the NDPH Information Governance and Data Access Policies, and under the oversight of the NDPH Data Access External Oversight committee. Applications are accepted only from bona-fide researchers for research intended to deliver measurable benefits to the health care system in the UK. Applications are reviewed by The Data Access Committee, with advice from the MWS Advisory Committee and, as required, from the NDPH Data Access External Oversight Committee.

Process for Applications for MWS linked Data via sub-licence:

1. Informal Approach to Principal Investigators to check feasibility

2. Complete Data Access Application Form

3. MWS data access administrative checks

a. Confirm applicant bona-fides for health research

b. Confirm legal status for data processing

c. Check data security arrangements one or more of:

i. ISO 27001 Accreditation

ii. Data Protection Security Toolkit (DSPT)

iii. System level Security Policy (SLSP)

4. Initial decisions on data access are the responsibility of the MWS Data Access Applications Review Panel (which includes the study PIs and other members of the MWS research team) at monthly data access request meetings

a. Decision in principle

5. Further advice will be sought (if required) from the MWS Advisory Committee, which includes independent members

6. NDPH External Oversight Committee: notification and advice

a. Appeal decision (if required)

7. Final approval or rejection communicated

8. Details of data sharing published on public facing website

Before approval is given for data transfer, data requestors are required to provide evidence to the study team and to the University Information Compliance Office of data security measures in place within their institution. Where the data requested includes linked Data supplied by NHS England, the University will ensure that data security measures are compliant with those required under this agreement by NHS England. Datasets are securely shared with third parties under University of Oxford Material Transfer Agreements (MTA), known also as Data Use Agreements, by transferring encrypted data and telephoning the password to the individual receiving the data. The details of the data transfer and expiry date are also logged onto an internal data transfer log with details of the sender and recipient.

All sublicense applicants that request data must sign a Data Use Agreement with the University of Oxford and agree with the terms stated in the Data Use Agreement. By signing these agreements, the applicant agrees to adhere to these terms and conditions and is also reminded of the penalties they are likely to incur if they do not comply with the terms they have agreed. A list of shared fields and purpose will be uploaded on a public facing website (either NDPH or MWS).

The Data Use Agreement (sub-licence) between The University of Oxford and the applicant’s institution, states the conditions for the data to be used which include:

• Sub-licence granted is non-exclusive, personal and non-transferable

• The sub-licensee accepts liability for the use of the Licensed data

• Data only to be used within permitted territory of use

• Data to be used for the Permitted Purpose only, with appropriate acknowledgement in publications of NHS England and MWS as the source of the data

• Sub-licensee has appropriate IT and data protection/security measures and agrees to maintain data confidentiality

• Sub-licensee agrees to be monitored for compliance by Oxford and/or by NHS England

• The sub-licensee must make no attempt to re-identify individuals in the dataset and must report any inadvertent re-identification to the study team/OU DPO immediately.

• Regular annual progress reports to be communicated

• Audits to be performed by University of Oxford and/or NHS England

• Publications arising from data sharing to be communicated to the MWS

• Dataset to be destroyed at end of approved study, or by stated maximum term of MTA.

In some cases, where it is deemed by the Study Investigators that sharing of MWS data will contribute to a research collaboration (usually in organisations where specialist expertise is required to interpret results) a Sub-licence Data Use Agreement with Collaboration will be required to be signed by both institutions. In these cases, the Data Use Agreement with Collaboration will include the terms of the Sub-licensing Agreement (Data User Agreement) stated above.

In line with this onward sharing model, the data sharing controls in place between NHS England and University of Oxford are replicated between University of Oxford and the other organisations. The University of Oxford is fully accountable for the actions of the parties involved in subsequent data share and use. The Data Use Agreement between the University of Oxford and the applicant’s institution mirrors the Data Sharing Agreement in place between NHS England and the University of Oxford. NHS England’s controller requirements are flowed down to the sub-licensees.

Data shared under sub licence must not be used for any marketing, insurance, or sales purposes.

The expected outputs of the processing will be:

> Reports of findings to government regulatory bodies (submitted infrequently as and when there is a relevant finding to report)

> A Milestone Report to the funder, CRUK

> Submissions to peer reviewed open-access journals such as the Lancet and Lancet group, International Journal of Epidemiology, Neurology, Journal of the American Medical Association (JAMA) and the Journal of the National Cancer Institute (JNCI). At least 20 full peer-reviewed publications are planned, to include topics such as risk factors for dementia; diet; alcohol and risk of cancer; the role of co-morbidity in cancer patients; mammographic density and breast cancer risk; and risk prediction models for breast, ovarian and gastrointestinal cancers

> Presentations at annual/biennial conferences e.g., National Cancer Research Institute, Society for Social Medicine and Population Health, Alzheimer’s Association International.

> Presentation of research findings to study participants and to the public. E.g., talks planned for U3A; 2023, WI; Participant Panel meetings, held annually in Oxford.

The outputs will not contain NHS England Data and will only contain aggregated information with small numbers suppressed as appropriate in line with the relevant disclosure rules for the dataset(s) from which the information was derived.

The outputs will be communicated to relevant recipients through the following dissemination channels:

> Journals

> MWS Study website updates to provide links to open access papers and offer free downloads of accessible summaries of findings. Study websites will also be updated to include documentation on the use of study data, including NHS England Data, via onward sharing

> Press/media engagement. Recent media engagement has been via the British Broadcasting Corporation (BBC), the Telegraph, and the Daily Mail

> Public outreach work (e.g., science open days, talks to participant groups)

> Patient Information Newsletters, provider to MWS participants monthly

> Social Media: X handles at NDPH (@Oxford_NDPH) and Oxford Medical Sciences Division (@OxfordMedSci) with a combined following of over 11,000 will promote all significant findings

> Sharing of research experience with UK Government and other research bodies and NHS Data providers. Contributions through NHS England Research Advisory Group, UK Biobank working groups, the Dementias Platform UK and the CLOSER longitudinal studies consortium, and UK cohort COVID19 working groups

> Results will also be communicated to relevant committees, e.g., funding committees and the National Institute for Health and Care Excellence (NICE)

The study has been running for over 20 years and publications are produced on an ongoing basis, with currently planned outputs expected throughout 2023-2025.

The MWS research using linked health data has already directly influenced health care data as evidenced by the past outputs:

1. Beral V et al. Breast cancer risk in the interval between menopause and starting hormone therapy.

This study provided evidence to suggest that Women using hormone replacement therapy are at increased risk of breast cancer; confirming and extending the Women's Health Initiative (WHI) clinical trial results published in 2002. These finding were shared with the Medicines and Healthcare Regulatory Agency and helped inform changes in prescribing guidance both for the UK and elsewhere. It is estimated that tens of thousands of cases of cancer worldwide have been avoided because of the subsequent fall in use of hormone therapy

Subsequent papers have extended and updated findings providing clarity on timing of use and long-term risks of past use.

2. Sweetland S, et al. Smoking, surgery, and venous thromboembolism risk in women: United Kingdom cohort study.

One of several papers from 2009-2013 using linked NHS data to quantify blood clot risk in relation to surgery along with other risk factors. The paper showed that risk of blood clots after surgery was far higher, and lasted for much longer, than had been previously thought. This work has helped to inform USA and European surgical care guidelines MWS results on how characteristics of individuals affect participation and outcomes of bowel cancer screening are being incorporated into screening programme development

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK.

Participants in the MWS are the first generation of women in the UK to have smoked to the same extent as men and their findings show that female smokers died about 10 years earlier than non-smokers. These findings have been used to support campaigns into smoking cessation.

4. Travis RC, et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies.

No evidence was found for an association between night shift work and breast cancer risk, with MWS findings similar to those from other perspective studies. These findings have contributed to active clinical and legal national and international debate on the relationship between shift work and health, ultimately benefiting the public.

5. Kent S, et al. Hospital costs in relation to body-mass index in 1 million women in England: a prospective cohort study.

This study provided clinicians with the estimated costs to the NHS for obesity-related conditions, and on breast cancer risk in relation to night shift work. This will benefit the NHS because it will allow efficient allocation of funding and will help make a case for a healthier lifestyle.

6. Floud et al. Body mass index, diet, physical inactivity, and the incidence of dementia in 1 million UK women.

The study has contributed: the first reliable evidence for women of the large excess risks associated with continued smoking and of the benefits of quitting; clear evidence of the adverse effects of low levels of alcohol consumption, both for cirrhosis and for cancer risk; large-scale evidence that night shift work has little or no effect on breast cancer risk; quantification of the magnitude of venous thromboembolism risk after surgery; and demonstration that at least a decade before dementia is diagnosed clinically, substantial changes in behaviour and other factors occur, illustrating the need for long-term follow-up for unbiased assessment the role of any causal factor.

7. Pirie K, Reeves GK et al. Cellular Telephone Use and the Risk of Brain Tumors: Update of the UK Million Women Study Journal of the National Cancer Institute.

Key findings:

• By 2011, almost 75% of women aged between 60 and 64 years used a mobile phone, and just below 50% of those aged between 75 and 79 years.

• Over the 14 year follow-up period, 3,268 (0.42%) of the women developed a brain tumour.

• There was no significant difference in the risk of developing a brain tumour between those who had never used a mobile phone, and mobile phone users. These included tumours in the temporal and parietal lobes, which are the most exposed parts of the brain.

• There was also no difference in the risk of developing glioma, acoustic neuroma, meningioma, pituitary tumours or eye tumours.

• There was no increase in the risk of developing any of these types of tumour for those who used a mobile phone daily, spoke for at least 20 minutes a week and/or had used a mobile phone for over 10 years.

• The incidence of right-sided and left-sided tumours was similar in mobile phone users, even though mobile phone use tends to be considerably greater on the right than the left side.

8. Wong A et al. Total sleep duration and daytime napping in relation to dementia detection risk: Results from the Million Women Study Alzheimer's & Dementia: The Journal of the Alzheimer's Association 2023 https://doi.org/10.1002/alz.13009

Key findings:

• 15,146 women had dementia recorded in their health records 15 or more years after the initial reporting on sleep duration;

• Short sleep duration (six hours or less) was associated with a slightly higher dementia risk (1.08 [1.04 – 1.12]) when compared with a normal sleep duration (seven-eight hours);

• Women who reported long sleep duration (nine hours or more) were not at increased risk of dementia (1.04 [0.99 – 1.09]);

• The associations did not appear to differ across dementia types (Alzheimer’s disease, vascular dementia, unspecified dementia) for either short or long sleep duration when compared to normal sleep duration.

In addition, MWS has contributed to clinical guidelines in UK and worldwide:

> MHRA guidelines on prescribing of menopausal hormone therapy since 2003 [https://webarchive.nationalarchives.gov.uk/20090217132200/http://www.mhra.gov.uk/Publications/Safetyguidance/CurrentProblemsinPharmacovigilance/CON007449]. Results from MWS on risk of breast cancer in MHT users (ref 1 above) contributed to major changes in prescribing guidelines in 2003, both in the UK and worldwide; it is estimated that subsequent prescribing changes mean that tens of thousands of breast cancers have been avoided worldwide. MWS continues to provide reliable evidence to the MHRA on MHT use and risks of breast and other cancers, and of other conditions. [https://www.gov.uk/drug-safety-update/hormone-replacement-therapy-hrt-further-information-on-the-known-increased-risk-of-breast-cancer-with-hrt-and-its-persistence-after-stopping, for clinicians]

The Million Women Study is a prospective study of 1.3 million UK women. The aims of the Million Women Study are to study the relevance of lifestyle, environmental, and genetic factors for major chronic diseases (e.g. stroke, heart disease, cancer, dementia), to help improve risk prediction and prevention of these diseases.

The Million Women Study is run by the Cancer Epidemiology Unit in the Nuffield Department of Population Health (NDPH) in the University of Oxford, with current and recent funding largely from Cancer Research UK and from the Medical Research Council.

The primary objective of the processing is to conduct research into the impact of lifestyle and other factors on health outcomes. The MWS is a national study of women’s health with current and recent funding by Cancer Research UK and the Medical Research Council. The study involves 1.3 million women in England and Scotland, recruited through NHS breast screening clinics between 1996 and 2001, who gave written consent for follow-up of their health through their medical records to examine how reproductive, lifestyle and other factors affect their future health.

While the initial stimulus of the MWS was to investigate use of different types of menopausal hormone therapy (MHT), sometimes also referred to as Hormone Replacement Therapy, (HRT) and risk of breast cancer, the study was designed from the outset to investigate the role of many other factors (such as smoking and obesity) and other health outcomes.

The primary aim of the study is to investigate common, largely modifiable, risk factors for serious and common diseases in women, under the leadership and guidance of the MWS Principal Investigators based in the Nuffield Department of Population Health (NDPH) at the University of Oxford, with advice from the MWS Advisory Committee and from the MWS Participant Panel. Research strategy takes into account the importance of age-appropriate research questions, in relation to new evidence and to public health and participant priorities; the availability of sufficiently long follow-up and number of cases for less common diseases; and additional information which becomes available for linkage and which may allow new analytic approaches .. As the study has progressed, and the cohort has aged, different conditions have become of greater relevance to study, for example: dementia, other neurodegenerative and neurological conditions, mental health, cardiovascular disease and stroke, as well as other cancers. In order to allow the investigators to understand the risks and burdens for common illnesses, the scope of the study has been expanded overtime.

The MWS research agenda is peer-reviewed every few years as part of applications for renewed funding. The program of research is determined by the MWS Principal Investigators based at Cancer Epidemiology Unit (CEU) in the Nuffield Department of Public Health (NDPH) at the University of Oxford, with advice from the MWS Advisory Committee and from the MWS Participant Panel.

Participants were initially asked to complete a questionnaire between 1996 and 2001 about their health and lifestyle, childbirth history, and contraceptive pill and MHT use. Participants also provided signed consent to follow up through their screening and other medical records. Postal questionnaires have been sent to all study participants around 3, 8, 12, and 15 and 20 years after recruitment, to keep track of participants’ health, to ask some new questions, and to keep participants informed of study progress. Selected women have also provided answers to online questionnaires, e.g. on diet, and to disease specific questionnaires, e.g. for cervical cancer and breast cancer. Tumour tissue samples are also being collected.

In addition, since 2006 some women have been recruited into a related study collecting blood and saliva samples for biochemical and genetic analyses, the MWS: Disease Susceptibility in Women study, which has separate ethical approval. Results of blood and questionnaire data from this study feed into the main MWS database. More recently, some women have been asked to take part in another ethically approved MWS related study, Changes to Lifestyle in Retirement. Results from this study also feed into the main analysis database.

Data provided by NHS Digital enable unbiased and virtually complete follow-up of the cohort for health outcomes and support continued contact with participants. Over time, the study has developed a comprehensive data resource to enable MWS Principal Investigators and their research team to discover which risk factors contribute to the development of particular diseases such as various types of cancer, cardiovascular disease and which do not.

Under this Agreement, the MWS at NDPH request to receive updates on participants’ health and health-related data.

This includes:

i. Hospital Episode Statistics Admitted Patient Care HES (APC) – details of participant admissions to hospital for any reason;

ii. Civil Registration Mortality data – details of participant deaths including date and cause and place of death;

iii. Cancer Registration data – details of any cancers registered to participants.

These data are provided by NHS Digital to allow MWS to integrate NHS Digital data reliably with other data in the study database. To address the GDPR Principle of Data Minimisation the University of Oxford have restricted this request to a cohort of 1.3 million women. The MWS have limited their request for HES data to Admitted Patient Care, as this includes full diagnostic coding. Outpatient and A&E records are not requested as the information is not as complete as in the requested datasets.

As MWS possess a large cohort, it is not feasible for the MWS to follow up participants directly on an individual basis and so the data provided via NHS Digital is crucial to the viability of the MWS, as the primary source for follow up of the cohort in England. The linked data provide the key mechanism for the MWS Principal Investigators to identify relevant health outcomes. These include cardiovascular disease, metabolic and endocrine disorders, neurodegenerative disease, cancers, stroke, musculo-skeletal conditions, mental health and any other conditions relevant to this population. Date and cause of death are necessary for risk factor analyses.

Without the cohort-linked data provided by NHS Digital the MWS investigators will not be able to reliably estimate the full burden of disease in the cohort. To assess the relationship between risk factors and disease, the MWS research team need to take into account any pre-existing conditions and confounding factors. For this reason, the study requires a complete picture of the medical history of the participant going back as far as NHS linked records permit, as well as new diagnoses going forward. To maintain accurate full records for study participants, data updates may be provided as full record datasets from start of available record to updated end date.

Linked data from NHS Digital are used together with personal, social, lifestyle and health information provided to the study by MWS participants completing the study questionnaires. Baseline data was collected for 1.3 million women via the recruitment questionnaire.

NHS Digital data are also used together with linked data from other sources (for example, Public Health England and Clinical Practice Research Datalink (CPRD) (participant level data linkage to GP records) and other ethically approved linkages (e.g. UK Biobank and the AgeX Trial). The MWS and all other studies to which it is linked have been granted ethical approval by relevant ethics committees and, where applicable, The Confidentiality Advisory Group (CAG) to permit these linkages.

Other datasets provided through NHS Digital include:

iv. Personal Demographic Service (PDS) updates – details of changes to participants’ demographics, postal addresses or registered GP practices.

It is very important for the MWS to be able to maintain accurate information about participants’ contact details including postal address and registered GP practice. These details were collected for the whole cohort at the time of recruitment but there will have inevitably been changes since then. There are several reasons for the requirement for PDS data:

• Ongoing approved linkages to, and current applications for, participant level primary care data (CPRD) require accurate information about registered GP practice (at practice as well as PCT level) to ensure that data about a particular participant are requested from the correct practice. Primary care data adds value to the HES Admitted Patient Care (APC) data: comparisons allow estimation of the completeness of coverage by HES for disease ascertainment, and for some conditions diagnosed and treated mainly in primary care, are the main source of outcomes.

• Part of the MWS follow-up strategy relies on regular re-contact with surviving participants to invite them to complete follow-up questionnaires about health-related exposures or outcomes. This allows for more complete follow-up for self-reported outcomes, and up to date information on exposures that are likely to change over time, which makes the findings more reliable. It is important that the MWS has correct and up-to-date participant contact details for this.

In summary, to produce scientifically valid results it is essential that the whole MWS cohort is linked to information from medical records. Examination of the datasets available from NHS Digital will provide the optimum amount of data required to produce valuable and meaningful results for the MWS. MWS requires linkage to the complete records in each of the datasets named above. With incomplete linkage, or with follow-up using only other methods such as participant report, there would be a high risk of the results being biased by selective lack of follow-up e.g. women who have died and many who are ill cannot complete follow-up questionnaires. Up to date details of deaths and of identifiers such as name and postal address are needed to ensure that when re-contacting participants, MWS minimise the risk of mail being sent to recently deceased participants or to out of date addresses. In order to produce accurate results in a long-term study of broad scope, it is not possible to minimise further the amount of data requested.

Participants have signed consent forms permitting access to their data to find out about the participant’s health in the future and to use information from those records. It is considered very unlikely that participants would object to the processing activities described in this application. Participants are kept informed of MWS activities through regular mailings and through their website and publications. During the past 22 years since participants joined the MWS, fewer than 200 of the 1.3 million participants have withdrawn from the study and asked that their information be deleted. Advice and feedback on data use are provided by the MWS Participant Panel and by the study’s Advisory Group, which has independent representation.

The processing described within this Agreement falls under ‘Article 6(1)(e) of the General Data Protection Regulation (GDPR) – processing is necessary for the performance of a task carried out in the public interest’ for research purposes is justified as:

• The study is directly related to population health benefits which are in the public interest.

• The University of Oxford is a ‘public authority’ as defined in the Data Protection Act 2018.

• Statute 1(3) of the ‘Statutes of the University of Oxford’ states that “The principal objects of the University are the advancement of learning by teaching and research and its dissemination by every means.” (http://www.admin.ox.ac.uk/statutes/1086-120.shtml)

The condition for processing special category personal data is met by the use of ‘Article 9(2)j - processing is necessary for archiving purposes in the public interest, scientific or historical research purposes’.

The University of Oxford is the sole data controller who also process the data for the Million Women Study as described in this Agreement. Cancer Research UK (CRUK) fund the study. CRUK do not have any role in determining the purpose and means of processing. The University of Oxford provide a Milestone Report to CRUK, any data contained within such report will be aggregated with small numbers suppressed (in line with HES Analysis Guidance).

The MWS is an ongoing open-ended cohort study. Results are disseminated nationally and internationally to scientific peers and to the public in peer-reviewed open-access papers in research journals, including The Lancet and The International Journal of Epidemiology; to national and international research and clinical colleagues in conference presentations; to government regulatory bodies in reports and through membership of relevant committees; and to participants and the wider public through the study website, through the media and through public outreach work (science open days, talks to participant groups). The study has been running for over 20 years and publications are produced on an ongoing basis. Publications do not identify individuals and contain only de-identified data in aggregate form, with small numbers suppressed in line with the HES Analysis Guidelines.

Study outputs demonstrate the benefit to health and social care of the use of data provided by NHS Digital in combination with other data collected by the MWS. Government and the NHS are able to incorporate MWS findings on disease risk in clinical guidance, and clinicians, researchers and the public are given reliable information about risk of disease in relation to risk factors, including modifiable lifestyle factors such as smoking, diet and exercise. Economic analyses help inform public health strategy. Data sharing maximises the use of MWS data for public good.

Examples of past outputs

• Among over 120 published papers, notable study publications include:

1. Beral V et al. Breast cancer risk in the interval between menopause and starting hormone therapy. J Natl Cancer Inst. 2011; 103(4):296-305.

2. Sweetland S, et al. Smoking, surgery, and venous thromboembolism risk in women: United Kingdom cohort study. Circulation. 2013;127(12):1276-82.

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013; 381(9861):133-41.

4. Travis RC, et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies. J Natl Cancer Inst. 2016; 108.

5. Kent S, et al. Hospital costs in relation to body-mass index in 1 million women in England: a prospective cohort study. Lancet Public Health. 2017; 2(5):e214-e222.

6. Floud et al. Body mass index, diet, physical inactivity, and the incidence of dementia in 1 million UK women. Neurology. 2020; 94(2):e123-e132.

• MWS contribution to clinical guidelines in UK and worldwide

1. MHRA guidelines on prescribing of menopausal hormone therapy since 2003 [https://webarchive.nationalarchives.gov.uk/20090217132200/http://www.mhra.gov.uk/Publications/Safetyguidance/CurrentProblemsinPharmacovigilance/CON007449]. Results from MWS on risk of breast cancer in MHT users (ref 1 above) contributed to major changes in prescribing guidelines in 2003, both in the UK and worldwide; it is estimated that subsequent prescribing changes mean that tens of thousands of breast cancers have been avoided worldwide. MWS continues to provide reliable evidence to the MHRA on MHT use and risks of breast and other cancers, and of other conditions. [https://www.gov.uk/drug-safety-update/hormone-replacement-therapy-hrt-further-information-on-the-known-increased-risk-of-breast-cancer-with-hrt-and-its-persistence-after-stopping, for clinicians]

[ https://www.nhs.uk/conditions/hormone-replacement-therapy-hrt/risks/, for patients]

2. USA and EU guidelines on risk of venous thromboembolism after surgery, 2016-2019 [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6963238/]. MWS aggregate analysis results provided in response to a data access request.

• MWS results are reported in the media and reach a worldwide audience

https://www.bbc.co.uk/news/health-45953442 - Height and cancer risk.

https://www.dailymail.co.uk/health/article-7805477/Being-obese-50s-raises-risk-dementia-21.html - obesity and dementia.

https://www.telegraph.co.uk/news/2019/08/30/women-hrt-urged-review-treatment-gp-following-new-breast-cancer/ new worldwide evidence [including MWS] on type and timing of HRT and breast cancer risk

Target publications from funded research programmes in the next 3 years (2022-2025):

At least 20 full peer-reviewed publications are planned, to include topics such as risk factors for dementia; diet; alcohol and risk of cancer; the role of co-morbidity in cancer patients; mammographic density and breast cancer risk; and risk prediction models for breast, ovarian and gastrointestinal cancers. As in the past, MWS will submit papers to journals such as the Lancet and Lancet group, International journal of Epidemiology, Neurology, JAMA and JNCI. In light of the SARS CoV-2 pandemic MWS have revised their study protocol to include analyses on risk factors for, and health and healthcare systems outcomes, from COVID-19 and plan to submit several additional papers on this topic.

All of these rely on data provided by NHS Digital and other NHS data providers. The study team will submit academic papers on the results of these and other analyses to open-access, high impact, peer-reviewed journals and will continue to liaise with NHS regulatory bodies as appropriate to ensure the results are used to help provide evidence for health guidelines and best clinical practice.

Other planned outputs for 2022-2025 include:

• The MWS website at www.millionwomenstudy.org will provide links to open access papers and offer free downloads of accessible summaries of findings. The twitter handles at NDPH (@Oxford_NDPH) and Oxford Medical Sciences Division (@OxfordMedSci) with a combined following of over 11,000 will promote all significant findings and will disseminate the findings at academic conferences and to lay audiences.

• Presentation of research findings from current and recent grant funded programmes at annual/biennial conferences e.g., National Cancer Research Institute, Society for Social Medicine and Population Health, Alzheimer’s Association International.

• Presentation of research findings to study participants and to the public [e.g., talks planned for U3A; 2022, WI; Participant Panel meetings, Oxford, annually].

• Sharing of research experience with UK Government and other research bodies and NHS data providers. Contributions through NHS Digital Research Advisory Group, UK Biobank working groups, the Dementias Platform UK and the CLOSER longitudinal studies consortium, and UK cohort COVID19 working groups.

The MWS research using linked health data has already directly influenced health care data as evidenced by the past outputs:

1. Beral V et al. Breast cancer risk in the interval between menopause and starting hormone therapy. J Natl Cancer Inst. 2011; 103(4):296-305.

This study provided evidence to suggest that Women using hormone replacement therapy are at increased risk of breast cancer; confirming and extending the Women's Health Initiative (WHI) clinical trial results published in 2002. These finding were shared with the Medicines and Healthcare Regulatory Agency and helped inform changes in prescribing guidance both for the UK and elsewhere. It is estimated that tens of thousands of cases of cancer worldwide have been avoided because of the subsequent fall in use of hormone therapy

Subsequent papers have extended and updated findings providing clarity on timing of use and long-term risks of past use.

2. Sweetland S, et al. Smoking, surgery, and venous thromboembolism risk in women: United Kingdom cohort study. Circulation. 2013;127(12):1276-82.

One of several papers from 2009-2013 using linked NHS data to quantify blood clot risk in relation to surgery along with other risk factors. The paper showed that risk of blood clots after surgery was far higher, and lasted for much longer, than had been previously thought. This work has helped to inform USA and European surgical care guidelines MWS results on how characteristics of individuals affect participation and outcomes of bowel cancer screening are being incorporated into screening programme development

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013; 381(9861):133-41.

Participants in the MWS are the first generation of women in the UK to have smoked to the same extent as men and their findings show that female smokers died about 10 years earlier than non-smokers. These findings have been used to support campaigns into smoking cessation.

4. Travis RC, et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies. J Natl Cancer Inst. 2016; 108.

No evidence was found for an association between night shift work and breast cancer risk, with MWS findings similar to those from other perspective studies. These findings have contributed to active clinical and legal national and international debate on the relationship between shift work and health, ultimately benefiting the public.

5. Kent S, et al. Hospital costs in relation to body-mass index in 1 million women in England: a prospective cohort study. Lancet Public Health. 2017; 2(5):e214-e222.

This study provided clinicians with the estimated costs to the NHS for obesity-related conditions, and on breast cancer risk in relation to night shift work. This will benefit the NHS because it will allow efficient allocation of funding and will help make a case for a healthier lifestyle.

6. Floud et al. Body mass index, diet, physical inactivity, and the incidence of dementia in 1 million UK women. Neurology. 2020; 94(2):e123-e132.

Midlife obesity may well be a cause of dementia. In contrast, the MWS found that behavioural changes due to preclinical disease could largely or wholly account for associations of low BMI, low caloric intake, and inactivity with dementia detection during the first decade of follow-up. These analyses were possible because of virtually complete follow-up by linkage to NHS hospital admissions data, which the MWS has shown to be reliable for such analyses of dementia. Dementia research is a key NHS priority and the study's unique dataset allows timely investigation of long term risks.

The value of the study in terms of public health benefit is already proven, most notably in its contribution to evidence-based guidelines for the prescribing and use of MHT, which have led to a fall in MHT use and subsequently to falling rates of breast cancer.

The study has contributed: the first reliable evidence for women of the large excess risks associated with continued smoking and of the benefits of quitting; clear evidence of the adverse effects of low levels of alcohol consumption, both for cirrhosis and for cancer risk; large-scale evidence that night shift work has little or no effect on breast cancer risk; quantification of the magnitude of venous thromboembolism risk after surgery; and demonstration that at least a decade before dementia is diagnosed clinically, substantial changes in behaviour and other factors occur, illustrating the need for long-term follow-up for unbiased assessment the role of any causal factor.

This Agreement permits the University of Oxford to receive identifiable HES, mortality, cancer registration and Demographics data for use in the Million Women Study (MWS). This Agreement also permits the University of Oxford to retain data disseminated under previous iterations of the Agreement.

The primary objective of the processing is to conduct research into the impact of lifestyle and other factors on health outcomes. The MWS is a national study of women’s health with current and recent funding by Cancer Research UK and the Medical Research Council. The study involves 1.3 million women in England and Scotland, recruited through NHS breast screening clinics between 1996 and 2001, who gave written consent for follow-up of their health through their medical records to examine how reproductive, lifestyle and other factors affect their future health.

While the initial stimulus of the MWS was to investigate use of different types of menopausal hormone therapy (MHT), sometimes also referred to as Hormone Replacement Therapy, (HRT) and risk of breast cancer, the study was designed from the outset to investigate the role of many other factors (such as smoking and obesity) and other health outcomes.

The primary aim of the study is to investigate common, largely modifiable, risk factors for serious and common diseases in women, under the leadership and guidance of the MWS Principal Investigators. As the study has progressed, and the cohort has aged, different conditions have become of greater relevance to study, for example: dementia, other neurodegenerative and neurological conditions, mental health, cardiovascular disease and stroke, as well as other cancers. In order to allow the investigators to understand the risks and burdens for common illnesses, the scope of the study has been expanded overtime.

The MWS research agenda is peer-reviewed every few years as part of applications for renewed funding. The program of research is determined by the MWS Principal Investigators based at Cancer Epidemiology Unit (CEU) in the Nuffield Department of Public Health (NDPH) at the University of Oxford, with advice from the MWS Advisory Committee and from the MWS Participant Panel. Research strategy takes into account the importance of age-appropriate research questions, in relation to new evidence and to public health and participant priorities; the availability of sufficiently long follow-up and number of cases for less common diseases; and additional information which becomes available for linkage, and which may allow new analytic approaches.

Participants were initially asked to complete a questionnaire between 1996 and 2001 about their health and lifestyle, childbirth history, and contraceptive pill and MHT use. Participants also provided signed consent to follow up through their screening and other medical records. Postal questionnaires have been sent to all study participants around 3, 8, 12, and 15 and 20 years after recruitment, to keep track of participants’ health, to ask some new questions, and to keep participants informed of study progress. Selected women have also provided answers to online questionnaires, e.g. on diet, and to disease specific questionnaires, e.g. for cervical cancer and breast cancer. Tumour tissue samples are also being collected.

Data provided by NHS Digital enable unbiased and virtually complete follow-up of the cohort for health outcomes and support continued contact with participants. Over time, the study has developed a comprehensive data resource to enable MWS Principal Investigators and their research team to discover which risk factors contribute to the development of particular diseases such as various types of cancer, cardiovascular disease and which do not.

Under this Agreement, the MWS at NDPH request to receive updates on participants’ health and health-related data.

This includes:

i. Hospital Episode Statistics Admitted Patient Care HES (APC) – details of participant admissions to hospital for any reason;

ii. Civil Registration Mortality data – details of participant deaths including date and cause and place of death;

iii. Cancer Registration data – details of any cancers registered to participants.

These data are provided by NHS Digital to allow MWS to integrate NHS Digital data reliably with other data in the study database. To address the GDPR Principle of Data Minimisation the University of Oxford have restricted this request to a cohort of 1.3 million women. The MWS have limited their request for HES data to Admitted Patient Care, as this includes full diagnostic coding. Outpatient and A&E records are not requested as the information is not as complete as in the requested datasets.

As MWS possess a large cohort, it is not feasible for the MWS to follow up participants directly on an individual basis and so the data provided via NHS Digital is crucial to the viability of the MWS, as the primary source for follow up of the cohort in England. The linked data provide the key mechanism for the MWS Principal Investigators to identify relevant health outcomes. These include cardiovascular disease, metabolic and endocrine disorders, neurodegenerative disease, cancers, stroke, musculo-skeletal conditions, mental health and any other conditions relevant to this population.

Without the cohort-linked data provided by NHS Digital the MWS investigators will not be able to reliably estimate the full burden of disease in the cohort. To assess the relationship between risk factors and disease, the MWS research team need to take into account any pre-existing conditions and confounding factors. For this reason, the study requires a complete picture of the medical history of the participant going back as far as NHS linked records permit, as well as new diagnoses going forward. To maintain accurate full records for study participants, data updates may be provided as full record datasets from start of available record to updated end date.

Linked data from NHS Digital are used together with personal, social, lifestyle and health information provided to the study by MWS participants completing the study questionnaires. Baseline data was collected for 1.3 million women via the recruitment questionnaire.

NHS Digital data are also used together with linked data from other sources (for example, Public Health England and Clinical Practice Research Datalink (CPRD) (participant level data linkage to GP records) and other ethically approved linkages (e.g. UK Biobank and the AgeX Trial). The MWS and all other studies to which it is linked have been granted ethical approval by relevant ethics committees and, where applicable, tThe Confidentiality Advisory Group (CAG) to permit these linkages.

Other datasets provided through NHS Digital include:

iv. Personal Demographic Service (PDS) updates – details of changes to participants’ demographics, postal addresses or registered GP practices.

It is very important for the MWS to be able to maintain accurate information about participants’ contact details including postal address and registered GP practice. These details were collected for the whole cohort at the time of recruitment but there will have inevitably been changes since then. There are several reasons for the requirement for PDS data:

• Ongoing approved linkages to, and current applications for, participant level primary care data (CPRD) require accurate information about registered GP practice (at practice as well as PCT level) to ensure that data about a particular participant are requested from the correct practice. Primary care data adds value to the HES Admitted Patient Care (APC) data: comparisons allow estimation of the completeness of coverage by HES for disease ascertainment, and for some conditions diagnosed and treated mainly in primary care, are the main source of outcomes.

• Part of the MWS follow-up strategy relies on regular re-contact with surviving participants to invite them to complete follow-up questionnaires about health-related exposures or outcomes. This allows for more complete follow-up for self-reported outcomes, and up to date information on exposures that are likely to change over time, which makes the findings more reliable. It is important that the MWS has correct and up-to-date participant contact details for this.

In summary, to produce scientifically valid results it is essential that the whole MWS cohort is linked to information from medical records. Examination of the datasets available from NHS Digital will provide the optimum amount of data required to produce valuable and meaningful results for the MWS. MWS requires linkage to the complete records in each of the datasets named above. With incomplete linkage, or with follow-up using only other methods such as participant report, there would be a high risk of the results being biased by selective lack of follow-up e.g. women who have died and many who are ill cannot complete follow-up questionnaires. Up to date details of deaths and of identifiers such as name and postal address are needed to ensure that when re-contacting participants, MWS minimise the risk of mail being sent to recently deceased participants or to out of date addresses. In order to produce accurate results in a long-term study of broad scope, it is not possible to minimise further the amount of data requested.

Participants have signed consent forms permitting access to their data to find out about the participant’s health in the future and to use information from those records. It is considered very unlikely that participants would object to the processing activities described in this application. Participants are kept informed of MWS activities through regular mailings and through their website and publications. During the past 22 years since participants joined the MWS, fewer than 200 of the 1.3 million participants have withdrawn from the study and asked that their information be deleted. Advice and feedback on data use are provided by the MWS Participant Panel and by the study’s Advisory Group, which has independent representation.

The average age of Million Women Study participants is now 78 years. Study participants are especially vulnerable to coronavirus infection. The study are currently documenting mortality from the disease and about 4000 have already died (mainly in April and May 2020) with COVID mentioned on their death certificate. However, there are considerably more deaths than expected from previous years from non-COVID causes. Having additional hospital admission (HES) data for April, May and June 2020, even if provisional, would help understand risk factors, not only for COVID, but also the extra deaths occurring in 2020 - 2021.

The processing described within this Agreement falls under ‘Article 6(1)(e) of the General Data Protection Regulation (GDPR) – processing is necessary for the performance of a task carried out in the public interest’ for research purposes is justified as:

• The study is directly related to population health benefits which are in the public interest.

• The University of Oxford is a ‘public authority’ as defined in the Data Protection Act 2018.

• Statute 1(3) of the ‘Statutes of the University of Oxford’ states that “The principal objects of the University are the advancement of learning by teaching and research and its dissemination by every means.” (http://www.admin.ox.ac.uk/statutes/1086-120.shtml)

The condition for processing special category personal data is met by the use of ‘Article 9(2)j - processing is necessary for archiving purposes in the public interest, scientific or historical research purposes’.

The University of Oxford is the sole data controller who also process the data for the Million Women Study as described in this Agreement. Cancer Research UK (CRUK) fund the study. CRUK do not have any role in determining the purpose and means of processing. The University of Oxford provide a Milestone Report to CRUK, any data contained within such report will be aggregated with small numbers suppressed (in line with HES Analysis Guidance).

The MWS is an ongoing open-ended cohort study. Results are disseminated nationally and internationally to scientific peers and to the public in peer-reviewed open-access papers in research journals, including The Lancet and The International Journal of Epidemiology; to national and international research and clinical colleagues in conference presentations; to government regulatory bodies in reports and through membership of relevant committees; and to participants and the wider public through the study website, through the media and through public outreach work (science open days, talks to participant groups). The study has been running for over 20 years and publications are produced on an ongoing basis. Publications do not identify individuals and contain only aggregated data with small numbers supressed in line with the HES Analysis Guidance.

Examples of past outputs

• Among over 120 published papers, notable study publications include:

1. Beral V et al. Breast cancer risk in the interval between menopause and starting hormone therapy. J Natl Cancer Inst. 2011; 103(4):296-305.

First large cohort study comparing breast cancer risk with different types of hormone replacement therapy. Current use associated with increased risk, greater for combined oestrogen-progestagen therapy. Confirmed and extended Womes Health Initiative (WHI) clinical trial results published in 2002. Subsequent papers have extended and updated findings providing clarity on timing of use and long-term risks of past use.

2. Sweetland S, et al. Smoking, surgery, and venous thromboembolism risk in women: United Kingdom cohort study. Circulation. 2013;127(12):1276-82.

One of several papers from 2009-2013 using linked NHS data to quantify VTE risk in relation to surgery along with other risk factors. First large prospective study in women, showing that smoking increases risk of VTE and adds to already substantial risk after surgery.

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013; 381(9861):133-41.

First prospective analysis of long-term risks of smoking in UK women, showing risks comparable to those found in men in the British Doctors’ Study; using linked NHS data and providing the first reliable estimates of smoking-related mortality risks for women, widely relevant for health care and policy.

4. Travis RC, et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies. J Natl Cancer Inst. 2016; 108.

No evidence was found for an association between night shift work and breast cancer risk, with MWS findings similar to those from other prospective studies.

5. Kent S, et al. Hospital costs in relation to body-mass index in 1 million women in England: a prospective cohort study. Lancet Public Health. 2017; 2(5):e214-e222.

This NIHR funded health economic analysis [and a linked analysis published in 2019 of primary care costs] provided novel reliable and comprehensive estimates of the impact of obesity on NHS care, using linked data to include conditions not previously considered relevant.

6. Floud et al. Body mass index, diet, physical inactivity, and the incidence of dementia in 1 million UK women. Neurology. 2020; 94(2):e123-e132.

Midlife obesity may well be a cause of dementia. In contrast, the MWS found that behavioural changes due to preclinical disease could largely or wholly account for associations of low BMI, low caloric intake, and inactivity with dementia detection during the first decade of follow-up. These analyses were possible because of virtually complete follow-up by linkage to NHS hospital admissions data, which the MWS has shown to be reliable for such analyses of dementia. Dementia research is a key NHS priority, and our unique dataset allows timely investigation of long term risks.

MWS contribution to published clinical guidelines in UK and worldwide

1. MHRA guidelines on prescribing of menopausal hormone therapy since 2003 [https://webarchive.nationalarchives.gov.uk/20090217132200/http://www.mhra.gov.uk/Publications/Safetyguidance/CurrentProblemsinPharmacovigilance/CON007449]. Results from MWS on risk of breast cancer in MHT users (ref 1 above) contributed to major changes in prescribing guidelines in 2003, both in the UK and worldwide; it is estimated that subsequent prescribing changes mean that tens of thousands of breast cancers have been avoided worldwide. MWS continues to provide reliable evidence to the MHRA on MHT use and risks of breast and other cancers, and of other conditions. [https://www.gov.uk/drug-safety-update/hormone-replacement-therapy-hrt-further-information-on-the-known-increased-risk-of-breast-cancer-with-hrt-and-its-persistence-after-stopping, for clinicians]

[ https://www.nhs.uk/conditions/hormone-replacement-therapy-hrt/risks/, for patients]

2. USA and EU guidelines on risk of venous thromboembolism after surgery, 2016-2019 [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6963238/]. MWS aggregate analysis results provided in response to a data access request.

• Million Women Study results are reported in the media and reach a worldwide audience

https://www.bbc.co.uk/news/health-45953442 - Height and cancer risk.

https://www.dailymail.co.uk/health/article-7805477/Being-obese-50s-raises-risk-dementia-21.html - obesity and dementia.

https://www.telegraph.co.uk/news/2019/08/30/women-hrt-urged-review-treatment-gp-following-new-breast-cancer/ new worldwide evidence [including MWS] on type and timing of HRT and breast cancer risk

Target publications from funded research programmes in the next 5 years (2020-2025)

At least 20 full peer-reviewed publications are planned, to include topics such as risk factors for dementia; diet, alcohol and risk of cancer; the role of co-morbidity in cancer patients; mammographic density and breast cancer risk; and risk prediction models for breast, ovarian and gastrointestinal cancers. As in the past, the MWS will submit papers to journals such as the Lancet and Lancet group, International Journal of Epidemiology, Neurology, JAMA and JNCI.

In light of the SARS CoV-2 pandemic, the MWS' study protocol was revised to include analyses on risk factors for and health and healthcare systems outcomes from COVID-19 and the plan is to submit several additional papers on this topic.

All of these rely on data provided by NHS Digital and other NHS data providers. The study team will submit academic papers on the results of these and other analyses to open-access, high impact, peer-reviewed journals and will continue to liaise with NHS regulatory bodies as appropriate to ensure the results are used to help provide evidence for health guidelines and best clinical practice.

Other planned outputs for 2021-2025 include:

• The MWS website at www.millionwomenstudy.org will provide links to open access papers and offer free downloads of accessible summaries of findings. The twitter handles at NDPH (@Oxford_NDPH) and Oxford Medical Sciences Division (@OxfordMedSci) with a combined following of over 11,000 will promote all significant findings and will disseminate the findings at academic conferences and to lay audiences.

• Presentation of research findings from current and recent grant funded programmes at annual/biennial conferences eg National Cancer Research Institute, Society for Social Medicine and Population Health, Alzheimer’s Association International.

• Presentation of research findings to study participants and to the public [eg talks planned for U3A; 2021, WI; Participant Panel meetings, Oxford, annually].

• Sharing of research experience with UK Government and other research bodies and NHS data providers. Contributions through NHS Digital Research Advisory Group, UK Biobank working groups, the Dementias Platform UK and the CLOSER longitudinal studies consortium, and UK cohort COVID19 working groups.

The MWS research using linked health data has already directly influenced health care data as evidenced by the past outputs:

1. Beral V et al. Breast cancer risk in the interval between menopause and starting hormone therapy. J Natl Cancer Inst. 2011; 103(4):296-305.

This study provided evidence to suggest that Women using hormone replacement therapy are at increased risk of breast cancer; confirming and extending the Womes Health Initiative (WHI) clinical trial results published in 2002. These finding were shared with the Medicines and Healthcare Regulatory Agency and helped inform changes in prescribing guidance both for the UK and elsewhere. It is estimated that tens of thousands of cases of cancer worldwide have been avoided because of the subsequent fall in use of hormone therapy

Subsequent papers have extended and updated findings providing clarity on timing of use and long-term risks of past use.

2. Sweetland S, et al. Smoking, surgery, and venous thromboembolism risk in women: United Kingdom cohort study. Circulation. 2013;127(12):1276-82.

One of several papers from 2009-2013 using linked NHS data to quantify blood clot risk in relation to surgery along with other risk factors. The paper showed that risk of blood clots after surgery was far higher, and lasted for much longer, than had been previously thought. This work has helpeed to inform USA and European surgical care guidelines MWS results on how characteristics of individuals affect participation and outcomes of bowel cancer screening are being incorporated into screening programme development

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013; 381(9861):133-41.

Participants in the MWS are the first generation of women in the UK to have smoked to the same extent as men and their findings show that female smokers died about 10 years earlier than non-smokers. These findings have been used to support campaigns into smoking cessation.

4. Travis RC, et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies. J Natl Cancer Inst. 2016; 108.

No evidence was found for an association between night shift work and breast cancer risk, with MWS findings similar to those from other perspective studies. These findings have contributed to active clinical and legal national and international debate on the relationship between shift work and health, ultimately benefiting the public.

5. Kent S, et al. Hospital costs in relation to body-mass index in 1 million women in England: a prospective cohort study. Lancet Public Health. 2017; 2(5):e214-e222.

This study provided clinicians with the estimated costs to the NHS for obesity-related conditions, and on breast cancer risk in relation to night shift work. This will benefit the NHS because it will allow efficient allocation of funding, and will help make a case for a healthier lifestyle.

6. Floud et al. Body mass index, diet, physical inactivity, and the incidence of dementia in 1 million UK women. Neurology. 2020; 94(2):e123-e132.

Midlife obesity may well be a cause of dementia. In contrast, the MWS found that behavioural changes due to preclinical disease could largely or wholly account for associations of low BMI, low caloric intake, and inactivity with dementia detection during the first decade of follow-up. These analyses were possible because of virtually complete follow-up by linkage to NHS hospital admissions data, which the MWS has shown to be reliable for such analyses of dementia. Dementia research is a key NHS priority and our unique dataset allows timely investigation of long term risks.

The value of the study in terms of public health benefit is already proven, most notably in its contribution to evidence-based guidelines for the prescribing and use of MHT, which have led to a fall in inappropriate MHT use and subsequently to falling rates of breast cancer.

The study has contributed: the first reliable evidence for women of the large excess risks associated with continued smoking and of the benefits of quitting; clear evidence of the adverse effects of low levels of alcohol consumption, both for cirrhosis and for cancer risk; large-scale evidence that night shift work has little or no effect on breast cancer risk; quantification of the magnitude of venous thromboembolism risk after surgery; and demonstration that at least a decade before dementia is diagnosed clinically, substantial changes in behaviour and other factors occur, illustrating the need for long-term follow-up for unbiased assessment the role of any causal factor.

The University of Oxford requires HES, mortality, cancer registration and PDS data for use in the Million Women Study (MWS).

The MWS is a national study of women’s health funded by Cancer Research UK and the Medical Research Council. The study involves 1.3 million women in England and Scotland, recruited through NHS breast screening clinics between 1996 and 2001, who gave written consent for follow up of their health through their medical records to examine how reproductive, lifestyle and other factors affect their future health. While the initial stimulus of the study was to investigate use of different types on menopausal hormone therapy and risk of breast cancer, from its inception the study was designed to investigate the role of other factors (such as smoking and obesity) and other health outcomes. The overall aim of the study is to investigate common, largely modifiable, risk factors for serious and common diseases in women. As the study has progressed, and the cohort has aged, different conditions have become of greater relevance: for example, dementia, other neurodegenerative and neurological conditions, mental health, cardiovascular disease and stroke, as well as other cancers. The research agenda is peer- reviewed every few years as part of the applications for renewed funding and developments and are determined by the MWS team in the Cancer Epidemiology Unit (CEU) at the University of Oxford, taking into account the importance of age-appropriate research questions, in relation to new evidence and to public health priorities; the availability of sufficiently long follow-up and number of cases for less common diseases; and additional information which becomes available for linkage and which may allow new analytic approaches.

Cancer diagnosis data and death data and HES admitted patient data are required in order to define end points and disease diagnoses in study participants during follow-up. To produce scientifically valid results, it is essential that the whole Million Women Study cohort is linked to information from medical records. With incomplete linkage, or with follow-up using other methods such as participant report, there would be a high risk of the results being biased by selective lack of follow-up e.g. women who have died and many who are ill cannot complete follow-up questionnaires. Up to date details of deaths and of identifiers such as name and postal address are also needed to ensure that when re-contacting participants, this will ensure no mail will be sent to recently-deceased participants and minimise mailing to out of date addresses.

The wealth of data in the MWS requires careful analysis, and any pre-existing conditions and confounding factors need to be considered. For this reason, it requires a complete picture of the medical history of the participant going back as far as NHS linked records permit, and new diagnoses going forward. In order to produce accurate results in a long –term study of broad scope, it is not possible to minimise further the amount of data requested.

Linked data from NHS Digital are used together with personal, social, lifestyle and health information provided to the study by participants completing the study questionnaires. There is baseline data collected for 1.3 million women via the recruitment questionnaire. Postal re-survey questionnaires were sent to all study participants around 3, 8, 12, and 15 years after recruitment, to keep track of participant’s health and to ask some new questions; and selected women also provided answers to an online diet questionnaire and to disease specific questionnaires, e.g. for cervical cancer and breast cancer. In addition, some women were recruited into a sub study collecting blood samples for biochemical and genetic analyses, the Million Women Study: Disease Susceptibility in Women study.

The average age of Million Women Study participants is now 77 years. Study participants are especially vulnerable to coronavirus infection. The study are currently documenting mortality from the disease and about 4000 have already died (mainly in April and May 2020) with COVID mentioned on their death certificate. However, there are considerably more deaths than expected from previous years from non-COVID causes. Having additional hospital admission (HES) data for April, May and June 2020, even if provisional, would help understand risk factors, not only for COVID, but also the extra deaths occurring in April, May and June 2020.

The Million Women Study is an ongoing open-ended cohort study. Results are disseminated nationally and internationally in peer-reviewed open-access papers in research journals including The Lancet and The International Journal of Epidemiology, in related presentations to national and international research and clinical colleagues, in reports to government regulatory bodies, on the study website, through the media and through public outreach work (science open days, talks to participant groups). The study has been running for over 20 years and publications are produced on an ongoing basis. Presentations of findings will be at conferences such as the National Cancer Research Institute (NCRI) Conference in 2019.

The publications do not identify individuals and contain only de-identified data in aggregate form (tabulations and figures showing analysis results at the minimum level of detail required, using small number suppression) in line with the HES Analysis Guide and with ONS terms and conditions relating to the data being shared under this Agreement.

Among over 120 published papers, notable study publications include:

1. Beral V; Million Women Study Collaborators. Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet. 2003;362(9382):419-27.

2. Sweetland S, Beral V, et al. Venous thromboembolism risk in relation to use of different types of postmenopausal hormone therapy in a large prospective study. J Thromb Haemost. 2012;10(11):2277-86.

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013 Jan 12;381(9861):133-41.

4. Travis RC,et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies. J Natl Cancer Inst 2016;108.

5. Kent S,et al. Hospital costs in relation to body-mass index in 1·1 million women in England: a prospective cohort study. Lancet Public Health. 2017;2(5):e214-e222.

Million Women Study results are reported in the media such as the BBC and national newspapers and its outputs reach a worldwide audience. e.g.

http://www.bbc.co.uk/news/health-35052404 - being unhappy or stressed will not make you ill

http://news.bbc.co.uk/1/hi/health/7906355.stm - a drink a day raises cancer risk

http://www.bbc.co.uk/news/health-14220382 - tall people more likely to develop cancer

Future anticipated work using HES, cancer registry and ONS mortality data:

Current funded work includes analyses of hormonal and other risk factors for subtypes of various common cancers (ovarian, colorectal, breast, endometrial and haematological); of diet and risk of cancer, fracture and vascular disease; of lifestyle factors, past health including mental health and long-term risk of dementia; of risk of osteoporosis and fracture in relation to use of bisphosphonate drugs; and of risk factors for different types of stroke. The study team will submit academic papers on the results of these analyses to open-access, high impact, peer-reviewed journals and will continue to liaise with NHS regulatory bodies as appropriate to ensure the results are used to help provide evidence for health guidelines and best clinical practice.

The Million Women Study website at www.millionwomenstudy.org provides links to open access papers and offers free downloads of accessible summaries of findings. The twitter handles at Nuffield Department of Population Health (@Oxford_NDPH) and Oxford Medical Sciences Division (@OxfordMedSci) with a combined following of over 11,000 will promote all significant findings and will disseminate the findings at academic conferences and to lay audiences, as above.

All future outputs will contain only data that is aggregated, with small numbers suppressed in line with the HES Analysis Guide. ONS terms and conditions relating to the data being shared under this agreement will be adhered to.

In the next year or so the following reports, analyses and presentations are planned:

Data Analyses and papers:

1. Long-term effects of dementia associated with midlife body mass index, diet and physical activity in one million UK women to be submitted to the journal Alzheimer’s and Dementia 2018/19.

2. Smoking and dementia incidence: prospective study of one million UK women to be submitted to The Lancet 2018/19.

3. Foods, macronutrients and breast cancer risk in postmenopausal women: a large UK cohort. To be submitted to the International Journal of Epidemiology 2018/19.

4. Diet and risk of different types of stroke. To be submitted to the journal Neurology 2018/19.

Upcoming Conference presentations:

1. Use of Endocrine therapy for breast cancer to be presented to The Early Breast Cancer Triallists Collaborative Group in Oxford June 2018.

2. Timing of alcohol consumption and risk of liver cirrhosis: prospective study of UK women at European Congress of Epidemiology in Lyon July 2018.

3. Histological type as a predictor of ovarian cancer survival at the British Association of Gynaecological Pathologists - Annual Meeting June 2018.

4. Integrating pathology into epidemiology: Challenges and opportunities. European Society of Pathology Academy July 2018.

The Million Women Study research using linked health data has already directly influenced health care. Results showed, for example, that women using hormone replacement therapy are at increased risk of breast cancer; this work, published in 2003, was shared with the Medicines and Healthcare Regulatory Agency and helped inform changes in prescribing guidance both for the UK and elsewhere. It is estimated that tens of thousands of cases of cancer worldwide have been avoided as a result of the subsequent fall in use of hormone therapy. Another example where NHS Digital data contributed to an influential paper was in showing that risk of blood clots after surgery was far higher, and lasted for much longer, than had been previously thought. This work is helping to inform European surgical care guidelines. MWS results on how characteristics of individuals affect participation and outcomes of bowel cancer screening are being incorporated into screening programme development. Recent work on the costs to the NHS of obesity-related conditions, and on breast cancer risk in relation to night shift work, has direct public health relevance. Participants in the Million Women Study are the first generation of women in the UK to have smoked to the same extent as a man and the findings show that female smokers died about 10 years earlier than non-smokers. These effects are much greater than had been reported previously and influence policy.

The University of Oxford requires HES, mortality, cancer registration and PDS data for use in the Million Women Study (MWS).

The MWS is a national study of women’s health funded by Cancer Research UK and the Medical Research Council. The study involves 1.3 million women in England and Scotland, recruited through NHS breast screening clinics between 1996 and 2001, who gave written consent for follow up of their health through their medical records to examine how reproductive, lifestyle and other factors affect their future health. While the initial stimulus of the study was to investigate use of different types on menopausal hormone therapy and risk of breast cancer, from its inception the study was designed to investigate the role of other factors (such as smoking and obesity) and other health outcomes. The overall aim of the study is to investigate common, largely modifiable, risk factors for serious and common diseases in women. As the study has progressed, and the cohort has aged, different conditions have become of greater relevance: for example, dementia, other neurodegenerative and neurological conditions, mental health, cardiovascular disease and stroke, as well as other cancers. The research agenda is peer- reviewed every few years as part of the applications for renewed funding and developments and are determined by the MWS team in the Cancer Epidemiology Unit (CEU) at the University of Oxford, taking into account the importance of age-appropriate research questions, in relation to new evidence and to public health priorities; the availability of sufficiently long follow-up and number of cases for less common diseases; and additional information which becomes available for linkage and which may allow new analytic approaches.

Cancer diagnosis data and death data and HES admitted patient data are required in order to define end points and disease diagnoses in study participants during follow-up. To produce scientifically valid results, it is essential that the whole Million Women Study cohort is linked to information from medical records. With incomplete linkage, or with follow-up using other methods such as participant report, there would be a high risk of the results being biased by selective lack of follow-up e.g. women who have died and many who are ill cannot complete follow-up questionnaires. Up to date details of deaths and of identifiers such as name and postal address are also needed to ensure that when re-contacting participants, this will ensure no mail will be sent to recently-deceased participants and minimise mailing to out of date addresses.

The wealth of data in the MWS requires careful analysis, and any pre-existing conditions and confounding factors need to be considered. For this reason, it requires a complete picture of the medical history of the participant going back as far as NHS linked records permit, and new diagnoses going forward. In order to produce accurate results in a long –term study of broad scope, it is not possible to minimise further the amount of data requested.

Linked data from NHS Digital are used together with personal, social, lifestyle and health information provided to the study by participants completing the study questionnaires. There is baseline data collected for 1.3 million women via the recruitment questionnaire. Postal re-survey questionnaires were sent to all study participants around 3, 8, 12, and 15 years after recruitment, to keep track of participant’s health and to ask some new questions; and selected women also provided answers to an online diet questionnaire and to disease specific questionnaires, e.g. for cervical cancer and breast cancer. In addition, some women were recruited into a sub study collecting blood samples for biochemical and genetic analyses, the Million Women Study: Disease Susceptibility in Women study.

The average age of Million Women Study participants is now 77 years. Study participants are especially vulnerable to coronavirus infection. The study are currently documenting mortality from the disease and about 4000 have already died (mainly in April and May 2020) with COVID mentioned on their death certificate. However, there are considerably more deaths than expected from previous years from non-COVID causes. Having additional hospital admission (HES) data for April, May and June 2020, even if provisional, would help understand risk factors, not only for COVID, but also the extra deaths occurring in April, May and June 2020.

The Million Women Study is an ongoing open-ended cohort study. Results are disseminated nationally and internationally in peer-reviewed open-access papers in research journals including The Lancet and The International Journal of Epidemiology, in related presentations to national and international research and clinical colleagues, in reports to government regulatory bodies, on the study website, through the media and through public outreach work (science open days, talks to participant groups). The study has been running for over 20 years and publications are produced on an ongoing basis. Presentations of findings will be at conferences such as the National Cancer Research Institute (NCRI) Conference in 2019.

The publications do not identify individuals and contain only de-identified data in aggregate form (tabulations and figures showing analysis results at the minimum level of detail required, using small number suppression) in line with the HES Analysis Guide and with ONS terms and conditions relating to the data being shared under this Agreement.

Among over 120 published papers, notable study publications include:

1. Beral V; Million Women Study Collaborators. Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet. 2003;362(9382):419-27.

2. Sweetland S, Beral V, et al. Venous thromboembolism risk in relation to use of different types of postmenopausal hormone therapy in a large prospective study. J Thromb Haemost. 2012;10(11):2277-86.

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013 Jan 12;381(9861):133-41.

4. Travis RC,et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies. J Natl Cancer Inst 2016;108.

5. Kent S,et al. Hospital costs in relation to body-mass index in 1·1 million women in England: a prospective cohort study. Lancet Public Health. 2017;2(5):e214-e222.

Million Women Study results are reported in the media such as the BBC and national newspapers and its outputs reach a worldwide audience. e.g.

http://www.bbc.co.uk/news/health-35052404 - being unhappy or stressed will not make you ill

http://news.bbc.co.uk/1/hi/health/7906355.stm - a drink a day raises cancer risk

http://www.bbc.co.uk/news/health-14220382 - tall people more likely to develop cancer

Future anticipated work using HES, cancer registry and ONS mortality data:

Current funded work includes analyses of hormonal and other risk factors for subtypes of various common cancers (ovarian, colorectal, breast, endometrial and haematological); of diet and risk of cancer, fracture and vascular disease; of lifestyle factors, past health including mental health and long-term risk of dementia; of risk of osteoporosis and fracture in relation to use of bisphosphonate drugs; and of risk factors for different types of stroke. The study team will submit academic papers on the results of these analyses to open-access, high impact, peer-reviewed journals and will continue to liaise with NHS regulatory bodies as appropriate to ensure the results are used to help provide evidence for health guidelines and best clinical practice.

The Million Women Study website at www.millionwomenstudy.org provides links to open access papers and offers free downloads of accessible summaries of findings. The twitter handles at Nuffield Department of Population Health (@Oxford_NDPH) and Oxford Medical Sciences Division (@OxfordMedSci) with a combined following of over 11,000 will promote all significant findings and will disseminate the findings at academic conferences and to lay audiences, as above.

All future outputs will contain only data that is aggregated, with small numbers suppressed in line with the HES Analysis Guide. ONS terms and conditions relating to the data being shared under this agreement will be adhered to.

In the next year or so the following reports, analyses and presentations are planned:

Data Analyses and papers:

1. Long-term effects of dementia associated with midlife body mass index, diet and physical activity in one million UK women to be submitted to the journal Alzheimer’s and Dementia 2018/19.

2. Smoking and dementia incidence: prospective study of one million UK women to be submitted to The Lancet 2018/19.

3. Foods, macronutrients and breast cancer risk in postmenopausal women: a large UK cohort. To be submitted to the International Journal of Epidemiology 2018/19.

4. Diet and risk of different types of stroke. To be submitted to the journal Neurology 2018/19.

Upcoming Conference presentations:

1. Use of Endocrine therapy for breast cancer to be presented to The Early Breast Cancer Triallists Collaborative Group in Oxford June 2018.

2. Timing of alcohol consumption and risk of liver cirrhosis: prospective study of UK women at European Congress of Epidemiology in Lyon July 2018.

3. Histological type as a predictor of ovarian cancer survival at the British Association of Gynaecological Pathologists - Annual Meeting June 2018.

4. Integrating pathology into epidemiology: Challenges and opportunities. European Society of Pathology Academy July 2018.

The Million Women Study research using linked health data has already directly influenced health care. Results showed, for example, that women using hormone replacement therapy are at increased risk of breast cancer; this work, published in 2003, was shared with the Medicines and Healthcare Regulatory Agency and helped inform changes in prescribing guidance both for the UK and elsewhere. It is estimated that tens of thousands of cases of cancer worldwide have been avoided as a result of the subsequent fall in use of hormone therapy. Another example where NHS Digital data contributed to an influential paper was in showing that risk of blood clots after surgery was far higher, and lasted for much longer, than had been previously thought. This work is helping to inform European surgical care guidelines. MWS results on how characteristics of individuals affect participation and outcomes of bowel cancer screening are being incorporated into screening programme development. Recent work on the costs to the NHS of obesity-related conditions, and on breast cancer risk in relation to night shift work, has direct public health relevance. Participants in the Million Women Study are the first generation of women in the UK to have smoked to the same extent as a man and our findings show that female smokers died about 10 years earlier than non-smokers. These effects are much greater than had been reported previously and influence policy.

23/7/2019 Amendment to the DSA to add only 1 complete year of HES for 2018/2019

The University of Oxford requires HES, mortality, cancer registration and PDS data for use in the Million Women Study (MWS).

The MWS is a national study of women’s health funded by Cancer Research UK and the Medical Research Council. The study involves 1.3 million women in England and Scotland, recruited through NHS breast screening clinics between 1996 and 2001, who gave written consent for follow up of their health through their medical records to examine how reproductive, lifestyle and other factors affect their future health. While the initial stimulus of the study was to investigate use of different types on menopausal hormone therapy and risk of breast cancer, from its inception the study was designed to investigate the role of other factors (such as smoking and obesity) and other health outcomes. The overall aim of the study is to investigate common, largely modifiable, risk factors for serious and common diseases in women. As the study has progressed, and the cohort has aged, different conditions have become of greater relevance: for example, dementia, other neurodegenerative and neurological conditions, mental health, cardiovascular disease and stroke, as well as other cancers. The research agenda is peer- reviewed every few years as part of the applications for renewed funding and developments and are determined by the MWS team in the Cancer Epidemiology Unit (CEU) at the University of Oxford, taking into account the importance of age-appropriate research questions, in relation to new evidence and to public health priorities; the availability of sufficiently long follow-up and number of cases for less common diseases; and additional information which becomes available for linkage and which may allow new analytic approaches.

Cancer diagnosis data and death data and HES admitted patient data are required in order to define end points and disease diagnoses in study participants during follow-up. To produce scientifically valid results, it is essential that the whole Million Women Study cohort is linked to information from medical records. With incomplete linkage, or with follow-up using other methods such as participant report, there would be a high risk of the results being biased by selective lack of follow-up e.g. women who have died and many who are ill cannot complete follow-up questionnaires. Up to date details of deaths and of identifiers such as name and postal address are also needed to ensure that when re-contacting participants, this will ensure no mail will be sent to recently-deceased participants and minimise mailing to out of date addresses.

The wealth of data in the MWS requires careful analysis, and any pre-existing conditions and confounding factors need to be considered. For this reason, it requires a complete picture of the medical history of the participant going back as far as NHS linked records permit, and new diagnoses going forward. In order to produce accurate results in a long –term study of broad scope, it is not possible to minimise further the amount of data requested.

Linked data from NHS Digital are used together with personal, social, lifestyle and health information provided to the study by participants completing the study questionnaires. There is baseline data collected for 1.3 million women via the recruitment questionnaire. Postal re-survey questionnaires were sent to all study participants around 3, 8, 12, and 15 years after recruitment, to keep track of participant’s health and to ask some new questions; and selected women also provided answers to an online diet questionnaire and to disease specific questionnaires, e.g. for cervical cancer and breast cancer. In addition, some women were recruited into a sub study collecting blood samples for biochemical and genetic analyses, the Million Women Study: Disease Susceptibility in Women study.

The Million Women Study is an ongoing open-ended cohort study. Results are disseminated nationally and internationally in peer-reviewed open-access papers in research journals including The Lancet and The International Journal of Epidemiology, in related presentations to national and international research and clinical colleagues, in reports to government regulatory bodies, on the study website, through the media and through public outreach work (science open days, talks to participant groups). The study has been running for over 20 years and publications are produced on an ongoing basis. Presentations of findings will be at conferences such as the National Cancer Research Institute (NCRI) Conference in 2019.

The publications do not identify individuals and contain only de-identified data in aggregate form (tabulations and figures showing analysis results at the minimum level of detail required, using small number suppression) in line with the HES Analysis Guide and with ONS terms and conditions relating to the data being shared under this Agreement.

Among over 120 published papers, notable study publications include:

1. Beral V; Million Women Study Collaborators. Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet. 2003;362(9382):419-27.

2. Sweetland S, Beral V, et al. Venous thromboembolism risk in relation to use of different types of postmenopausal hormone therapy in a large prospective study. J Thromb Haemost. 2012;10(11):2277-86.

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013 Jan 12;381(9861):133-41.

4. Travis RC,et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies. J Natl Cancer Inst 2016;108.

5. Kent S,et al. Hospital costs in relation to body-mass index in 1·1 million women in England: a prospective cohort study. Lancet Public Health. 2017;2(5):e214-e222.

Million Women Study results are reported in the media such as the BBC and national newspapers and its outputs reach a worldwide audience. e.g.

http://www.bbc.co.uk/news/health-35052404 - being unhappy or stressed will not make you ill

http://news.bbc.co.uk/1/hi/health/7906355.stm - a drink a day raises cancer risk

http://www.bbc.co.uk/news/health-14220382 - tall people more likely to develop cancer

Future anticipated work using HES, cancer registry and ONS mortality data:

Current funded work includes analyses of hormonal and other risk factors for subtypes of various common cancers (ovarian, colorectal, breast, endometrial and haematological); of diet and risk of cancer, fracture and vascular disease; of lifestyle factors, past health including mental health and long-term risk of dementia; of risk of osteoporosis and fracture in relation to use of bisphosphonate drugs; and of risk factors for different types of stroke. The study team will submit academic papers on the results of these analyses to open-access, high impact, peer-reviewed journals and will continue to liaise with NHS regulatory bodies as appropriate to ensure the results are used to help provide evidence for health guidelines and best clinical practice.

The Million Women Study website at www.millionwomenstudy.org provides links to open access papers and offers free downloads of accessible summaries of findings. The twitter handles at Nuffield Department of Population Health (@Oxford_NDPH) and Oxford Medical Sciences Division (@OxfordMedSci) with a combined following of over 11,000 will promote all significant findings and will disseminate the findings at academic conferences and to lay audiences, as above.

All future outputs will contain only data that is aggregated, with small numbers suppressed in line with the HES Analysis Guide. ONS terms and conditions relating to the data being shared under this agreement will be adhered to.

In the next year or so the following reports, analyses and presentations are planned:

Data Analyses and papers:

1. Long-term effects of dementia associated with midlife body mass index, diet and physical activity in one million UK women to be submitted to the journal Alzheimer’s and Dementia 2018/19.

2. Smoking and dementia incidence: prospective study of one million UK women to be submitted to The Lancet 2018/19.

3. Foods, macronutrients and breast cancer risk in postmenopausal women: a large UK cohort. To be submitted to the International Journal of Epidemiology 2018/19.

4. Diet and risk of different types of stroke. To be submitted to the journal Neurology 2018/19.

Upcoming Conference presentations:

1. Use of Endocrine therapy for breast cancer to be presented to The Early Breast Cancer Triallists Collaborative Group in Oxford June 2018.

2. Timing of alcohol consumption and risk of liver cirrhosis: prospective study of UK women at European Congress of Epidemiology in Lyon July 2018.

3. Histological type as a predictor of ovarian cancer survival at the British Association of Gynaecological Pathologists - Annual Meeting June 2018.

4. Integrating pathology into epidemiology: Challenges and opportunities. European Society of Pathology Academy July 2018.

The Million Women Study research using linked health data has already directly influenced health care. Results showed, for example, that women using hormone replacement therapy are at increased risk of breast cancer; this work, published in 2003, was shared with the Medicines and Healthcare Regulatory Agency and helped inform changes in prescribing guidance both for the UK and elsewhere. It is estimated that tens of thousands of cases of cancer worldwide have been avoided as a result of the subsequent fall in use of hormone therapy. Another example where NHS Digital data contributed to an influential paper was in showing that risk of blood clots after surgery was far higher, and lasted for much longer, than had been previously thought. This work is helping to inform European surgical care guidelines. MWS results on how characteristics of individuals affect participation and outcomes of bowel cancer screening are being incorporated into screening programme development. Recent work on the costs to the NHS of obesity-related conditions, and on breast cancer risk in relation to night shift work, has direct public health relevance. Participants in the Million Women Study are the first generation of women in the UK to have smoked to the same extent as a man and our findings show that female smokers died about 10 years earlier than non-smokers. These effects are much greater than had been reported previously and influence policy.

The University of Oxford requires HES, mortality, cancer registration and PDS data for use in the Million Women Study (MWS).

The MWS is a national study of women’s health funded by Cancer Research UK and the Medical Research Council. The study involves 1.3 million women in England and Scotland, recruited through NHS breast screening clinics between 1996 and 2001, who gave written consent for follow up of their health through their medical records to examine how reproductive, lifestyle and other factors affect their future health. While the initial stimulus of the study was to investigate use of different types on menopausal hormone therapy and risk of breast cancer, from its inception the study was designed to investigate the role of other factors (such as smoking and obesity) and other health outcomes. The overall aim of the study is to investigate common, largely modifiable, risk factors for serious and common diseases in women. As the study has progressed, and the cohort has aged, different conditions have become of greater relevance: for example, dementia, other neurodegenerative and neurological conditions, mental health, cardiovascular disease and stroke, as well as other cancers. The research agenda is peer- reviewed every few years as part of the applications for renewed funding and developments and are determined by the MWS team in the Cancer Epidemiology Unit (CEU) at the University of Oxford, taking into account the importance of age-appropriate research questions, in relation to new evidence and to public health priorities; the availability of sufficiently long follow-up and number of cases for less common diseases; and additional information which becomes available for linkage and which may allow new analytic approaches.

Cancer diagnosis data and death data and HES admitted patient data are required in order to define end points and disease diagnoses in study participants during follow-up. To produce scientifically valid results, it is essential that the whole Million Women Study cohort is linked to information from medical records. With incomplete linkage, or with follow-up using other methods such as participant report, there would be a high risk of the results being biased by selective lack of follow-up e.g. women who have died and many who are ill cannot complete follow-up questionnaires. Up to date details of deaths and of identifiers such as name and postal address are also needed to ensure that when re-contacting participants, this will ensure no mail will be sent to recently-deceased participants and minimise mailing to out of date addresses.

The wealth of data in the MWS requires careful analysis, and any pre-existing conditions and confounding factors need to be considered. For this reason, it requires a complete picture of the medical history of the participant going back as far as NHS linked records permit, and new diagnoses going forward. In order to produce accurate results in a long –term study of broad scope, it is not possible to minimise further the amount of data requested.

Linked data from NHS Digital are used together with personal, social, lifestyle and health information provided to the study by participants completing the study questionnaires. There is baseline data collected for 1.3 million women via the recruitment questionnaire. Postal re-survey questionnaires were sent to all study participants around 3, 8, 12, and 15 years after recruitment, to keep track of participant’s health and to ask some new questions; and selected women also provided answers to an online diet questionnaire and to disease specific questionnaires, e.g. for cervical cancer and breast cancer. In addition, some women were recruited into a sub study collecting blood samples for biochemical and genetic analyses, the Million Women Study: Disease Susceptibility in Women study.

The Million Women Study is an ongoing open-ended cohort study. Results are disseminated nationally and internationally in peer-reviewed open-access papers in research journals including The Lancet and The International Journal of Epidemiology, in related presentations to national and international research and clinical colleagues, in reports to government regulatory bodies, on the study website, through the media and through public outreach work (science open days, talks to participant groups). The study has been running for over 20 years and publications are produced on an ongoing basis. Presentations of findings will be at conferences such as the National Cancer Research Institute (NCRI) Conference in 2019.

The publications do not identify individuals and contain only de-identified data in aggregate form (tabulations and figures showing analysis results at the minimum level of detail required, using small number suppression) in line with the HES Analysis Guide and with ONS terms and conditions relating to the data being shared under this Agreement.

Among over 120 published papers, notable study publications include:

1. Beral V; Million Women Study Collaborators. Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet. 2003;362(9382):419-27.

2. Sweetland S, Beral V, et al. Venous thromboembolism risk in relation to use of different types of postmenopausal hormone therapy in a large prospective study. J Thromb Haemost. 2012;10(11):2277-86.

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013 Jan 12;381(9861):133-41.

4. Travis RC,et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies. J Natl Cancer Inst 2016;108.

5. Kent S,et al. Hospital costs in relation to body-mass index in 1·1 million women in England: a prospective cohort study. Lancet Public Health. 2017;2(5):e214-e222.

Million Women Study results are reported in the media such as the BBC and national newspapers and its outputs reach a worldwide audience. e.g.

http://www.bbc.co.uk/news/health-35052404 - being unhappy or stressed will not make you ill

http://news.bbc.co.uk/1/hi/health/7906355.stm - a drink a day raises cancer risk

http://www.bbc.co.uk/news/health-14220382 - tall people more likely to develop cancer

Future anticipated work using HES, cancer registry and ONS mortality data:

Current funded work includes analyses of hormonal and other risk factors for subtypes of various common cancers (ovarian, colorectal, breast, endometrial and haematological); of diet and risk of cancer, fracture and vascular disease; of lifestyle factors, past health including mental health and long-term risk of dementia; of risk of osteoporosis and fracture in relation to use of bisphosphonate drugs; and of risk factors for different types of stroke. The study team will submit academic papers on the results of these analyses to open-access, high impact, peer-reviewed journals and will continue to liaise with NHS regulatory bodies as appropriate to ensure the results are used to help provide evidence for health guidelines and best clinical practice.

The Million Women Study website at www.millionwomenstudy.org provides links to open access papers and offers free downloads of accessible summaries of findings. The twitter handles at Nuffield Department of Population Health (@Oxford_NDPH) and Oxford Medical Sciences Division (@OxfordMedSci) with a combined following of over 11,000 will promote all significant findings and will disseminate the findings at academic conferences and to lay audiences, as above.

All future outputs will contain only data that is aggregated, with small numbers suppressed in line with the HES Analysis Guide. ONS terms and conditions relating to the data being shared under this agreement will be adhered to.

In the next year or so the following reports, analyses and presentations are planned:

Data Analyses and papers:

1. Long-term effects of dementia associated with midlife body mass index, diet and physical activity in one million UK women to be submitted to the journal Alzheimer’s and Dementia 2018/19.

2. Smoking and dementia incidence: prospective study of one million UK women to be submitted to The Lancet 2018/19.

3. Foods, macronutrients and breast cancer risk in postmenopausal women: a large UK cohort. To be submitted to the International Journal of Epidemiology 2018/19.

4. Diet and risk of different types of stroke. To be submitted to the journal Neurology 2018/19.

Upcoming Conference presentations:

1. Use of Endocrine therapy for breast cancer to be presented to The Early Breast Cancer Triallists Collaborative Group in Oxford June 2018.

2. Timing of alcohol consumption and risk of liver cirrhosis: prospective study of UK women at European Congress of Epidemiology in Lyon July 2018.

3. Histological type as a predictor of ovarian cancer survival at the British Association of Gynaecological Pathologists - Annual Meeting June 2018.

4. Integrating pathology into epidemiology: Challenges and opportunities. European Society of Pathology Academy July 2018.

The Million Women Study research using linked health data has already directly influenced health care. Results showed, for example, that women using hormone replacement therapy are at increased risk of breast cancer; this work, published in 2003, was shared with the Medicines and Healthcare Regulatory Agency and helped inform changes in prescribing guidance both for the UK and elsewhere. It is estimated that tens of thousands of cases of cancer worldwide have been avoided as a result of the subsequent fall in use of hormone therapy. Another example where NHS Digital data contributed to an influential paper was in showing that risk of blood clots after surgery was far higher, and lasted for much longer, than had been previously thought. This work is helping to inform European surgical care guidelines. MWS results on how characteristics of individuals affect participation and outcomes of bowel cancer screening are being incorporated into screening programme development. Recent work on the costs to the NHS of obesity-related conditions, and on breast cancer risk in relation to night shift work, has direct public health relevance. Participants in the Million Women Study are the first generation of women in the UK to have smoked to the same extent as a man and our findings show that female smokers died about 10 years earlier than non-smokers. These effects are much greater than had been reported previously and influence policy.

The University of Oxford requires HES, mortality, cancer registration and PDS data for use in the Million Women Study (MWS).

The MWS is a national study of women’s health funded by Cancer Research UK and the Medical Research Council. The study involves 1.3 million women in England and Scotland, recruited through NHS breast screening clinics between 1996 and 2001, who gave written consent for follow up of their health through their medical records to examine how reproductive, lifestyle and other factors affect their future health. While the initial stimulus of the study was to investigate use of different types on menopausal hormone therapy and risk of breast cancer, from its inception the study was designed to investigate the role of other factors (such as smoking and obesity) and other health outcomes. The overall aim of the study is to investigate common, largely modifiable, risk factors for serious and common diseases in women. As the study has progressed, and the cohort has aged, different conditions have become of greater relevance: for example, dementia, other neurodegenerative and neurological conditions, mental health, cardiovascular disease and stroke, as well as other cancers. The research agenda is peer- reviewed every few years as part of the applications for renewed funding and developments and are determined by the MWS team in the Cancer Epidemiology Unit (CEU) at the University of Oxford, taking into account the importance of age-appropriate research questions, in relation to new evidence and to public health priorities; the availability of sufficiently long follow-up and number of cases for less common diseases; and additional information which becomes available for linkage and which may allow new analytic approaches.

Cancer diagnosis data and death data and HES admitted patient data are required in order to define end points and disease diagnoses in study participants during follow-up. To produce scientifically valid results, it is essential that the whole Million Women Study cohort is linked to information from medical records. With incomplete linkage, or with follow-up using other methods such as participant report, there would be a high risk of the results being biased by selective lack of follow-up e.g. women who have died and many who are ill cannot complete follow-up questionnaires. Up to date details of deaths and of identifiers such as name and postal address are also needed to ensure that when re-contacting participants, this will ensure no mail will be sent to recently-deceased participants and minimise mailing to out of date addresses.

The wealth of data in the MWS requires careful analysis, and any pre-existing conditions and confounding factors need to be considered. For this reason, it requires a complete picture of the medical history of the participant going back as far as NHS linked records permit, and new diagnoses going forward. In order to produce accurate results in a long –term study of broad scope, it is not possible to minimise further the amount of data requested.

Linked data from NHS Digital are used together with personal, social, lifestyle and health information provided to the study by participants completing the study questionnaires. There is baseline data collected for 1.3 million women via the recruitment questionnaire. Postal re-survey questionnaires were sent to all study participants around 3, 8, 12, and 15 years after recruitment, to keep track of participant’s health and to ask some new questions; and selected women also provided answers to an online diet questionnaire and to disease specific questionnaires, e.g. for cervical cancer and breast cancer. In addition, some women were recruited into a sub study collecting blood samples for biochemical and genetic analyses, the Million Women Study: Disease Susceptibility in Women study.

The Million Women Study is an ongoing open-ended cohort study. Results are disseminated nationally and internationally in peer-reviewed open-access papers in research journals including The Lancet and The International Journal of Epidemiology, in related presentations to national and international research and clinical colleagues, in reports to government regulatory bodies, on the study website, through the media and through public outreach work (science open days, talks to participant groups). The study has been running for over 20 years and publications are produced on an ongoing basis. Presentations of findings will be at conferences such as the National Cancer Research Institute (NCRI) Conference in 2019.

The publications do not identify individuals and contain only de-identified data in aggregate form (tabulations and figures showing analysis results at the minimum level of detail required, using small number suppression) in line with the HES Analysis Guide and with ONS terms and conditions relating to the data being shared under this Agreement.

Among over 120 published papers, notable study publications include:

1. Beral V; Million Women Study Collaborators. Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet. 2003;362(9382):419-27.

2. Sweetland S, Beral V, et al. Venous thromboembolism risk in relation to use of different types of postmenopausal hormone therapy in a large prospective study. J Thromb Haemost. 2012;10(11):2277-86.

3. Pirie K, Peto R, et al. The 21st century hazards of smoking and benefits of stopping: a prospective study of one million women in the UK. Lancet. 2013 Jan 12;381(9861):133-41.

4. Travis RC,et al. Night Shift Work and Breast Cancer Incidence: Three Prospective Studies and Meta-analysis of Published Studies. J Natl Cancer Inst 2016;108.

5. Kent S,et al. Hospital costs in relation to body-mass index in 1·1 million women in England: a prospective cohort study. Lancet Public Health. 2017;2(5):e214-e222.

Million Women Study results are reported in the media such as the BBC and national newspapers and its outputs reach a worldwide audience. e.g.

http://www.bbc.co.uk/news/health-35052404 - being unhappy or stressed will not make you ill

http://news.bbc.co.uk/1/hi/health/7906355.stm - a drink a day raises cancer risk

http://www.bbc.co.uk/news/health-14220382 - tall people more likely to develop cancer

Future anticipated work using HES, cancer registry and ONS mortality data:

Current funded work includes analyses of hormonal and other risk factors for subtypes of various common cancers (ovarian, colorectal, breast, endometrial and haematological); of diet and risk of cancer, fracture and vascular disease; of lifestyle factors, past health including mental health and long-term risk of dementia; of risk of osteoporosis and fracture in relation to use of bisphosphonate drugs; and of risk factors for different types of stroke. The study team will submit academic papers on the results of these analyses to open-access, high impact, peer-reviewed journals and will continue to liaise with NHS regulatory bodies as appropriate to ensure the results are used to help provide evidence for health guidelines and best clinical practice.

The Million Women Study website at www.millionwomenstudy.org provides links to open access papers and offers free downloads of accessible summaries of findings. The twitter handles at Nuffield Department of Population Health (@Oxford_NDPH) and Oxford Medical Sciences Division (@OxfordMedSci) with a combined following of over 11,000 will promote all significant findings and will disseminate the findings at academic conferences and to lay audiences, as above.

All future outputs will contain only data that is aggregated, with small numbers suppressed in line with the HES Analysis Guide. ONS terms and conditions relating to the data being shared under this agreement will be adhered to.

In the next year or so the following reports, analyses and presentations are planned:

Data Analyses and papers:

1. Long-term effects of dementia associated with midlife body mass index, diet and physical activity in one million UK women to be submitted to the journal Alzheimer’s and Dementia 2018/19.

2. Smoking and dementia incidence: prospective study of one million UK women to be submitted to The Lancet 2018/19.

3. Foods, macronutrients and breast cancer risk in postmenopausal women: a large UK cohort. To be submitted to the International Journal of Epidemiology 2018/19.

4. Diet and risk of different types of stroke. To be submitted to the journal Neurology 2018/19.

Upcoming Conference presentations:

1. Use of Endocrine therapy for breast cancer to be presented to The Early Breast Cancer Triallists Collaborative Group in Oxford June 2018.

2. Timing of alcohol consumption and risk of liver cirrhosis: prospective study of UK women at European Congress of Epidemiology in Lyon July 2018.

3. Histological type as a predictor of ovarian cancer survival at the British Association of Gynaecological Pathologists - Annual Meeting June 2018.

4. Integrating pathology into epidemiology: Challenges and opportunities. European Society of Pathology Academy July 2018.

The Million Women Study research using linked health data has already directly influenced health care. Results showed, for example, that women using hormone replacement therapy are at increased risk of breast cancer; this work, published in 2003, was shared with the Medicines and Healthcare Regulatory Agency and helped inform changes in prescribing guidance both for the UK and elsewhere. It is estimated that tens of thousands of cases of cancer worldwide have been avoided as a result of the subsequent fall in use of hormone therapy. Another example where NHS Digital data contributed to an influential paper was in showing that risk of blood clots after surgery was far higher, and lasted for much longer, than had been previously thought. This work is helping to inform European surgical care guidelines. MWS results on how characteristics of individuals affect participation and outcomes of bowel cancer screening are being incorporated into screening programme development. Recent work on the costs to the NHS of obesity-related conditions, and on breast cancer risk in relation to night shift work, has direct public health relevance. Participants in the Million Women Study are the first generation of women in the UK to have smoked to the same extent as a man and our findings show that female smokers died about 10 years earlier than non-smokers. These effects are much greater than had been reported previously and influence policy.