European Prospective Investigation into Cancer and Nutrition (EPIC) - Oxford ( ODR1516_018 )

The University of Oxford requires data from NHS England for the purpose of the EPIC-Oxford study - a nationwide longitudinal cohort study of 60,642 men and women aged 20 and above who were recruited between 1993 and 1999 from throughout the UK.

The study was designed to examine the effects of diet on long-term health, with a specific focus on vegetarians. 50% of the participants do not eat meat, with large numbers following lacto-vegetarian and vegan diets, and EPIC-Oxford is the only large prospective study in the world with dietary data and stored blood samples for a large number of vegetarians together with linkage for the whole cohort to medical records covering cancer diagnoses, hospitalisations and causes of death.

EPIC’s research on the long-term health of vegetarians is unique in the world and is supported by grants from the MRC (“Health of Vegetarians”), the Wellcome Trust (“Livestock, Environment and People”), Cancer Research UK (Aetiology of prostate cancer”) and the World Cancer Research Fund (“Cancer risk in vegetarians”). All this research funded by competitively awarded grants from government and charities, which is focused mainly on cardiovascular diseases, cancer, bone and joint health, and gastro-intestinal diseases, is completely dependent on continued ability to link the whole EPIC-Oxford cohort with the records from HES.

The study is needed to improve understanding of the effects of diet on health and thus inform advice to governments, health professionals and the public about dietary choices to maximise the potential for long-term good health. Further aims include examining the roles of other lifestyle factors (including shift-work) and of endogenous hormones in relation to health.

The study’s overall aim is to provide reliable evidence on choices people can make in adult life to help increase their chances of staying healthy into old age. The aim of the scientific research is to reliably inform the public and health providers and regulators about the statistical findings on risk factors including diet and lifestyle and environmental factors and risk of cancer and other medical conditions.

The data controller is the University of Oxford, and all data is processed in the Nuffield Department of Population Health (NDPH) at the University of Oxford. The University of Oxford is the sole data processor.

All data collected is processed lawfully under GDPR Article 6(1)(e) as a public task in the public interest and under Article 9(2)(j) (‘research’). The use of ‘Article 6(1)e – 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 risks and diet which is 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’ as, in line with the University’s principle objectives previously indicated, the University will be undertaking processing in line with the research protocol stated in this application. The research protocol has received due consideration for its public interest through peer review by a public funder, appropriate ethics committee review and supporting evidence of which is included as part of this application.

Participants were recruited into EPIC-Oxford through; GP practices; through the memberships of the Vegetarian Society and the Vegan Society and through health food shops and magazines. Participants are both male and female and aged 20 or over at recruitment. They were initially asked to complete a questionnaire about their diet and lifestyle, family history, a food frequency questionnaire and to donate a blood sample at recruitment between 1993 and 1999. All surviving participants also provided written consent to follow up their medical records at this time. Participants who were recruited via GP Practices were also asked for blood pressure and anthropometric measurements. Follow up questionnaires asking about diet, lifestyle and health were sent to all surviving participants approximately 5, 10 and 15 years after recruitment. Following recruitment, all surviving participants were also asked to complete a 7-day food diary to record all food consumed in one week.

When EPIC Oxford commenced in 1993 the records available were for cancer registrations and causes of death. Linkage to data from HES became possible after the completion of the recruitment to EPIC-Oxford, and linkage to HES was first established in 2008.

The following datasets are processed by the study and disseminated under two DARS agreements:

DARS-NIC-148322

-Demographics

-Civil Registrations of Death

-MRIS - Cohort Event Notification Report

-MRIS - Flagging Current Status Report

-Hospital Episode Statistics Admitted Patient Care (HES APC)

-MRIS - Cause of Death Report

-MRIS - Members and Postings Report

-Cancer Registration Data

DARS-NIC-656754

-NDRS Cancer Registration data

There are currently 60,642 EPIC-Oxford participants flagged at NHS England. The HES data required is record level pseudonymised data where the date fields only contain month and year. All morbidities and events are required since recruitment to the study in order to have a full record of cancers and deaths for the primary endpoint and to be able to exclude any co-morbidities and confounding variables from the analysis. A complete history of fact, cause and place of death including text fields are required to have a complete history of mortality for each participant to be able to conduct full analysis of the data and to exclude any confounding co-morbidities to produce accurate and reliable results.

Linkage to data for information on detailed NDRS cancer diagnoses and treatment is needed in order to examine the relationships of dietary, lifestyle and other potential risk factors with subsequent health. The aim is to contribute to knowledge of the epidemiology and aetiology of cancers. The priority endpoints which will be studied until 30/04/2026 and beyond are cancers of the breast, prostate, ovary, endometrium, thyroid and colorectum. Without the linked data provided by the NCRAS datasets, EPIC-Oxford researchers will not be able to fully understand the complete burden of the diseases of interest within the cohort. A complete history of the cancer diagnoses, as well as hospital admissions and fact and cause of death, is required to have a comprehensive documentation of conditions for each participant to be able to conduct full analysis of the data and also to exclude any confounding co-morbidities to produce accurate and reliable results.

The pathological and diagnostic data from the cancer outcomes dataset requested in this application will be used in epidemiological analyses of dietary, hormonal and other environmental and biological factors in relation to cancer risk and will also enable the study team to subdivide cancer types according to pathological features of the cancers. Some of the tumour sites being investigated are breast, prostate, ovarian, thyroid and endometrial cancer. The study team aim to expand our collection of data on tumours, allowing the for more tumour characteristics on more types of cancer, at lower cost. The study team require more detailed information on tumour characteristics to enable the study team to conduct further analyses of diet and other risk factors in relation to the incidence of sub-types of cancer, such as prostate cancers subdivided by TNM stage and Gleason score, and breast cancers subdivided by hormone receptor status.

Participants were recruited from England, Wales, Scotland and Northern Ireland, and may have moved since randomisation so coverage of England and Wales for cancer diagnosis and deaths and England for HES is crucial to follow-up all participants. Approval also exists for these data to be collected in Scotland (cancer, death and hospital admissions) and Wales (for hospital admissions only). It would be complicated, costly and burdensome to collect this information any other way, and full coverage of the cohort would not be achieved. The data has been restricted to cohort participants, with the minimum amount of information necessary to answer the research question e.g. only month and year provided for date fields for HES data. NHS England data is necessary because it provides the key mechanism for researchers to work on diseases that are of interest, and the mechanism to exclude any potentially confounding co-morbidities in these datasets. Under this Agreement HES data will be re-supplied for the whole cohort for the period 1997/98 to 2015/16 and new data supplied from 2016/17 up to the latest available. This is because the HES data (supplied under DARS-NIC-148322) is on the basis of participant informed consent rather than the previous CAG Section 251 support, which required national data opt-outs to be supplied.

The priority endpoints which will be studied over the next five years are ischaemic heart disease, stroke, fractures, joint disorders, digestive tract disorders and cancers of the breast, prostate and colorectum. Without the linked data provided by NHS England EPIC-Oxford researchers will not be able to fully understand the full burden of the diseases of interest within the cohort.

Study participants’ records were linked electronically to Hospital Episode Statistics for information on cause-specific hospital admissions, for example cancer diagnoses, cardiovascular disease, joint replacements and fractures. This was to examine the relationships between dietary, lifestyle and other potential risk factors with subsequent health. The aim is to contribute to knowledge of the epidemiology and aetiology of common diseases and other causes of hospital admissions. One of the primary outcomes is cause of mortality so continued receipt of this data will be required in the future.

The collection of these data provides minimal risk of harm to the participants.

Follow-up in a cohort study needs to be as complete as possible. This is to avoid both unnecessary reduction in size of the cohort (which reduces statistical power of data analyses), and loss-to-follow-up bias (which can cause misleading research results). To produce scientifically valid results, it is essential that the whole EPIC-Oxford cohort can be linked with information from medical records. If linkage was not complete there would be a high risk of the results being biased by showing falsely low rates of disease in some dietary groups.

EPIC-Oxford is committed to ensuring participants continue to be kept aware of progress and developments in the study, and to obtain their opinions on current and proposed research and how their data is being used in the study. The unanimous consensus of the study participant panel is that they strongly support the study’s aims and the research that is currently taking place and the future plans. The panel feel that they had been made aware at recruitment that they had signed up to a long term cohort study and they expected their health records to be followed up over many years. The investigators have not changed the scope of the study since the original participant information sheet and protocol were approved initially in 1993 and a further review by the Ethics committee and CAG took place in 2018. It was always intended that the cohort be followed up for diet and cancer risk and other chronic illnesses.

Updates to the protocol have expanded some of these groups, all within the scope of the original objectives. The participants are also aware that the study would run for a further 10 years.

EPIC-Oxford is an independent study with its own research aims with the University of Oxford acting as the sole data controller. From the initial set-up, the study is also part of a larger collaboration called EPIC-Europe. EPIC-Europe is a collaboration between 23 centres in 10 European countries with a total of 521000 participants. The aims of EPIC-Europe are broadly the same as EPIC-Oxford, to study the effects of diet and lifestyle and environmental factors on cancer and other chronic illnesses. The centres involved and the numbers of participants in each country details can be found at the EPIC-Europe web site (https://epic.iarc.fr/). EPIC-Europe is conducting a large number of analyses on the associations of diet with cancer risk, with a focus on the most common cancers such as cancers of the stomach, colorectum (large bowel), breast, prostate and lung. So far, research has focused on the role of fruit and vegetables, dietary fibre, meat, fish, dairy products and alcohol intake, as well as dietary-related factors such as obesity and physical activity, on cancer risk. More details of the findings from EPIC-Europe are found in the EPIC-Europe publications listed on the EPIC Europe website. This will be the focus of a separate data sharing application for EPIC Europe (DARS-NIC-340646-C8Z6J).

This nationwide cohort study was designed to examine the effects of diet on long-term health. All EPIC-Oxford participants provided written informed consent at recruitment to the study in the 1990s. When the study first commenced the records available were for cancer registrations and causes of death. Linkage to data from HES became possible after the completion of the recruitment to EPIC-Oxford, and linkage to HES was established in 2008.

The EPIC-Oxford study team has engaged with the EPIC participant panel by asking them ‘Given the time that has elapsed is it still reasonably expected that the study team are collecting hospital data as part of their medical records’. The participant panel were explicitly asked if they would expect ‘their medical records’ to cover HES data. The panel were also asked whether they felt that the consent given when enrolling into the study met their reasonable expectation that all medical records including deaths, cancer diagnosis and hospital admissions were part of their medical record.

The panel declared that it was and still is their expectation that hospital data, as well as data on cancers and deaths will be received by the EPIC-Oxford researchers as part of their medical record. The patient group ultimately feel that processing of the data is in line with the reasonable expectations of participants who consented to participate in the study.

The participant panel also contribute ideas for disseminating the results of EPIC-Oxford to the study participants e.g via vegetarian society and vegan society publications which the investigators are now looking into.

The national data opt-out does not apply where explicit consent has been obtained from the patient for the specific purpose.

Where individuals have opted out of disease registration by the National Disease Registration Service (NDRS), their data has been permanently removed from the registry and therefore will not be disseminated under this Data Sharing Agreement (DSA). https://digital.nhs.uk/ndrs/patients/opting-out

The identifying information shared with NHS England included names, dates of birth, NHS Number, gender, address and postcode. NHS England retains this information and consequently, there is no requirement for further data to flow from EPIC-Oxford to NHS England unless a participant has moved to England from Scotland and requires to be flagged, or a participant has withdrawn from the study and requires to be removed from the flagged cohort whereupon no further data about this participant will be supplied from NHS England.

A complete history of the cancer diagnoses, hospital admissions and fact and cause of death are required to have a complete history of conditions for each participant to be able to conduct full analysis of the data and to exclude any confounding co-morbidities to produce accurate and reliable results.

NHS England send identifying Demographic, Civil Registration mortality and Cancer data back to University of Oxford including supplied date of birth, Supplied Gender, Supplied Member Number, Cause of Death text and cancer registration details. Cause of death text is important because it enables researchers to determine whether or not the ICD codes provided refer to diseases/conditions that were on the causal pathway. Cancer registration number is important because it enables researchers to correctly identify which cancer is to be deleted in the event of a cancelled cancer; cancer site and morphology (type) data are essential and cancer anniversary year is required to check on the date of diagnosis. Supplied gender and date of birth are necessary for checking that the data really do refer to the participant identified by the member number.

The HES data provided by NHS England is pseudonymised record level data and the participants are not re-identified.

The Demographic, Civil Registration mortality and Cancer Data is held separately from the HES data, When an analysis file is created the fields are merged using a pseudonymised ID into a file containing no other identifying information. During this process there will be no attempt to re-identify any individuals who are taking part in the study.

These data from NHS England provide the study with the most accurate and cost-effective method of participant follow-up. Using any other means would be laborious and costly, and the results would not be as accurate or comprehensive than those that are collected centrally at NHS England.

The data is received from NHS England by a single named Statistician, who processes the files, storing the data, and creating an updated events data file from which analysis files for researchers and other statisticians who work on the study are generated. The analysis files contain the minimum of individual level data required to answer the research question, and are pseudonymised, containing no identifiable data. During this process there will be no attempt to re-identify any individuals who are taking part in the study.

Two named EPIC-Oxford Statisticians have access to the main administrative database, which includes identifiable information on participants (allowing for linkage to health-related records) and is located and maintained separately from the research database (which contains all the data of analysis files for researchers but no data from which any participant could be identified).

1. HES data and identifying Demographic, Civil Registration mortality and Cancer data are downloaded from the NHS England server to Oxford by a named statistician.

2. Pseudonymised NHS England data including event dates and particulars are stored in events databases in NDPH at the University of Oxford, with access limited to named individuals. These databases do not contain participant identifying information (name, date of birth, NHS number, gender, and postcode).

3. Identifying participant information linked to the study ID numbers are stored separately to the dataset for use in analysis and are held only for administrative purposes and for use in facilitating ongoing data linkage.

4. Data files for analysis by epidemiologists, statisticians and students are created by the named statisticians, merging data from the events databases with baseline, follow-up and biological data provided by participants as required.

5. The analysis dataset contains full date of death for individuals whose deaths were reported. These data files contain only pseudonymised data (gender, month and year of birth, month and year of any events, but no names or other identifiable data), and contain the minimum amount of data required to answer the research question. The analysis dataset (containing the study participant’s linked records from the sources specified above) will not be re-linked with the identifiers. All subsequent analyses use only subsets of the pseudonymised data. All such subsets are customised according to the characteristics relevant to the specific analysis, containing only the minimum data required for the specific purpose.

6. Data files for analysis are retained in a secure location on the server only for as long as necessary, for example, to answer questions from journal correspondents and commentators on published papers, and in order to re-run analyses as required.

The research questions and analyses are overseen by the Chief and Principal Investigators of EPIC Oxford who are employees of the University of Oxford and apply for funding to use these data in accordance with the protocol for the benefit of the public to be able to make choices regarding their diet and lifestyle and their risk of cancer or other chronic illnesses.

The datasets will be pseudonymised as described above before statistical analyses are undertaken. Various types of analyses are undertaken on an ongoing basis for the overarching purpose of assessing cancer incidence, health risks and overall mortality. The data will be held only at the NDPH at the University of Oxford. The data will only be accessed by authorised members of the EPIC-Oxford study team of epidemiologists and statisticians, all of whom are substantive employees of the University of Oxford, or non-contractual DPhil and MSc students who complete a University Research Services form agreeing to terms and conditions of the project, grant and relevant sections of the latest Data Sharing Agreement with NHS England. These are filed in the Research Coordinators office with signed copies sent to the Director of Research Services at University of Oxford. Every member of staff working on these data receives an IG induction and made aware of the terms of use of these data. The staff and students also complete an MRC course on Research and GDPR and confidentiality https://byglearning.co.uk/mrcrsc-lms/course/index.php?categoryid=1 for which a certificate with the appropriate pass mark needs to be submitted before access to data is authorised. The data will only be used for the objectives of the study as described within the Data Sharing Agreement. The EPIC-Oxford study will not share any data supplied by NHS England with any other institution or individual outside of the study team at Oxford University.

The IT security measures in the NDPH include:

• Access control: controlling access to resources using a default “deny-all” policy, user, authentication and policies to commission and decommission access on receipt of the appropriate permissions

• Boundary control: configuration and management of a perimeter firewall and interior firewalls within the network, monitoring, DDOS and other attack mitigation,

• Secure storage: NHS England -supplied data is stored on a dedicated, access-controlled server which facilitates encryption-at-rest and prompt data deletion.

The Data will be accessed by authorised personnel via remote access. The Data will remain on the servers at NDPH at the University of Oxford at all times

For remote access:

- Remote access will only be from secure locations situated within the territory of use (as further restricted elsewhere within the DSA if so done) stated within this DSA;

- Access controls granting users the minimum level of access required are in place;

- Remote access is only via secure connections (e.g., VPNs or secure protocols) to protect data;

- Multifactor authentication (MFA) is required for remote access;

- Device security, including up-to-date software and operating systems, antivirus software, and enabled firewalls are utilised for the remote access;

- All remote access is undertaken within the scope of the organisation’s DSPT (or other security arrangements as per this agreement) and complies with the organisation’s remote access policy.

Diet has been identified as the number one cause for the burden of disease worldwide, and by providing new evidence on the impact of diet on health EPIC-Oxford will contribute to reducing the work and cost to the NHS of diet-related ill-health. The aim of EPIC-Oxford is to improve information on diet in relation to the risk of cancer and other chronic diseases, which offers huge potential for improvements in public health in the UK. The results are published in peer-reviewed publications and presented at conferences, and are also reported through national media.

A major focus of research in EPIC-Oxford is the long-term health of vegetarians and vegans, and this research relates directly to the health of the more than 1.2 million people in the UK who follow vegetarian diets (NHS 2014). The long-term effects on health of a vegetarian diet are not well understood, and little is known about the health effects of a vegan diet. The earlier research in EPIC-Oxford has demonstrated lower risks of ischaemic heart disease, stomach cancer and perhaps haematological cancers in vegetarians compared with non-vegetarians, but understanding of these relationships is incomplete and much further research is needed to assess both the potential beneficial effects of a vegetarian diet and also possible hazards associated with low intakes of some nutrients, such as protein, long-chain n-3 fatty acids, vitamin B12, vitamin D and calcium (particularly in vegans). As well as peer-reviewed scientific publications, the EPIC-Oxford website (www.epic-oxford.org) is used to describe all findings, with lay summaries of findings when appropriate, copies of abstracts, and links to pdfs of full papers. The website provides information both for study participants and for a wider audience in the UK and worldwide and where appropriate findings will also be communicated to the NHS because the research will provide information to underpin their advice, e.g. as on their website: http://www.nhs.uk/Livewell/Vegetarianhealth/Pages/Goingvegetarian.aspx

During 2019 the study team have published four papers on the health of participants in EPIC-Oxford. In a paper in the BMJ, the study reported for the first time ever on the risk of stroke in vegetarians, and found that vegetarians had a higher risk for stroke than meat-eaters, although this was counter-balanced by a lower risk for ischaemic heart disease https://www.bmj.com/content/366/bmj.l4897.long. This paper was published with a linked editorial in the BMJ https://www.bmj.com/content/366/bmj.l5272 and a detailed commentary on the NHS website https://www.nhs.uk/news/food-and-diet/vegetarian-diet-linked-lower-risk-heart-disease-higher-risk-stroke/, as well as articles in 122 news outlets worldwide and widespread media coverage giving an Altmetric score of 2515, which is in the top 1% of all outputs https://bmj.altmetric.com/details/65924701.

Selected publications in 2019

Tong TYN, Appleby PN, Bradbury KE, Perez-Cornago A, Travis RC, Clarke R, Key TJ. Risks of ischaemic heart disease and stroke in meat eaters, fish eaters, and vegetarians over 18 years of follow-up: results from the prospective EPIC-Oxford study. BMJ. 2019 Sep 4;366:l4897.

Knuppel A, Papier K, Key TJ, Travis RC. EAT-Lancet score and major health outcomes: the EPIC-Oxford study. Lancet. 2019 Jul 20;394(10194):213-214.

Papier K, Appleby PN, Fensom GK, Knuppel A, Perez-Cornago A, Schmidt JA, Tong TYN, Key TJ. Vegetarian diets and risk of hospitalisation or death with diabetes in British adults: results from the EPIC-Oxford study. Nutr Diabetes. 2019 Feb 25;9(1):7.

Papier K, Tong TY, Appleby PN, Bradbury KE, Fensom GK, Knuppel A, Perez-Cornago A, Schmidt JA, Travis RC, Key TJ. Comparison of major protein-source foods and other food groups in meat-eaters and non-meat-eaters in the EPIC-Oxford Cohort. Nutrients. 2019 Apr 11;11(4). pii: E824.

The data used from NHS England combined with EPIC Oxford data demonstrates the benefits of the research being in the public interest. The public are able to choose whether their diet choices may affect their health, particularly in heart disease and stroke.

In the Lancet the study reported analyses in EPIC-Oxford on mortality in relation to compliance with the EAT-Lancet dietary score, showing beneficial associations for ischaemic heart disease and diabetes, but no association with stroke and no clear association with mortality. Other 2019 papers reported that vegetarians have a 37% lower risk of diabetes than meat-eaters, and that non-meat-eaters have higher intakes of high-protein meat alternatives (soya, legumes, pulses, nuts, seeds) and other plant-based foods (whole grains, vegetables, fruits) and lower intakes of refined grains, fried foods, alcohol and sugar-sweetened beverages than meat-eaters.

EPIC-Oxford research findings presented to clinicians, policymakers and other academics at conferences (listed below) and via the press as described above will directly benefit health care through the NHS by providing clinicians and other NHS health care professionals with up-to-date evidence-based guidance on the effects of diet on long term health and the risk of death. This will improve clinical health care and inform planners and policy makers to address demands on health and social care in the present and the future.

Examples of recent findings through the linkage of HES data is a recent high profile publication in the BMJ looking at the risks of ischaemic heart disease and stroke in meat eaters, fish eaters, and vegetarians over 18 years of follow-up https://www.ncbi.nlm.nih.gov/pubmed/31484644. This was a newsworthy article and reached a wide audience via the press https://www.bbc.co.uk/news/health-49579820 , with reports in 122 news outlets throughout the world https://bmj.altmetric.com/details/65924701/news. Earlier this year a study paper on diabetes showed that this risk of this disease is much lower in vegetarians than in meat eaters https://www.nature.com/articles/s41387-019-0074-0. Altogether, this research provides more information for government policy makers to advise the general public on eating a healthy diet.

Target publications for 2021/2023 are:

Vegetarian diets and risk of bone fractures: analyses during 2020, publish 2021

Target journal: BMC Medicine

Vegetarian diets and risk of joint disorders: analyses during 2020, publish 2021

Target journal: Rheumatology

Target conference: Society for Social Medicine and Population Health Annual Scientific Meeting, Cambridge 2020

Vitamin B12 and risk of haemorrhagic stroke: analyses during 2021, publish 2022

Target journal: European Heart Journal

Target conference: European Society of Cardiology Congress 2021

Vegetarian diets and risk of digestive tract disorders: analyses during 2021, publish 2022

Target journal: Gut

Target conference: Nutrition Society Summer Symposium, 2021

Diet and risk of cancers of the breast, prostate and colorectum: analyses 2020-2022, publish 2023

Target journal: British Journal of Cancer

Target conference: National Cancer Research Institute Conference 2022

All information regarding collection, processing and storage of participant data including details of participant panel discussions mentioned previously are on the EPIC –Oxford website at www.EPIC-Oxford.org . This website is updated regularly with publications, contains annual newsletters and details of participant panel meetings ensuring that participants are fully informed of how their data are being used. The website also contains details of how participants may opt-out of the study and contact details for the study team, who receive regular e-mails, letters and telephone messages from study participants to update their contact details, or report an illness or a death. Study participants also contact the EPIC-Oxford team about EPIC-Oxford research findings they see in the national press and online.

The website also regularly publishes results from research in EPIC-Oxford, including on diseases for which the information comes from HES (such as ischemic heart disease, stroke and fractures), and information on these scientific publications is included on the study website. Results are also described in EPIC-Oxford newsletters and have been covered by national media (online and print), and is covered by the Nuffield Department of Population Health website www.ndph.ox.ac.uk . Information regarding publications is also tweeted by the wider Twitter handle for the Nuffield Department of Population Health (@Oxford_NDPH).

Diet has been identified as the number one cause for the burden of disease worldwide, and by providing new evidence on the impact of diet on health, EPIC-Oxford expect to contribute to reducing the work and cost to the NHS of diet-related ill-health.

The aim of EPIC-Oxford is to provide high quality information on the relationships of diet with long-term health, with a focus on the health of vegetarians and vegans. Although much is already known about diet and health, important questions remain, and in particular it is essential to gain a deeper understanding of the health impacts of “plant-based diets” which are increasingly being recommended as part of strategies to combat climate change and for other environmental reasons. EPIC-Oxford is unique in the world in having a large cohort in which 50% of participants do not eat meat, with long follow-up, a biobank, and linkage to medical records to provide complete and objective follow-up.

Publications from EPIC-Oxford constitute a very substantial proportion of the world literature on the long-term health of vegetarians; the only other comparable prospective study is based in California, but this has no biobank or comprehensive linkage to electronic medical records. Based on the research from EPIC-Oxford and other studies, it is now well-established that vegetarian diets are compatible with good long-term health, and this research underpins the ability of the NHS to provide advice supporting healthy eating as a vegetarian; the NHS could not give such advice without the evidence to support the healthiness of vegetarian diets

Obesity, ischaemic heart disease and diabetes are diet-related diseases which put an enormous burden on the health of people in the UK and on the NHS. EPIC-Oxford has shown that, compared with meat-eaters, vegetarians are less likely to be obese (BMJ 1996 Sep 28;313(7060):816-7.), have lower blood cholesterol and a 22% lower risk of ischaemic heart disease (BMJ. 2019 Sep 4;366:l4897.), and a 37% lower risk of diabetes (Nutr Diabetes 2019 Feb 25;9(1):7.). These findings have not only been published in scientific journals and presented at conferences, both scientific and with a lay audience, but have received enormous media interest. For example, the findings on the low risk of ischaemic heart disease in vegetarians, which were published in the BMJ with an accompanying editorial, had an Altmetric score of 2533 which is in the 99th percentile of outputs, with coverage by 132 news outlets worldwide including the BBC and all the major UK newspapers; it is therefore very likely that the findings were read by millions of people in the UK, providing this reliable information direct to the public to increase knowledge and support beneficial dietary changes which in turn will reduce the associated work and cost to the NHS of diet-related ill-health.

Findings from EPIC-Oxford have also provided important evidence used by Public Health England’s Scientific Advisory Committee on Nutrition (SACN) in its expert evaluations of nutritional impacts on health, which underpin government policy through the Department of Health. For example, results from EPIC-Oxford together with EPIC-Europe on the relationship of folic acid with the risk of prostate cancer were part of the evidence considered in SACN’s report on folic acid, while other results from EPIC were evaluated in SACN’s review of carbohydrates. The SACN reports recommended folic acid fortification of the food supply, decreases in the recommended intake of free sugars, and increases in the recommended intake of dietary fibre for the UK population, thus having direct impacts on nutritional policy and the future health of the public