Epidemiological and health services research using routine NHS data:

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: + Civil Registrations of Death

Objective for processing

The former Unit of Health-Care Epidemiology (UHCE), now at the University of Oxford, was [11 words unchanged] undertook research using routinely collected hospital admissions data and mortality data. The original Unit’s overall aims were, and still are, to undertake epidemiological and health services research, in particular now by using routine NHS statistical data and by undertaking studies that use cohort methodologies. Although the original UHCE has now merged into successor research teams within both the Nuffield Department of Population Health (NDPH) and the Nuffield Department of Women’s and Reproductive Health (NDWRH), both within the Medical Sciences Division at the University of Oxford, the same research aims remain. [2 paragraphs unchanged] The UHCE was, historically, very closely associated with the Regional tier of the NHS and with the Department of Health. Although the UHCE is part of the University of Oxford, it was based on the Oxford Regional Health Authority (RHA) site from 1963 until the reorganisation of RHAs in the mid-1990s. In particular, the UHCE worked closely with the Oxford RHA on medical statistics, record linkage (including the Oxford Record Linkage Study (ORLS)), and health services research. From 1998-2005, the Unit had strong service links with the Department of Health’s National Centre for Health Outcomes Development (NCHOD) - the former Unit Director directed the work programme of the Oxford site of NCHOD. As part of the NCHOD work, the DH commissioned the UHCE to construct and analyse English national record-linked HES files, with HES-to-HES linkage and HES-to-mortality linkage, along the lines of the Oxford Record Linkage Study. The former Unit Director was also Co-founder and Scientific Director of the South East England Public Health Observatory from its inception in 2000 until 2005. Since the reorganisation of the NHS and the inception of (formally) Public Health England in 2013, the Unit retains formal links with the Office for Health Improvement and Disparities (OHID; one of the organisations that PHE dissolved into) through honorary research contracts. The contracts formally with PHE were replaced by contracts with OHID after PHE ceased as a legal entity. Whilst no individual-level data is ever shared with OHID, the fulfilment of these contracts by UHCE staff (in producing summary reports and up-to-date research publications) is reliant on the receipt of up-to-date HES and mortality data from NHS Digital. The former UHCE curated the longest running period of hospitalisation data in the UK. This includes the Oxford regional hospital data ,from 1963 and the Hospital in-patient Enquiry for England data from 1968 (a predecessor of HES). These datasets, now residing within UHCE’s successor, the Applied Health Research Unit (AHRU) within the Nuffield Department of Population Health, with the addition of the HES data this application pertains to, creates a dataset spanning over 60 years of hospital admissions in England. This length and breadth of information creates a unique resource that allows research into areas such as hospital trends, mortality rates, disease history and maternal disease links, which could take years to complete using other research methods.. This enables University of Oxford researchers to react extremely swiftly to the health questions of the day and generate evidence to help tackle issues arising for both the health care system and from patients. As a result, UHCE now has the longest running period of hospitalisation data in the UK, which includes the Oxford regional hospital data from 1963, the Hospital in-patient Enquiry for England data from 1968 (a predecessor of HES) and the HES data supplied by NHS Digital. This length and breadth of data places UHCE in a unique position, allowing the Unit to carry out research into areas such as hospital trends, mortality rates, disease history and maternal disease links, which could take years to complete using other research methods, but which UHCE can achieve extremely efficiently. This means UHCE can react extremely swiftly to the health questions of the day and issues arising for the health care system or from patients. For example, in the questions over the ‘week-end effect’, UHCE was able to publish on mortality rates for meningococcal meningitis. There are few other diseases which can act as such a good marker to show the difference between expert treatment or no or suboptimal treatment. The study showed no evidence of an adverse day of the week effect. The nature and scope of the work carried out using this data, described in further detail below, requires that the data be population-level. This pseudonymised data is required to support the following work themes: The UHCE has run a continuous work programme of rolling research, notably using hospital statistics, mortality data, and record linkage, from 1963 to the present. It is part of the University of Oxford’s Nuffield Department of Population Health and is located in the Big Data Institute at the University of Oxford. The HES Admitted Patient Care and civil registration mortality data required from NHS Digital is unfiltered personal pseudonymised health data and is required on a monthly basis. The nature and scope of the work carried out using this data, described in further detail below, requires that the data be population-level; that is, without filtering the subject pool by, e.g. geographical spread, condition cohort, etc. Due to the breadth of the data and the way it is used, there is no objective to be able to identify any person, e.g., for control or cohort groups. This pseudonymised data is required to support the following work themes: [1 paragraph unchanged] UHCE Research is undertaking research into trends in most hospital specialities, and for many [65 words unchanged] the impact of multi-morbidity; and, involving clinicians, attempting to explain the trends. UHCE is also undertaking In addition, studies of trends in the use of hospital care by particular demographic groups including children, adolescents, and the elderly. In addition to the study of individual diseases, UHCE this will also include studies of medical problems defined by behaviour and aetiology (e.g. self-poisoning [7 words unchanged] where appropriate studying age and cohort effects as well as period effects. [2 paragraphs unchanged] UHCE use the The data will be used to analyse the distribution of hospital admission rates across England. Where admission [88 words unchanged] less common conditions such as multiple sclerosis, motor neurone disease or haemophilia. UHCE intend The data are used to keep maintain up to date the ‘atlases’ of disease across England. [1 paragraph unchanged] This The data can be are also used to develop ‘a examine the science of prognosis’. The aim is to study mortality rates following admission for each diagnosis and operation. UHCE focuses There is a focus on diseases and operations for which there is likely to be interest in long-term trends, using the benefit of the five-decade six-decade runs of data. For example, in studies in the former Oxford region UHCE have shown substantial declines in 30-day and 90-day mortality after emergency admission for myocardial infarction and stroke. [7 paragraphs unchanged] Before any study commences, an authorisation procedure is followed to ensure that the project is suitable. In addition, compliance with the accompanying UHCE Publications Policy and NDPH Data Access Policy must be maintained. All data access requests must be authorised by the Head of UHCE the Applied Health Research Unit (AHRU) within the NDPH / Information Asset Owner. The authorisation and assurance process occurs as follows: [1 paragraph unchanged] 2. Data access requirements confirmed and authorised by Head of UHCE AHRU / Information Asset Owner [4 paragraphs unchanged] iii. UHCE NDPH data is suitable for the project [3 paragraphs unchanged] 3. UHCE NDPH Data Declaration signed [5 paragraphs unchanged] In any one year NDPH undertakes, on average, around twenty distinct research [66 words unchanged] datasets both directly (minimised as appropriate) and indirectly (via automated user interface tools). tools as described below). UHCE The research is supported by research grants from NIHR Oxford Biomedical Research Centre and Health Data Research UK [12 words unchanged] Social Care (England) and the devolved administrations, and leading medical research charities). [1 paragraph unchanged] The University of Oxford is the Data Controller and also processes the data for this study. All processing is undertaken within the Nuffield Department of Population Health or the Nuffield Department of Women’s and Reproductive Health, both within the Medical Sciences Division (MSD) at Oxford University, and NDPH Oxford University makes all decisions regarding the data. No other organisations determine the purpose described in this Agreement or process the data for this or any other purpose.

Processing activities

[1 paragraph unchanged] The University of Oxford will amend this agreement if the requested data is needed for research which does not fall under the themes described here. All processing occurs at the within either NDPH site or NDWRH at the University of Oxford specified in this agreement. Oxford.. No record-level data is ever sent outside of NDPH. these two departments. [1 paragraph unchanged] UHCE undertakes The research on hospital statistics, and on mortality, using six different datasets. These are:- [1 paragraph unchanged] b. the Oxford Record Linkage Study, phase 2 (1998 - 2022 present – a regional subset of linked HES) (ORLS2); [2 paragraphs unchanged] e. Hospital Episode Statistics (HES) for England, linked, (1999 - 2022) present) (provisional); f. Death registry (civil registration mortality) data, standalone (1995-2022) (1995-present) (Source: ONS). The five non-death datasets constitute the longest run of hospital data in England. The UHCE dataset for the ORLS (consisting of ORLS1 and ORLS2 subsets) spanning 1963-2016/17 is the longest run of hospital data in England at regional level and, despite the linkage break between the two subsets in 1998, is the only long-running dataset in England with record linkage going back so many decades (to 1963). Death data has been provided directly by ONS from 1995 to 2022. NHS England have provided death data from 2012 onwards, linked to HES. Death data has been provided directly by ONS from 1995 to 2022. NHS Digital have provided death data from 2012 onwards, linked to HES. [1 paragraph unchanged] In the development and use of large datasets of routine health data, UHCE undertakes original research; it collaborates with others on research projects; and it provides support as required to the NHS, DH and their information functions (although record-level data is not shared with or accessed by them). [1 paragraph unchanged] NHS Digital England will supply pseudonymised HES and civil registration mortality (death registry) data to UHCE via a secure file transfer system. The pseudonymised data from NHS Digital includes encrypted fields (NHS Number, Date of Birth, Postcode, local patient identifier) which would be considered identifiable in their raw form. However, these data are encrypted by NHS Digital to a pseudonymised format before dissemination to UHCE giving UHCE no possibility of re-identifying. These data are required to undertake HES to HES and HES to ONS linkage internally within UHCE for the following reasons: The data are held in secure file servers and individual SQL databases, subject to individual user control. Individual-level records are only ever provided to researchers holding an employment or study contract with NDPH or NDWRH at the University of Oxford. - Production of a linked mother-baby hospitalisation and death dataset for maternity and perinatal studies - Objective comparison between matched data used previously within UHCE and that supplied currently from NHS Digital. The University of Oxford do not hold Study ID. HES records for babies born in hospital don’t have the HES ID/Token_person_ID information for the mother, and vice versa. The encrypted identifiers are therefore required to link mothers to babies in order to create a linked mother-baby HES dataset. The data are held in individual SQL databases, subject to individual user control. The HES and civil registration mortality data are processed on receipt so as to have the same database configuration, in order to ‘look like’ the hospital and mortality data in the ORLS. This is done to facilitate the running of the same software across all the datasets held in UHCE. The datasets are all de-identified and are held securely within the UHCE, and individual-level records are only ever provided to researchers holding an employment or study contract with UHCE. Except where specified, the datasets are not linked to one another as individual-level records. When there is a need to construct studies based on data that span the time frame of the years covered by the different datasets – typically in, for example, studies of hospital admissions across five decades – the UHCE software packages are invoked to run the analyses within each dataset to produce aggregated statistical results. At the stage of the aggregated statistics, UHCE software templates are used to bring together the tabular results that span individual results from within each dataset – e.g. electronic tables for admission rates in the years 1968-1985, 1989-1998, 1989/90-2018/19 are brought together – into combined tables for the whole period 1968-2018/19. [1 paragraph unchanged] • Generation of UHCE NDPH data Raw HES and Death data is transformed into UHCE data. an NDPH Hospitalisation dataset. This involves substitution of some values, most notably encrypted identifiers; addition of UHCE NDPH record identifiers, derivation of new values, reordering and restructuring, and change of storage medium (from text file to database and analysis file) • Linkage Analysis of UHCE NDPH hospitalisation data Linkage activities are limited to: o HES-Death linkage to monitor linkage outcomes of same linkage as already carried out by UHCE and NHS Digital o HES-HES linkage to produce mother-baby data for use as described above • Analysis of UHCE data [1 paragraph unchanged] Output: Method: UHCE Our team have developed suites of software, developed over many years in the UHCE, years, which can be used to run analyses across multiple datasets in a [13 words unchanged] to ‘see’ individual level records in order to ‘queue and run’ the software. The UHCE has developed a ‘front end’ software, and the data do not need to be linked to form the analytical software such that, for each new run, the member of staff uses a menu. The menu gives the operator a choice of selecting the ICD code and diagnostic code (or equivalent for operations), the required age groups, selection of gender, selection of people, the calendar or financial years required, whether to select the primary diagnosis or all diagnostic fields, whether to select all cases or just electives or just emergencies, and so on. complete picture. A suite of templates and data manipulation tools then analyses the data in the individual datasets separately (ORLS1, ORLS2, HIPE, HES) [25 words unchanged] set of electronic tables and combines them into overall tables and graphs giving giving, for instance, a (seemingly) continuous runs of trends. trends covering the whole length of all datasets. Similarly, in studies that require data for the full length of the ORLS (1963-latest available), UHCE software packages are invoked to run the analyses within ORLS1 and, separately, within ORLS2 to produce aggregated statistical results based on the data within each of the two datasets separately. At the stage of the production of the aggregated statistics, UHCE software templates are used to bring together the tabular results that span individual results e.g. electronic tables for admission rates in the years 1963-1998 and for 1999 - latest available – into a combined table for the whole period 1963 - latest available. Results can also be produced via custom analysis. With this method of analysis, researchers work directly with data to obtain a specific analytical outcome. This is employed when the suites of software described above are not able to provide the required analytical method, or when custom analysis is swifter. In this case, a custom extract is used, minimised as necessary. Results can also be produced via custom analysis. With this method of analysis, researchers work directly with data to obtain a specific analytical outcome. This is employed when the suites of software described above are not able to provide the required analytical method, or when custom analysis is swifter. [3 paragraphs unchanged] The information systems used by the Unit of Health-Care Epidemiology are secure and comply with the principles outlined in BS7799 (The Code [23 words unchanged] granted via permission from the Data Asset Owner, and full disk encrypted. All the datasets proposed for use in these studies are pseudonymised. They [9 words unchanged] the smallest geographical unit associated with each record is that of the person’s subject’s Lower Super Output Area of residence. UHCE Our researchers have no requirement to re-identify the individuals within the datasets and will make no attempt to re-identify. UHCE Our researchers will not share any record level data and all outputs will be aggregated with small numbers suppressed in line with the HES analysis guide.

Expected output

UHCE is an academic health The research unit, and the outputs of the Unit’s research activities generated are solely for non-commercial purposes for the advancement of knowledge in public [7 words unchanged] Accordingly, all outputs are limited to the fulfilment of these non-commercial purposes. The scientific output outputs produced by the research team are disseminated to researchers, policy makers and the general public. Findings are [11 words unchanged] including publishing under open access licence, at conferences, and in reports to Public NHS England, the Department of Health England. UHCE and Social Care and other relevant bodies. The research team also aims aim to demonstrate its work to a wider public, beyond the scientific community, by regularly participating in public engagement events. A list of publications that arise from the study are regularly submitted to the funders, including National Institute for Heath and Care Research Oxford Biomedical Research Centre and Public Health England, Data Research UK, as annual reports. Team members use research outputs as examples when teaching [24 words unchanged] is open to academics and research staff within the University of Oxford. UHCE works The researchers work closely with a dedicated team teams of communication specialists in the Nuffield Department of Population Health, Oxford Cancer and across the University which coordinates the dissemination of research outputs and operates the communication channels, which includes a regular newsletter. newsletters. They maintain bilateral engagements with funders and stake holders. The communication team works with the external media and runs the social media platforms for the department. All research outputs are intended for a non-commercial use, and UHCE do not expect there is no intention to commercialise any tools or outputs produced during or beyond the project’s completion. [1 paragraph unchanged] An examples of a recent outputs is as follows: Below is a list of exemplar projects which will be undertaken using the requested data. Study on the determinants of the decline in mortality from acute stroke in England: linked national database study of 795,869 adults. After publication in May 2019, the study findings attracted considerable media attention from the national news agencies, and were widely reported on television, radio and in the newspapers, creating positive publicity for the NHS and firmly establishing the rise in young stroke incidence as a national public health issue. The study made front page news in leading national newspapers “The Times” and “The Guardian”. 1. A study investigating the use of colonoscopy and flexible sigmoidoscopy across the English NHS to determine both cancer pick-up rates and complications following the procedures (for example, bowel perforations and bleeding). It is estimated this will be ready for publication in a journal such as Gut in the late 2026. Below is a detailed list of UHCE’s expected publications and other research activities which will be generated using NHS Digital data (all studies listed below have been added since the last independent review of this Data Sharing Agreement in 2019, except projects 10 and 11). 2. Monitoring cancer-related mortality trends and end-of-life care in England. For publication in multiple peer-reviewed journals such as the British Journal of Cancer in 2026-28. 1. A study investigating the impact of the COVID-19 pandemic on the management and outcome of colorectal cancer in the English NHS. In partnership with NHS Digital, the University of Oxford's group has already published one paper (in Lancet Gastroenterology and Hepatology) on this topic in January 2021 using SUS+ data and is working to ensure the results are updated each month. The next phase of the work is to investigate whether the changes in care are linked to outcome so the University of Oxford plan to apply the same and adapted algorithms to the linked HES-mortality file. This is expected to enable the study team, for example, to determine whether the observed increase in stoma formation has been reversed or whether there have been changes in surgical mortality. It is anticipated this work will be published in a similar journal in the late 2022/ early 2023. 3. Studies of the impact of the COVID-19 pandemic on health and healthcare in England for a range of conditions including cardiovascular disease, cancer, infectious diseases, hip fracture, and alcohol-related liver disease, for publication in multiple peer-reviewed journals, such as The Lancet, The BMJ and the Journal of Epidemiology and Community Health, up until 2026/7. 2. Study of rates of COVID-19 related mortality in patients with underlying comorbidities compared with the general population in England. Ongoing outputs have been used in the national COVID-19 response and have now been extended to a collaboration with the RECOVERY trial. Publication in a peer-reviewed journal is expected in in mid-late 2022. 4. Study of inflammatory bowel disease as a risk factor for colorectal cancer. For publication in a peer-reviewed journal such as Lancet Gastroenterology and Hepatology in late-2026. 3. Study investigating trends in major operations for extensive/metastatic pelvic cancers. Algorithm for identifying these operations being developed in partnership with NHS England and peer-reviewed publication to be submitted to a journal such as British Journal of Surgery in late 2022. 5. NIHR-funded project “PReventing Severe Maternal Morbidity and mortality through a national programme of interventions” (PRiSMM), to monitor and evaluate the introduction of a new maternity care bundle being rolled out nationally https://www.england.nhs.uk/long-read/the-maternal-care-bundle/. Monitoring and reporting of maternal morbidity will be reported in a dashboard, starting in Summer 2026. Further work to describe the impact of the Maternal Care Bundle will be conducted in 26/27, with scientific publications to follow. 4. A study investigating the use of colonoscopy and flexible sigmoidoscopy across the English NHS to determine both cancer pick-up rates and complications following the procedures (for example, bowel perforations and bleeding). It is estimated this will be ready for publication in a journal such as Gut in the late 2022 / early 2023. 6. Postgraduate teaching MSc in Global Health Sciences and Epidemiology course in the University of Oxford, using published examples from the linked HES. This is an ongoing annual commitment. 5. An ongoing series of international comparison studies using English data, as a part of a collaborative project with academic partners in Australia, New Zealand and Canada. In the first instance, it aims to compare the determinants of the reduction in mortality rates from myocardial infarction (published) and stroke (current project) in England, Australia, New Zealand and Canada. It uses routinely-collected datasets in all 4 jurisdictions and the analytical methods developed by the WHO MONICA study - a large cohort study - on routine hospital data. The study on myocardial infarction was published in the Lancet Public Health in early March 2022. The current one is a comparison of the determinants of the reduction of stroke mortality and is expected to begin in mid-2022 and to be published in a peer-reviewed journal in 2023. 7. The study team work closely with the Medical School at the University of Oxford and offer supervised research placements for undergraduate medical students. This is an ongoing commitment. 6. Study of the risk of circulatory conditions in patients with Marfan Syndrome. For publication in a peer-reviewed journal in late-2022. 7. Monitoring socioeconomic inequalities in premature and preventable mortality. For publication in peer-reviewed journal such as Science in late-2022. 8. Study of distribution of neonatal surgery in England in response to the Getting It Right First Time (GIRFT) report for Paediatric Surgery, to be published in a peer-reviewed journal in late-2022. 9. Study of ethnic disparities in maternal health, characterised by deprivation and geography. Collaboration with the National Perinatal Epidemiology Unit, reporting to the Minister of State for Mental Health, Suicide Prevention and Patient Safety. For publication in late-2022. 10. Study of long-term trends in hospital admission for stroke and stroke mortality in England. For publication in a peer-reviewed journal such as Stroke in 2022/23. 11. Study of trends and geographical variations in 30-day case fatality rates after acute stroke using linked HES and mortality data. For publication in a peer-reviewed journal such as Stroke in 2022/23. 12. NIHR-funded study of timing of stoma closure in neonates. For publication in a paediatric journal such as Archives of Disease in Childhood in 2022/23. 13. Study of the association between eating disorders, such as anorexia and bulimia, with the risk of cancers. This work is under review with the British Journal of Psychiatry. 14. Several other studies of disease associations employing similar methodology to the above two studies: cancer outcomes in patients with cystic fibrosis; cardiovascular disease outcomes in patients with systemic lupus erythematosus; aortic dissection and aneurysm in patients with giant cell arteritis; cerebral venous thrombosis in pregnancy; bleeds and venous thromboembolism after hip fracture. For publication in peer-reviewed journals in 2022/23. 15. Study of regional long-term trends of hospital admissions for myocardial infarction in England. The University of Oxford's group has recently published on very long-term trends in hospital admissions for myocardial infarction in England as a whole and by sex and age (Journal of Epidemiology & Community Health, July 2021). This study intends to examine rates in different regions of the country, the results of which are expected to be published in a peer-reviewed journal such as Heart in 2023. 16. Study of long-term trends in the incidence and case-fatality of myocardial infarction by ethnicity using linked HES and mortality records. This work is expected to be published in a peer reviewed journal in 2024. 17. A study on the long-term trend in hospital admissions for primary care sensitive conditions for children aged under 15 years. This study is expected to be published in a peer-reviewed journal such as Archives of Disease in Childhood. 18. Postgraduate teaching MSc in Global Health Sciences and Epidemiology course in the University of Oxford, using published examples from the linked HES. This is an ongoing annual commitment. 19. The study team work closely with the Medical School at the University of Oxford and offer supervised research placements for undergraduate medical students. This is an ongoing commitment.

Expected measurable benefits

With reference to each of the individual studies listed above in section 5c, data received from NHS Digital England is expected to provide the following benefits: 1. The published study has determined there is presently a diagnostic deficit of over 4,000 colorectal cancers as a result of the COVID-19 pandemic. In addition, for those people who have been diagnosed, standards of care have changed dramatically. More people are living with stomas than would be expected and many more have had delays to their surgical procedures or alternatives such as radiotherapy. The impact of these changes on outcome is not yet understood. This study will aim to investigate how NHS colorectal cancer services have changed during the pandemic and provide higher resolution information to inform the recovery. 1. Colonoscopy is the main diagnostic test for colorectal cancer and is an extremely commonly used procedure in the NHS. Prior to COVID-19, colonoscopy services were under pressure due to inadequate capacity to keep up with demand but, as the procedures were initially deemed aerosol generating, the current situation has become even more pressured. Evidence is urgently needed, therefore, to increase the quality and efficiency of colonoscopic services. This study hopes to provide this information. It is expected to both enable the pick-up rates of cancer to be determined and help quantify the number of adverse outcomes linked to the procedures. At present, such information is not available for the NHS, which is preventing optimisation of services. 2. This study has directly informed the roll-out of the COVID-19 vaccination priority strategies (current and future) by determining which clinically vulnerable populations will benefit most from early vaccination. Findings were reported directly to the Joint Committee on Vaccinations and Immunisation. 2. Monitoring cancer-related mortality trends and end-of-life care in England will provide critical evidence to evaluate how well the NHS supports people with advanced cancer and to identify inequalities in outcomes and care. Investigation of trends in multiple cause-coded mortality data will allow a more complete understanding of the role that other health conditions play in cancer-related mortality and the complexity of patients’ health at the end of life. Analysing these data alongside indicators such as place of death and emergency/elective care prior to death will enable policymakers and healthcare providers to detect gaps in services, better understand comorbidities and care needs, and guide the planning and allocation of palliative services. 3. The surgical management of metastatic cancers involves major operations, such as pelvic exenterations (the removal of all pelvic organs), that have significant resource implications, are associated with significant morbidity and mortality outcomes for which are poorly understood. Given the scarcity of information on the use of these operations specialist cancer commissioning within NHS England are seeking further information about their use. University of Oxford’s proposed study will aim to provide the first population-based information on these operations in England and so inform the commissioning of cancer services. 3. Monitoring the impact of the COVID-19 pandemic on health and healthcare will provide essential evidence for informed decision-making and effective resource allocation in the NHS. By systematically tracking both the direct effects of infection and the wider consequences for healthcare services and population health, this work will enable policymakers and healthcare providers to identify emerging pandemic-related pressures, prioritise interventions, and target public health messaging where it is most needed. 4. Colonoscopy is the main diagnostic test for colorectal cancer and is an extremely commonly used procedure in the NHS. Prior to COVID-19, colonoscopy services were under pressure due to inadequate capacity to keep up with demand but, as the procedures were initially deemed aerosol generating, the current situation has become even more pressured. Evidence is urgently needed, therefore, to increase the quality and efficiency of colonoscopic services. This study hopes to provide this information. It is expected to both enable the pick-up rates of cancer to be determined and help quantify the number of adverse outcomes linked to the procedures. At present, such information is not available for the NHS, which is preventing optimisation of services. 4. Investigating inflammatory bowel disease as a risk factor for colorectal cancer incidence and mortality will provide important evidence to understand how these inflammatory conditions contribute to cancer incidence and outcomes at the population level. Characterising this association in detail will enable policymakers to assess whether surveillance, screening, prevention, and treatment strategies are effectively reducing risk. 5. The collaborative project on heart attack and stroke is establishing an international network of researchers working on routine hospital statistics in England, Australia, New Zealand and Canada to investigate cardiovascular disease incidence and mortality rates. This facilitated comparison of routine hospital statistics in different countries provide specific measures of the factors affecting trends in mortality rates in these countries. The recently published study reported how much of the reduction in mortality rates from heart attack was due to hospital care versus preventive medicine, and how the contribution of the two differ between the countries. This will hopefully inform and enable better health care planning and resource allocation for the treatment and prevention of heart attack in each country. The success of this project has demonstrated that the methods developed by large and inexpensive studies, such as the WHO MONICA study, can be used by those working with routine hospital statistics. The current study intends to investigate stroke mortality using the same methodology and will hopefully make similar contributions to health care planning and resource allocation for stroke. 5. PReventing Severe Maternal Morbidity and mortality through a national programme of interventions (PRiSMM) will provide national data on maternal morbidity, by ethnicity, deprivation and geographic area. This information will be updated with new data releases, and findings provided in a dashboard that can be used to study trends, inform policy and monitor safety. Morbidity reporting is due by Summer 2026 in a timeline agreed with DHSC. Further work to evaluate the national Maternal Care Bundle, will provide evidence of clinical and cost effectiveness, and inform subsequent iterations of the policy. 6. Marfan Syndrome is a hereditary condition which affects connective tissues. Among most severe complications of Marfan are aortic aneurysms (an enlargement of the aorta). This study is expected to quantify the risk of a range of vascular complications in these patients and report any changes in risks that occurred in recent years compared to the past. The hypothesis is that as the care and prognosis of these individuals have improved, the risk of developing a ruptured aneurysm and other severe and life-threatening vascular conditions declined. These findings are expected to be published in peer-reviewed journal in 2022. 6. Members of the NDPH team teach the bioinformatics and epidemiology modules on MSc in Global Health Sciences and Epidemiology course run by the University of Oxford, using examples of studies using HES. The postgraduate students are the next generation of academics, public health professionals and policy makers. It is important for them to learn about the value of electronic medical data available in England. The course highlights the importance of routine health data for research, as it teaches about multiple ways of using the data. The routine electronic medical records could be used on their own, or linked to cohort studies and clinical trials, which could potentially reduce the costs of running these studies. 7. This study is part of a larger international research project to compare social inequalities in premature and preventable deaths over time, within and across countries. Previous research largely reported the social inequality based on individual-level factors such as education, occupation. This study intends to generate aggregate-level evidence in each of the participating countries to help monitor the social inequalities in premature and preventable mortality at whole-population level. For English data, the University of Oxford will report these by levels of the indices of multiple deprivation. Given there is a substantial difference in the life expectancy of people with regard to their socioeconomic status, it is hoped that this work will provide valuable evidence to help inform interventions to eliminate these inequalities. The training provided by the study team for medical students will hopefully ensure that future doctors in England have the skills and experience required for working with and understanding routine electronic health records. Students will work under the supervision of experienced staff using existing infrastructure. Through teaching future doctors, the study team aim to make a valuable contribution to the digital transformation of care in the NHS and delivering the commitments within the NHS Long Term Plan. 8. The Getting it Right First Time report for Paediatric Surgery has recently been published: https://www.gettingitrightfirsttime.co.uk/surgical-specialty/paediatric-surgery/. It makes striking recommendations for the reorganisation of services, including potentially halving the number of surgical centres. University of Oxford intend to perform a detailed analysis of current hospital service provision and combine this with evidence from other publications and family perspectives to discuss issues around how to optimise quality whilst maintaining good access to services for families, which it is hoped will directly inform the reorganisation of neonatal surgical services. More generally, obtaining this new release (and, indeed, all future new releases – i.e. keeping ‘up-to-date’) of linked HES data and mortality data for all England is intended to enable the research team to produce the most up-to-date analyses of trends and variation in population-based hospital admission rates, incidence and mortality. Without the data collected for the most recent years and months, the research output is less informative and outdated, especially during the COVID-19 pandemic; this will limit the impact of the outputs in the scientific community and among policy makers. UHCE publishes work in international journals and presents at international conferences, and this contributes to building a strong reputation for the UK as a country with established systems of collecting and analysing electronic medical records. Being a part of the University of Oxford, NDPH research output is taken into consideration when ranking academic institutions; with the new data, the Unit will be able to continue producing results and publish papers for recent years, not only historical studies, and this will help to make a substantial contribution to maintaining Oxford’s reputation for scientific excellence. 9. Following publication of a recent MBRRACE report by the National Perinatal Epidemiology Unit (https://www.bmj.com/content/372/bmj.n152), there is growing interest from within central government (https://www.parallelparliament.co.uk/mp/nadine-dorries/debate/Commons/2021-04-19/debates/6935B9C7-6419-4E7B-A813-E852A4EE4F5C/BlackMaternalHealthcareAndMortality) and among the general public (Channel 4’s Dispatches: The Black Maternity Scandal) in uncovering ethnic disparities in maternal health. In collaboration with the National Perinatal Epidemiology Unit, these summary findings were reported to the Minister of State for Mental Health, Suicide Prevention and Patient Safety in late 2021 and will be published in a peer-reviewed journal in late 2022. It is hoped that any disparities uncovered by the study will be used to focus efforts to level up maternity outcomes for women from ethnic backgrounds where inequities are found. 10. The study of long-term trends in hospital admission rates and mortality from stroke measures the instance of stroke in the population. The pattern of hospitalisation trends has important implications for planning of health-care policies and allocation of resources. Monitoring of recent trends in stroke mortality offers an opportunity to record and report any slowdown or plateauing in trends. Analysis of trends in population subgroups, by age, ethnicity or socio-economic status, will help to identify population at risk of developing stroke or dying from stroke, and it is hoped this information could be used for targeted interventions. 11. The analysis of geographical differences in 30-day case fatality rates after acute stroke in England would measure the variations in short-term fatality in different parts of the country, and the findings will help to identify areas with particularly high or low case-fatality rates. This information, which will be reported to Public Health England and is intended to be used by the local health authorities in regions with high rates to improve the provision of care of stroke patients. 12. Optimal timing for neonatal stoma closure remains unclear. In this study University of Oxford will aim to establish current practice on timing of stoma closure using HES, as part of a multidisciplinary NIHR-funded research programme to determine feasibility of a clinical trial comparing ‘early’ and ‘late’ stoma closure, which is hoped will directly inform clinical practice in this area. 13. In collaboration with the Department of Psychiatry at the University of Oxford, the study team conduct research on mental health problems. The study team are working closely with the psychiatrist to do studies with high clinical value that will have an impact on clinical practice. Improving mental health is high on the national agenda, and HES data can be used as a resource for important work. In particular, the study team work on mental health conditions that affect children and young people. The focus of their work for the next year are eating disorders. Eating disorders are increasing in the population, and most commonly affect young people. However, the long-term effects of these conditions are unknown. It is possible that changes in metabolism, hormones, or irritation of the upper gastrointestinal tracts from excessive vomiting, might result in increased risk of cancer. However low body mass index might be protective against some cancers. The literature on this possible association is limited. It is hoped these findings could be used to inform patients suffering from eating disorders and doctors looking after these patients, as well as used for identifying populations at particularly high risk of cancers that might benefit from earlier screening. 14. Characterising associations between diseases can help to uncover shared aetiologies or an increased need for clinical surveillance. As explained elsewhere, the Unit has the ability to run these disease association analyses very swiftly, and the (separate) Oxford Region dataset can be used to complement such findings to determine how these disease associations may have evolved over time. Owing to strong and longstanding links with clinical teams within the Oxford University Hospital NHS Foundation Trust and beyond, University of Oxford are able to target specific disease association studies, such as these, which are of particular and current interest to the clinical community. 15. The recently published very long-term study found that after a steady decline in admissions for heart attack from the 1990s, there was an overall modest increase in admission rates between 2012 and 2016. Analysis of regional variations in hospital admissions for heart attack in England will assess variations in hospital admission rates and the extent of this rise in rates in different parts of the country. This work will report trends beyond 2016, which is important to assess whether the rates continued to rise or not. 16. The study of long-term trends in the incidence and case-fatality from heart attack intends to measure changes in the occurrence of myocardial infarction in the population. The study team will look at trends in rates by sex, age, socioeconomic status, ethnic group and geographical areas. The findings are expected to help inform coronary disease prevention strategies and public awareness campaigns to help focus on those most at risk and has implications for health-care resource planning for those areas of the country with the largest burden of coronary disease. 17. The study of long-term trends in hospital admissions for conditions expected to be treated by general practitioners in children aged under 15 years aims to measure the number and rate of hospital admissions for these conditions. The study team will look at trends by sex, age and area socioeconomic status. This is expected to help inform the delivery and structure of acute paediatric care in both primary and secondary care. 18. Members of the UHCE team teach the bioinformatics and epidemiology modules on MSc in Global Health Sciences and Epidemiology course run by the University of Oxford, using examples of studies using HES. The postgraduate students are the next generation of academics, public health professionals and policy makers. It is important for them to learn about the value of electronic medical data available in England. The course highlights the importance of routine health data for research, as it teaches about multiple ways of using the data. The routine electronic medical records could be used on their own, or linked to cohort studies and clinical trials, which could potentially reduce the costs of running these studies. 19. The training provided by the study team for medical students will hopefully ensure that future doctors in England have the skills and experience required for working with and understanding routine electronic health records. Students will work under the supervision of experienced staff using existing infrastructure. Through teaching future doctors, the study team aim to make a valuable contribution to the digital transformation of care in the NHS and delivering the commitments within the NHS Long Term Plan. More generally, obtaining this new release (and, indeed, all future new releases – i.e. keeping ‘up-to-date’) of linked HES data and mortality data for all England is intended to enable the research team to produce the most up-to-date analyses of trends and variation in population-based hospital admission rates, incidence and mortality. Without the data collected for the most recent years and months, the research output is less informative and outdated, especially during the COVID-19 pandemic; this will limit the impact of UHCE work in the scientific community and among policy makers. UHCE publishes work in international journals and presents at international conferences, and this contributes to building a strong reputation for the UK as a country with established systems of collecting and analysing electronic medical records. Being a part of the University of Oxford, UHCE research output is taken into consideration when ranking academic institutions; with the new data, the Unit will be able to produce results and publish papers for recent years, not only historical studies, and this will help to make a substantial contribution to maintaining Oxford’s reputation for scientific excellence.

Benefits reported

This database and the research carried out within NDPH (Nuffield Department of Population Health) which uses it significantly contributes to the body of evidence and knowledge [117 words unchanged] access to longitudinal patient history when defining patient populations or identifying comorbidities. [1 paragraph unchanged] 1. The work on Early investigation of Covid-19 mortality in patients who are clinically vulnerable (study 2 in 5c, [94 words unchanged] of Covid-19 vaccines to clinically vulnerable patients, most notably those on dialysis. [2 paragraphs unchanged] 4. A national study of over 650,000 patients admitted to hospital for [20 words unchanged] continued disparities in incidence by level of deprivation over the last decade. As part of this work, we are providing This provides continued surveillance of demographic-specific hip fracture rates and health inequalities as health systems recover from the pandemic and as policy evolves to track the success of prevention measures. [1 paragraph unchanged] 6. A national study of trends in admissions for pneumothorax before, during [18 words unchanged] most extensive analyses to date on the incidence of inpatient-treated spontaneous pneumothorax. Our The findings provide novel insights into the impact of the pandemic on the [38 words unchanged] pneumothorax management strategies, particularly in anticipation of potential future public health crises. [1 paragraph unchanged] 1. Identifying ethnic disparities in maternal care and maternal health outcomes (commissioned [15 words unchanged] has provided evidence in support of levelling up maternal health inequities by ethnicity; ethnicity and has directly informed the newly-implemented maternity care bundle (referenced as Geddes-Barton et al, 2024 in https://www.england.nhs.uk/long-read/the-maternal-care-bundle); 2. Related work on deprivation-related disparities in maternal care and maternal health [31 words unchanged] need for interventions that support pregnant women into stable and secure housing (in press with (published in BJOG). The research we are undertaking being undertaken on ethnic and deprivation inequalities in maternal health and perinatal outcomes is in collaboration with the National Perinatal Epidemiology Unit and feeds directly into the NIHR Policy Research Unit in Maternal and Neonatal Health (https://www.npeu.ox.ac.uk/pru-mnhc), directly helping to inform government policy decisions. [2 paragraphs unchanged]

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Benefits reported

This database and the research carried out within NDPH (Nuffield Department of Population Health) which uses it significantly contributes to the body of evidence and knowledge available which leads available, leading to changes in treatment, care and policies which are of benefit to the patient and the health care system. With 50 over 60 years’ worth of hospital data, research into areas such as admission trends, case fatality rates, disease associations and maternal disease links, that could otherwise take years to complete, can be achieved extremely very efficiently. This means UHCE that NDPH can react extremely swiftly to matters that may suddenly come into the [44 words unchanged] access to longitudinal patient history when defining patient populations or identifying comorbidities. In previous years, the University of Oxford have been able to react quickly, for example, to questions over the ‘week-end effect’, where UHCE was able to report rapidly on mortality rates for meningococcal meningitis (there are few other diseases which can act as such a good marker to show the difference between expert treatment or no or suboptimal treatment; the study showed no evidence of an adverse day of the week effect). In recent years, NDPH has notably been able to react swiftly to the impact of the COVID-19 pandemic and its aftermath on cause-specific hospital admissions and mortality. For example: In more recent years/months, the University of Oxford have been able to react swiftly to the impact of the COVID-19 pandemic on cause-specific hospital admissions and mortality. [1 paragraph unchanged] 2. The University of Oxford's NDPH report on the enormous and sustained decreases in hospital admissions for childhood [39 words unchanged] months, to minimise the burden on health systems and protect vulnerable children. The University of Oxford continues NDPH is continuing to provide ongoing updates using the most contemporary data to continually monitor [22 words unchanged] behavioural restrictions as well as the effects of any disruptions to the routine vaccine schedule. schedule and vaccine uptake in the years since the pandemic. 3. The University of Oxford have Work within NDPH has also defined and highlighted through publication the massive under-diagnosis of colorectal cancer [99 words unchanged] helping to minimise delays in diagnosis (which frequently lead to poor outcomes). Aside from the studies undertaken during the pandemic (described above), other recent work has included reacting to other issues in the political or public health spotlight, such as ethnic disparities in maternal care and maternal health outcomes (commissioned by the Minister of State for Mental Health, Suicide Prevention and Patient Safety), work which has provided evidence in support of levelling up maternal health inequities by ethnicity; proposed changes to the provision of neonatal surgery in England (in response to the GIRFT report on paediatric surgery), work which has provided evidence for reconfiguring neonatal surgery with due consideration to provider transfer times, maternity provision, and workforce skill maintenance/training; and the very high mortality rates in liver disease patients (commissioned by The Lancet), a Review which provides eight main recommendations for improving outcomes and survival rates for patients in the UK with both acute and chronic liver disease. 4. A national study of over 650,000 patients admitted to hospital for hip fracture in England, which showed an excess of hip fracture presentations to NHS hospitals since the COVID-19 pandemic, and continued disparities in incidence by level of deprivation over the last decade. As part of this work, we are providing continued surveillance of demographic-specific hip fracture rates and health inequalities as health systems recover from the pandemic and as policy evolves to track the success of prevention measures. Further examples of prior work, in order to illustrate the kind of studies that can be undertaken rapidly in the future and their impact, are as follows: 5. A national study of over 750,000 first time admissions for pulmonary embolism in England, which disentangles the long-term upward trend from the disruptions of the pandemic years, accounting for COVID-19 as a co-existing condition. The study, published in The Lancet Regional Health – Europe, offers an unprecedented insight into the evolution of PE incidence over the past 16 years and the impact of the COVID-19 pandemic. The published findings are accompanied by an updateable public-facing dashboard directed at clinicians and policymakers to guide prevention and treatment strategies, the benefits of which have been recognised by the scientific community, for example in an associated commentary in the same journal: “Zhong and colleagues’ commitment to ongoing, public-facing updates is commendable and should serve as a model for transparent, adaptive health surveillance. As health systems worldwide transition from pandemic crisis response to long-term resilience, integrated data infrastructures that capture the continuum of care for conditions like PE will be fundamental, not only for pandemic preparedness but also for closing gaps in everyday cardiovascular care.” https://doi.org/10.1016/j.lanepe.2025.101450 1. Congenital viral infections in England over five decades: CMV diagnoses have increased dramatically since 2007 when the NHS Newborn Hearing Screening Programme was introduced, providing evidence for the need to identify cases as soon as possible so that antiviral drugs can be administered early enough to prevent life-long hearing loss and neurodevelopmental disability. Reaction to this study later prompted calls (published in The Lancet) for policy makers to begin testing for congenital CMV in those who do not pass their newborn baby hearing screen, which, if implemented, could provide immediate benefits to affected infants by improving developmental outcomes. 6. A national study of trends in admissions for pneumothorax before, during and after the pandemic. The study, published in The Lancet Regional Health - Europe, represents one of the most extensive analyses to date on the incidence of inpatient-treated spontaneous pneumothorax. Our findings provide novel insights into the impact of the pandemic on the incidence of pneumothorax, showing a significant peak in COVID-19 related pneumothoraces during late 2020 and early 2021 and highlighting the role of COVID-19 as a contributing factor. These insights are crucial for healthcare providers and policymakers in optimising pneumothorax management strategies, particularly in anticipation of potential future public health crises. 2. Determinants of the decline in mortality from acute stroke in England: While reporting reductions of 50% in stroke mortality and 20% in stroke incidence overall, this study highlighted the significant role of acute hospital care in the reduction of stroke death overall, but also a worrying increase in stroke incidence in people younger than 55 years. This study prompted a number of large national news outlets to put acute stroke in the spotlight and created a space for public discussion. The study highlighted to policy makers, health professionals and the general public that guidelines for the primary prevention of stroke and public awareness campaigns should include the message that stroke is not just a condition of the elderly. Aside from the studies undertaken during and relating to the pandemic (described above), other recent work has included reacting to other issues in the political or public health spotlight: 3. Hospital admissions for viral meningitis in children in England over five decades: This study of 50-year trends in hospital admission rates for viral meningitis in childhood documented the impact of MMR on viral meningitis and an upsurge in the 2000s when MMR coverage dropped; it also documented trends in several other viral aetiologies. The study was reported directly to (formally) Public Health England for use as evidence in communication with the general public concerning the need to maintain the highest possible MMR vaccination levels to keep incidence of these diseases to a minimum. 1. Identifying ethnic disparities in maternal care and maternal health outcomes (commissioned by the Minister of State for Mental Health, Suicide Prevention and Patient Safety), work which has provided evidence in support of levelling up maternal health inequities by ethnicity; 4. Risk of pneumonia and pneumococcal disease in people hospitalized with diabetes mellitus: This study directly informs the position statements of high-profile diabetes charities in the UK. Following publication of this study, a position statement from Diabetes UK read: “All people with diabetes over the age of two years should be offered the pneumococcal vaccine.” 2. Related work on deprivation-related disparities in maternal care and maternal health outcomes; for example, a recent study revealing the extent to which homelessness is associated with an increased the risk of severe maternal morbidity, preterm birth, and low birth weight, highlighting the need for interventions that support pregnant women into stable and secure housing (in press with BJOG). 5. Associations between birthweight, gestational age at birth and subsequent type 1 diabetes in children under 12: High birthweight for gestational age and low gestational age at birth were both found to be significantly associated with subsequent type 1 diabetes. These findings demonstrated the potential role of gestational and early life environmental risk factors in the pathogenesis of type 1 diabetes, including the potential roles of insulin sensitivity and gut microbiota. According to Diabetes UK, “scientists across the world are working hard to understand just how important each suspected trigger could be, and who they could affect in the future. The findings of this study contribute to the wider scientific understanding of this disease. The research we are undertaking on ethnic and deprivation inequalities in maternal health and perinatal outcomes is in collaboration with the National Perinatal Epidemiology Unit and feeds directly into the NIHR Policy Research Unit in Maternal and Neonatal Health (https://www.npeu.ox.ac.uk/pru-mnhc), directly helping to inform government policy decisions. 6. Mortality from meningococcal disease by day of the week: English national linked database study: This study directly informed the legal debate about the 7-day NHS and the "weekend effect". 3. Proposed changes to the provision of neonatal surgery in England (in response to the GIRFT report on paediatric surgery), work which has provided evidence for reconfiguring neonatal surgery with due consideration to provider transfer times, maternity provision, and workforce skill maintenance/training; This study featured in various mainstream news outlets at the time and was referenced in the High Court judicial review case between NHS junior doctors, the British Medical Association and the Secretary of State for Health in relation to the new NHS contract for junior doctors. The study was described by Mr Justice Green as a "trenchant" piece of evidence. 4. The very high mortality rates in liver disease patients (commissioned by The Lancet), a Review which provides eight main recommendations for improving outcomes and survival rates for patients in the UK with both acute and chronic liver disease. 7. Time trends in ophthalmia neonatorum and dacryocystitis of the newborn in England: This study demonstrated to Public Health England that linked hospital data are the best available data for routinely monitoring the national incidence of newborn conjunctivitis. 8. Breast cancer mortality trends in England and the assessment of the effectiveness of mammography screening: population-based study: This study directly informed the public debate about breast cancer screening. Publication of this study was immediately followed by an NHS news release, which stated: “There is a great deal of information on both the pros and cons of screening. This study which shows no effect of mammographic screening on population-based breast cancer in England, provides additional valuable population data to inform the breast cancer screening debate.”

Unchanged: Objective for processing, Processing activities, Expected output, Expected measurable benefits.

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

The Unit of Health-Care Epidemiology (UHCE), now at the University of Oxford, [15 words unchanged] routinely collected hospital admissions data and mortality data. The Unit’s overall aims are were, and still are, to undertake epidemiological and health services research, in particular now by using routine NHS statistical data and by undertaking studies that use cohort methodologies. In order to carry out these studies, the personal data (which includes special categories of data) will [40 words unchanged] necessary for archiving purposes in the public interest, scientific or historical research purposes… purposes. The University of Oxford’s primary purpose is the advancement of education and research which are deemed to deliver a public benefit. This database and the research which uses it significantly contributes to the body of evidence and knowledge available which leads to changes in treatment, care and policies which are of benefit to the patient and the health care system; therefore the use of this data in this way is considered to be in the public interest. All research is conducted in accordance with the University’s Code of Practice for Research. The results from studies as a result of this data will benefit health and social care and therefore is considered to be in the public interest. The Unit UHCE was, historically, very closely associated with the Regional tier of the NHS [156 words unchanged] until 2005. Since the reorganisation of the NHS and the inception of (formally) Public Health England in 2013, the Unit retains formal links with the Office for Health Improvement and Disparities (OHID; one of the organisations that PHE dissolved into) through honorary research contracts. The contracts formally with PHE were replaced by contracts with OHID after PHE ceased as a legal entity. Whilst no individual-level data is ever shared with PHE, OHID, the fulfillment fulfilment of these contracts by UHCE staff (in producing summary reports and up-to-date research publications) is reliant on the receipt of up-to-date HES and mortality data from NHS Digital. [1 paragraph unchanged] The UHCE has run a continuous work programme of rolling research, notably [9 words unchanged] 1963 to the present. It is part of the University of Oxford’s Department of Public Health (now, as from 2013, the Nuffield Department of Population Health) Health and is located in the Big Data Institute at the University of [13 words unchanged] NHS Digital is unfiltered personal pseudonymised health data and is required on an annual a monthly basis. The nature and scope of the work carried out using this data, described in further detail below, requires that the data be population-level; that is, without filtering the subject pool by, e.g. geographical spread, condition cohort, etc. Due to the breadth of the data and the way it is used, there is no objective to be able to identify any person, e.g. e.g., for control or cohort groups. This pseudonymised data is required to support the following work themes: Pseudonymised data is required to support the following work projects: [5 paragraphs unchanged] UHCE use the data to analyse the distribution of hospital admission rates [95 words unchanged] such as multiple sclerosis, motor neurone disease or haemophilia. UHCE intend to update keep up to date the ‘atlases’ of disease across England. [7 paragraphs unchanged] 6. Health inequalities The research in this theme examines the sociodemographic inequalities in morbidity and mortality. This involves both individual-level (e.g., age, sex, ethnicity) and aggregate-level (e.g., regions, area-level measures of deprivation) analysis to help monitor the changes in social inequalities at population level over time. Ongoing projects include examination of trends in social inequalities in premature and preventable mortality, and disparities by age in the healthcare access for knee replacement and colorectal surgery. These projects will also help estimate the disruption caused by the Covid-19 pandemic. [18 paragraphs unchanged] In any one year, year NDPH undertakes, on average, around twenty distinct research projects utilising various different analysis techniques upon UHCE data across a broad range of the above six research themes. These range from small projects undertaken by standard users carrying out [41 words unchanged] both directly (minimised as appropriate) and indirectly (via automated user interface tools). [2 paragraphs unchanged] The University of Oxford is the Data Controller and also processes the [17 words unchanged] University, and NDPH makes all decisions regarding the data. No other organisations determine the purpose described in this Agreement or process the data for this or any other purpose.

Processing activities

Only substantive employees individuals substantively employed by the University of Oxford, or who are students of the University of Oxford Oxford, will have access to the data and only for the purposes described in this Agreement. There will be approximately ten students with access to the data. Students will only access the data under direct supervision of a substantive employee of the University of Oxford and will only process the data for the purposes described in the Agreement. The University of Oxford will only use are held responsible for the students’ access to the data under this Agreement for research which falls under the themes described above. All processing occurs at the NDPH site at the University of Oxford specified in this agreement. No record-level data is ever sent outside of NDPH. The University of Oxford will amend this agreement if the requested data is needed for research which does not fall under the themes described here. All processing occurs at the NDPH site at the University of Oxford specified in this agreement. No record-level data is ever sent outside of NDPH. [2 paragraphs unchanged] a. the Oxford Record Linkage Study, phase 1 (1963/4-1998/9) (1963-1999) (ORLS1); b. the Oxford Record Linkage Study, phase 2 (1998/9-2016/17 (1998 - 2022 – a regional subset of linked HES) (ORLS2); [1 paragraph unchanged] d. Hospital Episode Statistics (HES) for England, unlinked, 1989/90-1997/98; unlinked (1989 - 1999); e. Hospital Episode Statistics (HES) for England, linked, 1998/99-2020/21 (1999 - 2022) (provisional); f. Death registry (civil registration mortality) data. data, standalone (1995-2022) (Source: ONS). [1 paragraph unchanged] Death data has been provided directly by ONS from 1995 to 2017. 2022. NHS Digital have provided death data from 2012 onwards, linked to HES. [7 paragraphs unchanged] The University of Oxford do not hold Study ID. HES records for babies born in hospital don’t have the HES ID/Token_person_ID information for the mother, and vice versa. The encrypted identifiers are therefore required to link mothers to babies in order to create a linked mother-baby HES dataset. [1 paragraph unchanged] The datasets are all de-identified and are held securely within the UHCE, and individual-level records are only ever provided to researchers holding a an employment or study contract with UHCE. Except where specified, the datasets are not linked to one another as [83 words unchanged] – e.g. electronic tables for admission rates in the years 1968-1985, 1989-1998, 1989/90-2020/21 1989/90-2018/19 are brought together – into combined tables for the whole period 1968-2020/21. 1968-2018/19. [2 paragraphs unchanged] Raw HES and Death data is transformed into UHCE data. This involves substitution of some values, most notably encrypted identifiers; addition of UHCE record identifiers, derivation of new values, reordering and restructuring, and change of storage medium (from text file to database and analysis file) [7 paragraphs unchanged] UHCE have suites of software, developed over many years in the UHCE, [123 words unchanged] of templates and data manipulation tools then analyses the data in the four files individual datasets separately (ORLS1, ORLS2, HIPE, HES) automatically. When the results in each separate [24 words unchanged] them into overall tables and graphs giving (seemingly) continuous runs of trends. Similarly, in studies that require data for the full length of the ORLS (1963-2018/19), (1963-latest available), UHCE software packages are invoked to run the analyses within ORLS1 and, [44 words unchanged] e.g. electronic tables for admission rates in the years 1963-1998 and for 1999-2018/19 1999 - latest available – into a combined table for the whole period 1963-2018/19. 1963 - latest available. [4 paragraphs unchanged] The information systems used by the Unit of Health-Care Epidemiology are secure [33 words unchanged] granted via permission from the Data Asset Owner, and full disk encrypted. No remote access is allowed. [3 paragraphs unchanged]

Expected output

UHCE is an academic health research unit, and the outputs of the Unit’s research activities are solely for non-commercial purposes for the advancement of knowledge in public health, disease epidemiology and health service provision. Accordingly, all outputs are limited to the fulfilment of these non-commercial purposes. The scientific output produced by the research team are disseminated to researchers, policy makers and the general public. Findings are presented through publications in peer-reviewed academic, medical and public health journals, including publishing under open access licence, at conferences, and in reports to Public Health England. UHCE also aims to demonstrate its work to a wider public, beyond the scientific community, by participating in public engagement events. For example in May 2019, it is participating in an open public event organised by the Oxford Biomedical Research Centre at the John Radcliffe Hospital, Oxford. Since its work is based on routine hospital statistics, the aim is to discuss with patients and their carers the importance of electronic patient records, and how available information is used to facilitate scientific understanding of conditions and improve patient care pathways. A list of publications that arise from the study are regularly submitted to the funders, including National Institute for Heath Oxford Biomedical Research Centre and Public Health England, as annual reports. UHCE is an academic health research unit, and the outputs of the Unit’s research activities are solely for non-commercial purposes for the advancement of knowledge in public health, disease epidemiology and health service provision. Accordingly, all outputs are limited to the fulfilment of these non-commercial purposes. Team members use research outputs as examples when teaching postgraduate students and medical students at the University of Oxford. In such cases, the research outputs fully comply with the HES Analysis Guide with small numbers being suppressed appropriately and no additional processing of NHS Digital data takes place to facilitate this. Team members also teach a workshop on using HES data for research, which is open to academics and research staff within the University of Oxford. This involves a made-up example of NHS Digital data but no actual NHS Digital data are used in these workshops. The scientific output produced by the research team are disseminated to researchers, policy makers and the general public. Findings are presented through publications in peer-reviewed academic, medical and public health journals, including publishing under open access licence, at conferences, and in reports to Public Health England. UHCE also aims to demonstrate its work to a wider public, beyond the scientific community, by regularly participating in public engagement events. A list of publications that arise from the study are regularly submitted to the funders, including National Institute for Heath Oxford Biomedical Research Centre and Public Health England, as annual reports. Team members use research outputs as examples when teaching postgraduate students and medical students at the University of Oxford. They also teach a series of workshops on using HES data for research, which is open to academics and research staff within the University of Oxford. [3 paragraphs unchanged] Below is a detailed list of UHCE’s expected publications and other research activities. An examples of a recent outputs is as follows: 1. Study of secular trends in, and perinatal factors associated with, congenital viral infections in England over 50 years (CMV, HSV, VZV and rubella), including the use of linked HES (and HES Maternity). For presentation at the European Society for Paediatric Infectious Diseases and for publication in a peer-reviewed journal such as Lancet Infectious Diseases in late 2019/early 2020. Study on the determinants of the decline in mortality from acute stroke in England: linked national database study of 795,869 adults. After publication in May 2019, the study findings attracted considerable media attention from the national news agencies, and were widely reported on television, radio and in the newspapers, creating positive publicity for the NHS and firmly establishing the rise in young stroke incidence as a national public health issue. The study made front page news in leading national newspapers “The Times” and “The Guardian”. 2. Europeristat / Euronet-SAMM project: feasibility study of incidence of severe acute maternal morbidity (SAMM) in England using linked HES (including HES Maternity). International comparison with other routinely-collected European data sets in collaboration with academic partners across Europe. For publication in a peer-reviewed journal such as BJOG in late 2019. Below is a detailed list of UHCE’s expected publications and other research activities which will be generated using NHS Digital data (all studies listed below have been added since the last independent review of this Data Sharing Agreement in 2019, except projects 10 and 11). 3. Study of the burden of heart failure in pregnant women and survival rates, and secular changes from 1999 to 2016, using linked HES and mortality data. Collaboration with the National Perinatal Epidemiology Unit. For publication in the third quarter of 2020. 1. A study investigating the impact of the COVID-19 pandemic on the management and outcome of colorectal cancer in the English NHS. In partnership with NHS Digital, the University of Oxford's group has already published one paper (in Lancet Gastroenterology and Hepatology) on this topic in January 2021 using SUS+ data and is working to ensure the results are updated each month. The next phase of the work is to investigate whether the changes in care are linked to outcome so the University of Oxford plan to apply the same and adapted algorithms to the linked HES-mortality file. This is expected to enable the study team, for example, to determine whether the observed increase in stoma formation has been reversed or whether there have been changes in surgical mortality. It is anticipated this work will be published in a similar journal in the late 2022/ early 2023. 4. Study of equitability of access to common surgical treatments by individuals with psychiatric disease in England using HES. For publication in a peer-reviewed community health journal such as Journal of Epidemiology and Community Health in late 2019/early 2020. 2. Study of rates of COVID-19 related mortality in patients with underlying comorbidities compared with the general population in England. Ongoing outputs have been used in the national COVID-19 response and have now been extended to a collaboration with the RECOVERY trial. Publication in a peer-reviewed journal is expected in in mid-late 2022. 5. Study of trends and geographical variations in 30-day case fatality rates after acute stroke using linked HES and mortality data. For publication in a peer-reviewed journal such as Stroke in the second quarter of 2020. 3. Study investigating trends in major operations for extensive/metastatic pelvic cancers. Algorithm for identifying these operations being developed in partnership with NHS England and peer-reviewed publication to be submitted to a journal such as British Journal of Surgery in late 2022. 6. Analysis of English data, as a part of an international collaborative project on the determinants of the reduction in mortality rates from myocardial infarction and stroke in England, Australia, New Zealand and Canada. International comparison with other routinely-collected data sets in collaboration with academic partners in Australia, New Zealand and Canada. This project aims to utilise the analytical methods developed by the WHO MONICA study - a large cohort study - on routine hospital data. The paper on mortality from myocardial infarction is expected to be completed in late 2019 and published in a peer-reviewed journal such as Circulation in late 2019. The paper on mortality from stroke is expected to be published in late 2020/first quarter 2021. 4. A study investigating the use of colonoscopy and flexible sigmoidoscopy across the English NHS to determine both cancer pick-up rates and complications following the procedures (for example, bowel perforations and bleeding). It is estimated this will be ready for publication in a journal such as Gut in the late 2022 / early 2023. 7. Study of long-term trends in hospital admission for stroke and stroke mortality in England. For publication in a peer-reviewed journal such as Stroke in the second quarter of 2020. 5. An ongoing series of international comparison studies using English data, as a part of a collaborative project with academic partners in Australia, New Zealand and Canada. In the first instance, it aims to compare the determinants of the reduction in mortality rates from myocardial infarction (published) and stroke (current project) in England, Australia, New Zealand and Canada. It uses routinely-collected datasets in all 4 jurisdictions and the analytical methods developed by the WHO MONICA study - a large cohort study - on routine hospital data. The study on myocardial infarction was published in the Lancet Public Health in early March 2022. The current one is a comparison of the determinants of the reduction of stroke mortality and is expected to begin in mid-2022 and to be published in a peer-reviewed journal in 2023. 8. Analysis 6. Study of hospitalisation trends for childhood onset schizophrenia the risk of circulatory conditions in collaboration patients with the Department of Psychiatry, Warneford Hospital. Marfan Syndrome. For publication in a peer-reviewed journal such as British Journal of Psychiatry in the first quarter of 2020. late-2022. 9. Study of disease associations between hypogonadism (low testosterone), acromegaly (IGF1) and prostate cancer using linked HES. For publication in a peer-reviewed cancer journal such as BJC in late 2019. 7. Monitoring socioeconomic inequalities in premature and preventable mortality. For publication in peer-reviewed journal such as Science in late-2022. 10. Study of long-term trends in hospital admission rates and mortality rates of aortic aneurysm and aortic dissection. Expected outputs to include presentation at the European Society of Cardiology Congress and publication in a peer-reviewed journal such as Circulation in late 2019/early 2020. 8. Study of distribution of neonatal surgery in England in response to the Getting It Right First Time (GIRFT) report for Paediatric Surgery, to be published in a peer-reviewed journal in late-2022. 11. Study of the association between diabetes mellitus and aortic aneurysm rupture in people from different ethnic groups. For publication in a peer-reviewed journal such as Diabetologia in late 2019/first quarter 2021. 9. Study of ethnic disparities in maternal health, characterised by deprivation and geography. Collaboration with the National Perinatal Epidemiology Unit, reporting to the Minister of State for Mental Health, Suicide Prevention and Patient Safety. For publication in late-2022. 12. Teaching a series of workshops for the Oxford University public health / epidemiology students and staff on maximising the use of linked HES and mortality data in epidemiological and health services research. Three workshops are planned for 2018-2020, the first of which was held in October 2018 in the Big Data Institute. 10. Study of long-term trends in hospital admission for stroke and stroke mortality in England. For publication in a peer-reviewed journal such as Stroke in 2022/23. 11. Study of trends and geographical variations in 30-day case fatality rates after acute stroke using linked HES and mortality data. For publication in a peer-reviewed journal such as Stroke in 2022/23. 12. NIHR-funded study of timing of stoma closure in neonates. For publication in a paediatric journal such as Archives of Disease in Childhood in 2022/23. 13. Study of the association between eating disorders, such as anorexia and bulimia, with the risk of cancers. This work is under review with the British Journal of Psychiatry. 14. Several other studies of disease associations employing similar methodology to the above two studies: cancer outcomes in patients with cystic fibrosis; cardiovascular disease outcomes in patients with systemic lupus erythematosus; aortic dissection and aneurysm in patients with giant cell arteritis; cerebral venous thrombosis in pregnancy; bleeds and venous thromboembolism after hip fracture. For publication in peer-reviewed journals in 2022/23. 15. Study of regional long-term trends of hospital admissions for myocardial infarction in England. The University of Oxford's group has recently published on very long-term trends in hospital admissions for myocardial infarction in England as a whole and by sex and age (Journal of Epidemiology & Community Health, July 2021). This study intends to examine rates in different regions of the country, the results of which are expected to be published in a peer-reviewed journal such as Heart in 2023. 16. Study of long-term trends in the incidence and case-fatality of myocardial infarction by ethnicity using linked HES and mortality records. This work is expected to be published in a peer reviewed journal in 2024. 17. A study on the long-term trend in hospital admissions for primary care sensitive conditions for children aged under 15 years. This study is expected to be published in a peer-reviewed journal such as Archives of Disease in Childhood. 18. Postgraduate teaching MSc in Global Health Sciences and Epidemiology course in the University of Oxford, using published examples from the linked HES. This is an ongoing annual commitment. 19. The study team work closely with the Medical School at the University of Oxford and offer supervised research placements for undergraduate medical students. This is an ongoing commitment.

Expected measurable benefits

With reference to each of the individual studies listed above in section 5c: 5c, data received from NHS Digital is expected to provide the following benefits: 1. Defining congenital infection in the absence of routine surveillance is highly challenging. Data regarding laboratory or clinically confirmed cases of congenital infection are not routinely collected and so current estimates of diseases are unknown. HES data provide a unique and current opportunity to define the current incidence of important congenital infections such as CMV. A study of secular trends in congenital viral infections in England over 50 years (CMV, HSV, VZV and rubella), will put the current hospitalisation rates of these diseases into a 50-year historical context as a matter of public record. Previous studies have shown that changing population demographics, in particular increases in ethnic diversity, lead to higher rates of congenital CMV. By defining the incidence of congenital infections such as CMV, and the perinatal/demographic factors associated with its incidence, it is expected that publication and dissemination of this study (to be presented at the European Society for Paediatric Infectious Diseases and published in a peer-reviewed journal such as Lancet Infectious Diseases in late 2019) will immediately inform current clinical care (through promoting awareness amongst midwifes to impart appropriate guidance to pregnant women), NHS policy (inform calculations of the cost of congenital CMV to the NHS, so policy makers can weigh this against the benefits of introducing routine testing for all babies), promote awareness (raise awareness of congenital CMV among healthcare workers, aiming to reduce missed opportunities for diagnosis and treatment), and inform future research (inform future treatment strategies including vaccine trials). This study is also a response to Public Health England’s Maternity Transformation Programme to improve neonatal outcomes (PHE Remit Letter, 2018-2019). 1. The published study has determined there is presently a diagnostic deficit of over 4,000 colorectal cancers as a result of the COVID-19 pandemic. In addition, for those people who have been diagnosed, standards of care have changed dramatically. More people are living with stomas than would be expected and many more have had delays to their surgical procedures or alternatives such as radiotherapy. The impact of these changes on outcome is not yet understood. This study will aim to investigate how NHS colorectal cancer services have changed during the pandemic and provide higher resolution information to inform the recovery. 2. The feasibility study of incidence of severe acute maternal morbidity (SAMM) in England using linked HES (including HES Maternity) will allow international comparison with other routinely-collected data sets in collaboration with academic partners across Europe. To date, there is no SAMM registry or specific monitoring data in Europe. Hospital discharge databases offer the opportunity for such monitoring. Some countries already use hospital discharge databases to study SAMM or some of its components, but each country uses its own criteria and codes, preventing rigorous comparisons. The benefit of this study (to be published in a peer-reviewed journal such as BJOG in late 2019) is to further the CROWN Initiative, that maternal health surveillance across countries should be based on a set of common SAMM indicators assessed in comparable databases. The study is also a response to Public Health England’s Maternity Transformation Programme to improve maternal outcomes (PHE Remit Letter, 2018-2019). 2. This study has directly informed the roll-out of the COVID-19 vaccination priority strategies (current and future) by determining which clinically vulnerable populations will benefit most from early vaccination. Findings were reported directly to the Joint Committee on Vaccinations and Immunisation. 3. Better understanding of the health outcomes in women who developed heart failure during or after pregnancy could improve maternal health and reduce mortality. This study is also a response to Public Health England’s Maternity Transformation Programme to improve maternal outcomes (PHE Remit Letter, 2018-2019). Findings will be reported to PHE and published in a peer-reviewed journal such as BJOG in 2020. 3. The surgical management of metastatic cancers involves major operations, such as pelvic exenterations (the removal of all pelvic organs), that have significant resource implications, are associated with significant morbidity and mortality outcomes for which are poorly understood. Given the scarcity of information on the use of these operations specialist cancer commissioning within NHS England are seeking further information about their use. University of Oxford’s proposed study will aim to provide the first population-based information on these operations in England and so inform the commissioning of cancer services. 4. The study of common surgical treatments in individuals with psychiatric disease (for publication in the Journal of Epidemiology and Community Health in late 2019 and dissemination at the following Public Health England annual conference) is expected to identify potential issues of inequitability of access to healthcare among individuals with mental illness. Addressing health inequalities and improving the lives of people living with mental illness are two of Public Health England’s top priorities currently (PHE Remit Letter, 2018-2019). 4. Colonoscopy is the main diagnostic test for colorectal cancer and is an extremely commonly used procedure in the NHS. Prior to COVID-19, colonoscopy services were under pressure due to inadequate capacity to keep up with demand but, as the procedures were initially deemed aerosol generating, the current situation has become even more pressured. Evidence is urgently needed, therefore, to increase the quality and efficiency of colonoscopic services. This study hopes to provide this information. It is expected to both enable the pick-up rates of cancer to be determined and help quantify the number of adverse outcomes linked to the procedures. At present, such information is not available for the NHS, which is preventing optimisation of services. 5. The analysis of geographical differences in 30-day case fatality rates after acute stroke in England would measure the variations in short-term fatality in different parts of the country, and the findings will help to identify areas with particularly high or low case-fatality rates. This information, which will be reported to Public Health England and published in a peer-reviewed journal such as Stroke in early 2020, is intended to be used by the local health authorities in regions with high rates to improve the provision of care of stroke patients. 5. The collaborative project on heart attack and stroke is establishing an international network of researchers working on routine hospital statistics in England, Australia, New Zealand and Canada to investigate cardiovascular disease incidence and mortality rates. This facilitated comparison of routine hospital statistics in different countries provide specific measures of the factors affecting trends in mortality rates in these countries. The recently published study reported how much of the reduction in mortality rates from heart attack was due to hospital care versus preventive medicine, and how the contribution of the two differ between the countries. This will hopefully inform and enable better health care planning and resource allocation for the treatment and prevention of heart attack in each country. The success of this project has demonstrated that the methods developed by large and inexpensive studies, such as the WHO MONICA study, can be used by those working with routine hospital statistics. The current study intends to investigate stroke mortality using the same methodology and will hopefully make similar contributions to health care planning and resource allocation for stroke. 6. The collaborative project on myocardial infarction and stroke will help to establish an international network of researchers working on routine hospital statistics in England, Australia, New Zealand and Canada to investigate cardiovascular disease incidence and mortality rates. This will facilitate comparison of routine hospital statistics in different countries in relation to these diseases, providing specific measures of the factors that affected trends in mortality rates in these countries. The study will measure how much of the reduction in mortality rates from myocardial infarction was due to hospital care versus preventive medicine, and how the contribution of the two differ between the countries, to enable better health care planning and resource allocation for the treatment and prevention of myocardial infarction. Study findings are expected to be published and available to the public in late 2019. The success of this project will demonstrate that the methods developed by large and expensive studies, such as the WHO MONICA study, could be used by others working with routine hospital statistics. 6. Marfan Syndrome is a hereditary condition which affects connective tissues. Among most severe complications of Marfan are aortic aneurysms (an enlargement of the aorta). This study is expected to quantify the risk of a range of vascular complications in these patients and report any changes in risks that occurred in recent years compared to the past. The hypothesis is that as the care and prognosis of these individuals have improved, the risk of developing a ruptured aneurysm and other severe and life-threatening vascular conditions declined. These findings are expected to be published in peer-reviewed journal in 2022. 7. The study of long-term trends in hospital admission rates and mortality from stroke measures the burden of stroke in the population. The pattern of hospitalisation trends has important implications for planning of health-care policies and allocation of resources. Monitoring of recent trends in stroke mortality offers an opportunity to record and report any slowdown or plateauing in trends. Analysis of trends in population subgroups, by age, ethnicity or socio-economic status, will help to identify population at risk of developing stroke or dying from stroke, and this information could be used for targeted interventions. 7. This study is part of a larger international research project to compare social inequalities in premature and preventable deaths over time, within and across countries. Previous research largely reported the social inequality based on individual-level factors such as education, occupation. This study intends to generate aggregate-level evidence in each of the participating countries to help monitor the social inequalities in premature and preventable mortality at whole-population level. For English data, the University of Oxford will report these by levels of the indices of multiple deprivation. Given there is a substantial difference in the life expectancy of people with regard to their socioeconomic status, it is hoped that this work will provide valuable evidence to help inform interventions to eliminate these inequalities. 8. The study of temporal trends in schizophrenia in children could help to better understand the disease as well as factors influencing the provision of mental health care for children. 8. The Getting it Right First Time report for Paediatric Surgery has recently been published: https://www.gettingitrightfirsttime.co.uk/surgical-specialty/paediatric-surgery/. It makes striking recommendations for the reorganisation of services, including potentially halving the number of surgical centres. University of Oxford intend to perform a detailed analysis of current hospital service provision and combine this with evidence from other publications and family perspectives to discuss issues around how to optimise quality whilst maintaining good access to services for families, which it is hoped will directly inform the reorganisation of neonatal surgical services. 9. Experimental, clinical, and epidemiological evidence suggests that testosterone and insulin-like growth factor-I (IGF-I) are associated with prostate cancer risk. HES data offer the opportunity to study associations between sex-hormone-related disease and prostate cancer. Acromegaly is associated with clinically high IGF-I concentrations, and Klinefelter’s syndrome, testicular hypofunction and hypopituitarism are associated with clinically low circulating testosterone concentrations. This study, to be published in a peer-reviewed journal such as the British Journal of Cancer in late 2019, is expected to provide valuable insight into prostate cancer aetiology. 9. Following publication of a recent MBRRACE report by the National Perinatal Epidemiology Unit (https://www.bmj.com/content/372/bmj.n152), there is growing interest from within central government (https://www.parallelparliament.co.uk/mp/nadine-dorries/debate/Commons/2021-04-19/debates/6935B9C7-6419-4E7B-A813-E852A4EE4F5C/BlackMaternalHealthcareAndMortality) and among the general public (Channel 4’s Dispatches: The Black Maternity Scandal) in uncovering ethnic disparities in maternal health. In collaboration with the National Perinatal Epidemiology Unit, these summary findings were reported to the Minister of State for Mental Health, Suicide Prevention and Patient Safety in late 2021 and will be published in a peer-reviewed journal in late 2022. It is hoped that any disparities uncovered by the study will be used to focus efforts to level up maternity outcomes for women from ethnic backgrounds where inequities are found. 10. The study of mortality trends in aortic aneurysm will show whether the introduction of the national screening programme in 2013 has had measurable effects upon mortality rates in the population. This study, which will be reported to Public Health England and is expected to be presented at the European Society of Cardiology Congress and published in a peer-reviewed journal, will help to evaluate the investment of resources in the national screening programme for aortic aneurysm. 10. The study of long-term trends in hospital admission rates and mortality from stroke measures the instance of stroke in the population. The pattern of hospitalisation trends has important implications for planning of health-care policies and allocation of resources. Monitoring of recent trends in stroke mortality offers an opportunity to record and report any slowdown or plateauing in trends. Analysis of trends in population subgroups, by age, ethnicity or socio-economic status, will help to identify population at risk of developing stroke or dying from stroke, and it is hoped this information could be used for targeted interventions. 11. Diabetes mellitus (DM) is a known risk factor for the majority of cardiovascular disease, however literature suggest that aortic aneurysm may be relatively lower in people with DM than in those without DM. HES data offer the opportunity to investigate this potential inverse association in the English population. As explained elsewhere, the Unit has the ability to run these disease association analyses very swiftly, and the (separate) ORLS dataset can be used to complement such findings to determine how these disease associations may have evolved over time. A relatively low risk of aortic aneurysm in people with DM might suggest that there are unknown genetic, physiological or treatment mechanisms involved in the relationship between diabetes and aortic aneurysm, a finding that would have immediate implications for further research and potentially long-term implications for the understanding of aortic aneurysm pathology. 11. The analysis of geographical differences in 30-day case fatality rates after acute stroke in England would measure the variations in short-term fatality in different parts of the country, and the findings will help to identify areas with particularly high or low case-fatality rates. This information, which will be reported to Public Health England and is intended to be used by the local health authorities in regions with high rates to improve the provision of care of stroke patients. 12. A series of workshops for PhD students of epidemiology and public health research staff at the University of Oxford will help to disseminate the importance of routine health data, and how the data collected by the NHS can be used for epidemiological and health services research. This could potentially reduce the costs of clinical trials and cohort studies, if the recruiting and follow-up of patients could be achieved using routine hospital statistics. 12. Optimal timing for neonatal stoma closure remains unclear. In this study University of Oxford will aim to establish current practice on timing of stoma closure using HES, as part of a multidisciplinary NIHR-funded research programme to determine feasibility of a clinical trial comparing ‘early’ and ‘late’ stoma closure, which is hoped will directly inform clinical practice in this area. More generally, obtaining this new release (and, indeed, all future new releases – i.e. keeping ‘up-to-date’) of linked HES data will enable the research team to produce the most up-to-date analyses of trends and variation in hospital admission rates, incidence and mortality. Without the data collected for the most recent years, the research output is less informative and outdated; this will limit the impact of UHCE work in the scientific community and among policy makers. UHCE publishes work in international journals and presents at international conferences, and this contributes to building a strong reputation for the UK as a country with established system of collecting and analysing electronic medical records. Being a part of the University of Oxford, UHCE research output is taken into consideration when ranking academic institutions; with the new data, the Unit will be able to produce results and publish papers for recent years, not only historical studies, and this will help to make a substantial contribution to maintaining Oxford’s reputation for scientific excellence. 13. In collaboration with the Department of Psychiatry at the University of Oxford, the study team conduct research on mental health problems. The study team are working closely with the psychiatrist to do studies with high clinical value that will have an impact on clinical practice. Improving mental health is high on the national agenda, and HES data can be used as a resource for important work. In particular, the study team work on mental health conditions that affect children and young people. The focus of their work for the next year are eating disorders. Eating disorders are increasing in the population, and most commonly affect young people. However, the long-term effects of these conditions are unknown. It is possible that changes in metabolism, hormones, or irritation of the upper gastrointestinal tracts from excessive vomiting, might result in increased risk of cancer. However low body mass index might be protective against some cancers. The literature on this possible association is limited. It is hoped these findings could be used to inform patients suffering from eating disorders and doctors looking after these patients, as well as used for identifying populations at particularly high risk of cancers that might benefit from earlier screening. 14. Characterising associations between diseases can help to uncover shared aetiologies or an increased need for clinical surveillance. As explained elsewhere, the Unit has the ability to run these disease association analyses very swiftly, and the (separate) Oxford Region dataset can be used to complement such findings to determine how these disease associations may have evolved over time. Owing to strong and longstanding links with clinical teams within the Oxford University Hospital NHS Foundation Trust and beyond, University of Oxford are able to target specific disease association studies, such as these, which are of particular and current interest to the clinical community. 15. The recently published very long-term study found that after a steady decline in admissions for heart attack from the 1990s, there was an overall modest increase in admission rates between 2012 and 2016. Analysis of regional variations in hospital admissions for heart attack in England will assess variations in hospital admission rates and the extent of this rise in rates in different parts of the country. This work will report trends beyond 2016, which is important to assess whether the rates continued to rise or not. 16. The study of long-term trends in the incidence and case-fatality from heart attack intends to measure changes in the occurrence of myocardial infarction in the population. The study team will look at trends in rates by sex, age, socioeconomic status, ethnic group and geographical areas. The findings are expected to help inform coronary disease prevention strategies and public awareness campaigns to help focus on those most at risk and has implications for health-care resource planning for those areas of the country with the largest burden of coronary disease. 17. The study of long-term trends in hospital admissions for conditions expected to be treated by general practitioners in children aged under 15 years aims to measure the number and rate of hospital admissions for these conditions. The study team will look at trends by sex, age and area socioeconomic status. This is expected to help inform the delivery and structure of acute paediatric care in both primary and secondary care. 18. Members of the UHCE team teach the bioinformatics and epidemiology modules on MSc in Global Health Sciences and Epidemiology course run by the University of Oxford, using examples of studies using HES. The postgraduate students are the next generation of academics, public health professionals and policy makers. It is important for them to learn about the value of electronic medical data available in England. The course highlights the importance of routine health data for research, as it teaches about multiple ways of using the data. The routine electronic medical records could be used on their own, or linked to cohort studies and clinical trials, which could potentially reduce the costs of running these studies. 19. The training provided by the study team for medical students will hopefully ensure that future doctors in England have the skills and experience required for working with and understanding routine electronic health records. Students will work under the supervision of experienced staff using existing infrastructure. Through teaching future doctors, the study team aim to make a valuable contribution to the digital transformation of care in the NHS and delivering the commitments within the NHS Long Term Plan. More generally, obtaining this new release (and, indeed, all future new releases – i.e. keeping ‘up-to-date’) of linked HES data and mortality data for all England is intended to enable the research team to produce the most up-to-date analyses of trends and variation in population-based hospital admission rates, incidence and mortality. Without the data collected for the most recent years and months, the research output is less informative and outdated, especially during the COVID-19 pandemic; this will limit the impact of UHCE work in the scientific community and among policy makers. UHCE publishes work in international journals and presents at international conferences, and this contributes to building a strong reputation for the UK as a country with established systems of collecting and analysing electronic medical records. Being a part of the University of Oxford, UHCE research output is taken into consideration when ranking academic institutions; with the new data, the Unit will be able to produce results and publish papers for recent years, not only historical studies, and this will help to make a substantial contribution to maintaining Oxford’s reputation for scientific excellence.

Benefits reported

The Unit was, historically, very closely associated with the Regional tier of the NHS and with the Department of Health. Although the UHCE is part of Oxford University, it was based on the Oxford RHA’s site from 1963 until the reorganisation of Regional Health Authorities in the mid-1990s. In particular, the UHCE worked closely with the Oxford RHA on medical statistics, record linkage (including the Oxford Record Linkage Study (ORLS)), and health services research. From 1998-2005, the Unit had strong service links with the Department of Health’s National Centre for Health Outcomes Development (NCHOD) - the former Unit Director directed the work programme of the Oxford site of NCHOD. As part of the NCHOD work, the DH commissioned the UHCE to construct and analyse English national record-liked HES files, with HES-to-HES linkage and HES-to-mortality linkage, along the lines of the Oxford Record Linkage Study. The former Unit Director was also Co-founder and Scientific Director of the South East England Public Health Observatory from its inception in 2000 until 2005. Since the reorganisation of the NHS and the inception of Public Health England in 2013, the Unit retains formal links with PHE through honorary research contracts and the fulfilment of these contracts is reliant on the receipt of up-to-date HES and mortality data from NHS Digital. This database and the research which uses it significantly contributes to the body of evidence and knowledge available which leads to changes in treatment, care and policies which are of benefit to the patient and the health care system. With 50 years’ worth of hospital data, research into areas such as admission trends, case fatality rates, disease associations and maternal disease links, that could take years to complete, can be achieved extremely efficiently. This means UHCE can react extremely swiftly to matters that may suddenly come into the spotlight of public health or to new issues that may arise for the health care system or for specific specialties. Combining historical data with the most contemporary data available allows us to characterise the ‘current’ picture but with reference to historical context and with access to longitudinal patient history when defining patient populations or identifying comorbidities. The UHCE has run a continuous work programme of rolling research, notably using hospital statistics, mortality data, and record linkage, from 1963 to the present. It is part of the University of Oxford’s Department of Public Health (now, as from 2013, the Nuffield Department of Population Health) and is located in the Big Data Institute at the University of Oxford. In previous years, the University of Oxford have been able to react quickly, for example, to questions over the ‘week-end effect’, where UHCE was able to report rapidly on mortality rates for meningococcal meningitis (there are few other diseases which can act as such a good marker to show the difference between expert treatment or no or suboptimal treatment; the study showed no evidence of an adverse day of the week effect). With 50 years’ worth of hospital data, research into areas such as hospital trends, mortality rates, disease history and maternal disease links, that could take years to complete, can be achieved extremely efficiently. This means UHCE can react extremely swiftly to issues arising for the health care system or from patients. For example, in the questions over the ‘week-end effect’, UHCE was able to publish on mortality rates for meningococcal meningitis. There are few other diseases which can act as such a good marker to show the difference between expert treatment or no or suboptimal treatment. The study showed no evidence of an adverse day of the week effect. In more recent years/months, the University of Oxford have been able to react swiftly to the impact of the COVID-19 pandemic on cause-specific hospital admissions and mortality. This database and the research which uses it significantly contributes to the body of evidence and knowledge available which leads to changes in treatment, care and policies which are of benefit to the patient and the health care system. 1. The work on Covid-19 mortality in patients who are clinically vulnerable (study 2 in 5c, 5dii) was reported to the Joint Committee on Vaccinations and Immunisation and directly helped inform the early roll-out of Covid-19 vaccines to clinically vulnerable patients, most notably those on dialysis; importantly, this study relied equally on both the contemporary data (needed for identifying COVID-19 mortality outcomes) and the historical data (needed for identifying clinically vulnerable patients according to their prior hospitalisation history, such as patients who may have undergone organ transplant surgery many years ago). This study, through its reporting to the JCVI and the UK Kidney Association, directly helped inform the early roll-out of Covid-19 vaccines to clinically vulnerable patients, most notably those on dialysis. By way of further illustration, the following provides a small sample of publications and the impact they have had. 2. The University of Oxford's report on the enormous and sustained decreases in hospital admissions for childhood infections since the onset of the pandemic has highlighted an important public health phenomenon to the medical profession and health policy makers and provides evidence for implementing non-pharmacological interventions that could be sustained beyond the pandemic, especially during winter months, to minimise the burden on health systems and protect vulnerable children. The University of Oxford continues to provide ongoing updates using the most contemporary data to continually monitor the impact of any changes to natural immunity in the paediatric population that may have occurred during two years of societal and behavioural restrictions as well as the effects of any disruptions to the routine vaccine schedule. 1. Seminog OO, Scarborough P, Wright FL, Rayner M, Goldacre MJ. Determinants of the decline in mortality from acute stroke in England: linked national database study of 795 869 adults. BMJ 2019;365:l1778. 3. The University of Oxford have also defined and highlighted through publication the massive under-diagnosis of colorectal cancer during the pandemic and the increasing backlog of patients with colorectal cancer who require surgery during the pandemic. This work has directly quantified the scale of the backlog caused by the COVID-19 pandemic (i.e., number of individuals who would be expected to have been diagnosed with colorectal cancer but were not). This information has been used by the Department of Health and Social Care, clinical specialist bodies (such as the Association of Coloproctologists of Great Britain and Ireland) and hospital Trusts directly to take steps to reduce this backlog. This has positive benefits for those with the illness by helping to minimise delays in diagnosis (which frequently lead to poor outcomes). Key point: While reporting reductions of 50% in stroke mortality and 20% in stroke incidence overall, this study highlighted both a contrasting increase in stroke incidence in younger people and the significant role of hospitals in the reduction of stroke death overall, and the study’s widespread coverage significantly increased public awareness of these points. Aside from the studies undertaken during the pandemic (described above), other recent work has included reacting to other issues in the political or public health spotlight, such as ethnic disparities in maternal care and maternal health outcomes (commissioned by the Minister of State for Mental Health, Suicide Prevention and Patient Safety), work which has provided evidence in support of levelling up maternal health inequities by ethnicity; proposed changes to the provision of neonatal surgery in England (in response to the GIRFT report on paediatric surgery), work which has provided evidence for reconfiguring neonatal surgery with due consideration to provider transfer times, maternity provision, and workforce skill maintenance/training; and the very high mortality rates in liver disease patients (commissioned by The Lancet), a Review which provides eight main recommendations for improving outcomes and survival rates for patients in the UK with both acute and chronic liver disease. Detail: After publication in May 2019 the study findings attracted considerable media attention from the national news agencies, and were widely reported on television, radio and in the newspapers, creating positive publicity for the NHS and firmly establishing the rise in young stroke incidence as a national public health issue. The study made front page news in leading national newspapers “The Times” and “The Guardian”. Further examples of prior work, in order to illustrate the kind of studies that can be undertaken rapidly in the future and their impact, are as follows: 2. Seagroatt V, Goldacre MJ. Crohn's disease, ulcerative colitis, and measles vaccine in an English population, 1979-1998. J Epidemiol Community Health. 2003 Nov;57(11):883-7 1. Congenital viral infections in England over five decades: CMV diagnoses have increased dramatically since 2007 when the NHS Newborn Hearing Screening Programme was introduced, providing evidence for the need to identify cases as soon as possible so that antiviral drugs can be administered early enough to prevent life-long hearing loss and neurodevelopmental disability. Reaction to this study later prompted calls (published in The Lancet) for policy makers to begin testing for congenital CMV in those who do not pass their newborn baby hearing screen, which, if implemented, could provide immediate benefits to affected infants by improving developmental outcomes. Key point: This study directly influenced official NHS public engagement policy in relation to the measles vaccine. 2. Determinants of the decline in mortality from acute stroke in England: While reporting reductions of 50% in stroke mortality and 20% in stroke incidence overall, this study highlighted the significant role of acute hospital care in the reduction of stroke death overall, but also a worrying increase in stroke incidence in people younger than 55 years. This study prompted a number of large national news outlets to put acute stroke in the spotlight and created a space for public discussion. The study highlighted to policy makers, health professionals and the general public that guidelines for the primary prevention of stroke and public awareness campaigns should include the message that stroke is not just a condition of the elderly. Detail: Publication of this study was followed immediately by a NHS Immunisation Information press release (11 Dec 2003), which stated: “A new study has confirmed that the introduction of measles vaccine in this country played no part in causing Crohn’s Disease and Ulcerative Colitis. The theory that measles vaccine was linked to bowel disease and then autism depended on a belief that measles virus damaged the bowel. This study adds to the available evidence that says that this is not the case.” 3. Hospital admissions for viral meningitis in children in England over five decades: This study of 50-year trends in hospital admission rates for viral meningitis in childhood documented the impact of MMR on viral meningitis and an upsurge in the 2000s when MMR coverage dropped; it also documented trends in several other viral aetiologies. The study was reported directly to (formally) Public Health England for use as evidence in communication with the general public concerning the need to maintain the highest possible MMR vaccination levels to keep incidence of these diseases to a minimum. 3. Martin NG, Iro MA, Sadarangani M, Goldacre R, Pollard AJ, Goldacre MJ. Hospital admissions for viral meningitis in children in England over five decades: a population-based observational study. Lancet Infect Dis. 2016;16:1279-87 4. Risk of pneumonia and pneumococcal disease in people hospitalized with diabetes mellitus: This study directly informs the position statements of high-profile diabetes charities in the UK. Following publication of this study, a position statement from Diabetes UK read: “All people with diabetes over the age of two years should be offered the pneumococcal vaccine.” Key point: The study was reported directly to Public Health England for use as evidence in communication with the general public concerning the need to maintain the highest possible MMR vaccination levels. 5. Associations between birthweight, gestational age at birth and subsequent type 1 diabetes in children under 12: High birthweight for gestational age and low gestational age at birth were both found to be significantly associated with subsequent type 1 diabetes. These findings demonstrated the potential role of gestational and early life environmental risk factors in the pathogenesis of type 1 diabetes, including the potential roles of insulin sensitivity and gut microbiota. According to Diabetes UK, “scientists across the world are working hard to understand just how important each suspected trigger could be, and who they could affect in the future. The findings of this study contribute to the wider scientific understanding of this disease. Detail: This study of 50-year trends in hospital admission rates for viral meningitis in childhood documented the impact of MMR on viral meningitis and an upsurge in the 2000s when MMR coverage dropped following the Andrew Wakefield scandal; it also documented trends in several other viral aetiologies. 6. Mortality from meningococcal disease by day of the week: English national linked database study: This study directly informed the legal debate about the 7-day NHS and the "weekend effect". 4. Seminog OO, Goldacre MJ. Risk of pneumonia and pneumococcal disease in people hospitalized with diabetes mellitus: English record-linkage studies. Diabet Med. 2013 Dec;30(12):1412-9. doi: 10.1111/dme.12260. Epub 2013 Jul 24 This study featured in various mainstream news outlets at the time and was referenced in the High Court judicial review case between NHS junior doctors, the British Medical Association and the Secretary of State for Health in relation to the new NHS contract for junior doctors. The study was described by Mr Justice Green as a "trenchant" piece of evidence. Key point: This study directly informs the position statements of high-profile diabetes charities in the UK. 7. Time trends in ophthalmia neonatorum and dacryocystitis of the newborn in England: This study demonstrated to Public Health England that linked hospital data are the best available data for routinely monitoring the national incidence of newborn conjunctivitis. Detail: Following publication of this study, position statement from Diabetes UK: “All people with diabetes over the age of two years should be offered the pneumococcal vaccine.” 8. Breast cancer mortality trends in England and the assessment of the effectiveness of mammography screening: population-based study: This study directly informed the public debate about breast cancer screening. Publication of this study was immediately followed by an NHS news release, which stated: “There is a great deal of information on both the pros and cons of screening. This study which shows no effect of mammographic screening on population-based breast cancer in England, provides additional valuable population data to inform the breast cancer screening debate.” 5. Goldacre RR. Associations between birthweight, gestational age at birth and subsequent type 1 diabetes in children under 12: a retrospective cohort study in England, 1998-2012. Diabetologia. 2018;61(3):616-625 Key point: High birthweight for gestational age and low gestational age at birth were both found to be significantly associated with subsequent type 1 diabetes. Detail: These findings demonstrated the potential role of gestational and early life environmental risk factors in the pathogenesis of type 1 diabetes, including the potential roles of insulin sensitivity and gut microbiota. 6. Goldacre MJ, Maisonneuve JJ. Mortality from meningococcal disease by day of the week: English national linked database study. J Public Health (Oxf). 2013 Sep;35(3):413-21. doi: 10.1093/pubmed/fdt004. Epub 2013 Feb 1 Key point: This study directly informed the legal debate about the 7-day NHS and the "weekend effect". Detail: This study featured in various mainstream news outlets at the time and was referenced in the High Court judicial review case between NHS junior doctors, the British Medical Association and the Secretary of State for Health in relation to the new NHS contract for junior doctors. The study was described by Mr Justice Green as a "trenchant" piece of evidence. 7. Dharmasena A, Hall N, Goldacre R, Goldacre MJ. Time trends in ophthalmia neonatorum and dacryocystitis of the newborn in England, 2000-2011: database study. Sex Transm Infect. 2015;91:342-5 Key point: This study demonstrated to Public Health England that linked hospital data are the best available data for routinely monitoring the national incidence of newborn conjunctivitis. Detail: The annual figures for this notifiable disease that were reported during the study period under statutory health protection regulations drastically underestimated the actual occurrence of this disease among individuals in hospital (only 1 in 20 cases were reported to Public Health England). 8. Mukhtar TK, Yeates DR, Goldacre MJ. Breast cancer mortality trends in England and the assessment of the effectiveness of mammography screening: population-based study. J R Soc Med. 2013 Jun;106(6):234-42 Key point: This study directly informed the public debate about breast cancer screening. Detail: Publication of this study was immediately followed by an NHS news release, which stated: “There is a great deal of information on both the pros and cons of screening…This study provides additional valuable population data to inform the breast cancer screening debate.”

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: + HES-ID to MPS-ID HES Admitted Patient Care

Processing activities

[7 paragraphs unchanged] e. Hospital Episode Statistics (HES) for England, linked, 1998/99-2018/19 1998/99-2020/21 (provisional); [12 paragraphs unchanged] Except where specified, the datasets are not linked to one another as [83 words unchanged] – e.g. electronic tables for admission rates in the years 1968-1985, 1989-1998, 1989/90-2018/19 1989/90-2020/21 are brought together – into combined tables for the whole period 1968-2018/19. 1968-2020/21. [11 paragraphs unchanged] Similarly, in studies that require data for the full length of the [58 words unchanged] e.g. electronic tables for admission rates in the years 1963-1998 and for 1999-20118/19 1999-2018/19 – into a combined table for the whole period 1963-2018/19. [8 paragraphs unchanged]

Unchanged: Objective for processing, Expected output, Expected measurable benefits, Benefits reported.