ASCEND (A Study of Cardiovascular Events iN Diabetes)

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

Objective for processing

[1 paragraph unchanged] The aim of ASCEND is to determine reliably whether low dose aspirin and/or supplementation with omega-3 fatty acids (FA), safely prevents cardiovascular events (such as heart attacks and strokes), and deaths in patients with diabetes, who have not previously been diagnosed with arterial disease. Aspirin is recommended for people with established arterial disease. However, since it also causes bleeding, the balance of benefits and possible harms were not clear in this group of people with diabetes. The following is a summary of the aim of the research project provided by University of Oxford: The aim of ASCEND is to determine reliably whether low dose aspirin and/or supplementation with omega-3 fatty acids (FA), safely prevents cardiovascular events (such as heart attacks and strokes), and deaths in patients with diabetes, who have not previously been diagnosed with arterial disease. Aspirin is recommended for people with established arterial disease. However, since it also causes bleeding, the balance of benefits and possible harms were not clear in this group of people with diabetes. This question is relevant to many millions of people worldwide with diabetes. [117 words unchanged] stroke prevention. The ASCEND study was designed to help answer this question. [3 paragraphs unchanged] The main research objective for the ASCEND long-term follow-up, is to determine [43 words unchanged] as on heart attacks and strokes, will also be valuable to assess. For example, in the WOSCOP*s study of pravastatin to prevent cardiovascular disease, a protective effect on heart failure was only identified after 5-10 years since this complication of cardiovascular disease takes years to develop. Other analyses may be undertaken if new related, scientific questions arise. *The West of Scotland Coronary Prevention Study (WOSCOPS) was a randomized, placebo-controlled, primary prevention trial of pravastatin in men aged 45 to 64 (mean age of 55 years) with no history of myocardial infarction at randomization. Subjects were enrolled from 1989 to 1991 with the final follow-up visits in 1995. The ASCEND trial provides one of the first opportunities to prospectively test the hypothesis that aspirin prevents gastrointestinal and overall cancer incidence and death, both during the scheduled treatment period and during the ongoing long-term follow-up. Aspirin did not reduce the risk of cancer during the scheduled treatment phase of the study but this was too early to give a reliable answer as cancers may emerge later, hence the main analyses are pre-specified to be undertaken in 2024 and 2029 (5 and 10 years into the end of the scheduled treatment period). The ASCEND trial provides one of the first opportunities to prospectively test the hypothesis that aspirin prevents gastrointestinal and overall cancer incidence and death, both during the scheduled treatment period and during the ongoing long-term follow-up. Aspirin did not reduce the risk of cancer during the scheduled treatment phase of the study but this was too early to give a reliable answer as cancers may emerge later, hence the main analyses are pre-specified to be undertaken in 2024 and 2029 (5 and 10 years into the end of the scheduled treatment period). Note: It had originally been intended to undertake the 5-year analysis in 2022 and 10-year in 2027. However, a lag in the provision of cancer data means that a complete data set is unlikely to be available until Q1 2024. The following NHS England Data will be accessed: Cancer data already held includes the period of the study as well as from the period before the study started. Participants were eligible to take part if they had a history of cancer more than 5 years previously. If they report a new cancer, it is important to ascertain whether this is a new cancer or the recurrence of one that was diagnosed some years before. The analyses includes assessing the impact of the study treatments on newly diagnosed cancers so it is important to be able to be sure that reported cancers are new. The source of the medical events of interest to these analyses is from the NHS England Cancer and Civil Registrations of Deaths data and also the Hospital Episode Statistics (HES) Outpatient (OP), and HES Admitted Patient Care (APC) datasets. ASCEND represents a unique opportunity to assess the effects of aspirin, and of omega-3 FA, on dementia and cognitive impairment cost effectively. Results of the effects of aspirin on cognitive outcomes up to 2-years after the scheduled treatment period (to allow for delays in recording of dementia in hospitalisation datasets) were presented at the American Heart Association in 2021 and published in the European Heart Journal in 2022. This analysis did not demonstrate a clear effect of aspirin on dementia and cognitive impairment and long-term follow-up is needed to fully answer this questions. Alzheimer’s Research UK provided a grant to allow an assessment of cognitive function in the ASCEND trial participants around the time of the end of the scheduled treatment period. It is known that there is a link between dementia and diseases of the heart or circulation. Risk factors for circulatory diseases such as smoking, diabetes and high blood pressure are also more common in people who develop dementia. Participants voluntarily took part and either completed a cognitive function test conducted via telephone by a research nurse or an on-line version sent via email. Linkage with long-term dementia risk would allow methodological questions to be answered such as assessing the utility of different cognitive function measurements. Lipoprotein(a) (or Lp(a)) samples from participants’ blood samples will be sent to a laboratory in the United States for analysis. Samples will be sent with only the participant study identifier and no personal data. All data will be returned to the University of Oxford for statistical analysis. During the main trial analysis, study outcomes reported by participants or their doctors were adjudicated (i.e. study clinicians reviewed copies of relevant medical notes obtained from the participant’s GP or other sources and assessed whether the outcome really took place using standard criteria). Comparison with outcomes identified using the routinely collected healthcare datasets from NHS England and other health data providers enables methods for future trials to be developed. An analysis from ASCEND assessing the utility of cardiovascular outcomes identified from hospitalisation and mortality datasets compared with the ‘gold-standard’ adjudicated outcomes has been published (doi:10.1001/jamanetworkopen.2021.39748) but showed that these dataset missed around 15% of heart attacks, 20% of strokes and 60% of transient ischaemic attacks or ‘mini’ strokes. Further analyses assessing other potential sources of data about these outcomes (e.g. from a combination of HES Outpatients, Diagnostic Imaging and initiation of particular therapies using the NHSBSA medicines data) would help to find out the best way of identifying cardiovascular outcomes in routinely collected healthcare datasets. In addition, other outcomes such as major bleeding will be assessed. These results will inform the methods to be used for the long-term ASCEND follow-up and other studies. Thus, the long-term follow-up through linkage with datasets from NHS England and other data providers allows several additional important scientific questions to be answered from the participants’ contribution during the scheduled treatment period. DATA LINKAGE Participants were resident in and recruited from across the UK, and may have moved between nations since recruitment, and therefore the ASCEND study team would like to link all participants (including those resident in Scotland during the scheduled treatment period) to data held across England and Wales. The long-term follow-up is planned to continue until 2037 with information continuing to be collected from NHS England, and equivalent NHS registries in Scotland and Wales, about causes of death, cancer diagnoses, hospital admissions, attendances at out-patient clinics or A&E departments, and critical care episodes. Additionally, information about mental health, strokes, prescriptions, diabetes, and diagnostic imaging along with data on diabetes complications from the National Diabetes Audit will be collected. These data will be used to assess whether any benefits of aspirin and the omega-3 fish oils observed within the trial, continue long-term, or whether new benefits emerge. These analyses will include all 15,480 people within the UK, who were formerly entered (“randomised”) into the study between June 2005 and July 2011 (unless they withdrew consent). The main analyses will involve comparisons among all those previously allocated aspirin compared with those previously allocated placebo (dummy) and similarly between those previously allocated fish oil supplements versus those allocated placebo. The study team request linkage only to those datasets and data fields relevant to the research, for only the consented participants of the study. The study team will also link to equivalent datasets received from Digital Health & Care Wales (DHCW), the Secure Anonymised Information Linkage dataset (SAIL), the Scottish NHS Central Register (NHSCR) and Public Health Scotland (PHS). The study team intend to link data from participant DNA analyses to NHS England Data. Linkage to longer-term outcomes (including those identified from NHS England Data) is of considerable importance in fully understanding response to aspirin, other cardiovascular protective medications such as statins and cardiovascular risk and its consequences as a whole. Genetic data may be able to add insights into our understanding of the effects of statins, determinants and mechanisms of risk factors and disease outcomes, that can contribute to the development of better patient treatments and precision medicine approaches. NHS England Data will not be shared with any organisation doing the genetic work, and all analysis will be done at the University of Oxford. [2 paragraphs unchanged] • HES Admitted Patient Care, HES Outpatient, Accident and Emergency (replaced by [14 words unchanged] as a source of information about heart attacks, strokes and heart failure. There may also be some information about cancers (from HES APC & OP). These datasets will also be a part of the work to assess the utility of routinely collected healthcare data for identification of cardiovascular and other outcomes in trials. HES APC and OP contain information about attendance at relevant clinics and medical events that will be used in the cognitive function analyses for the dementia research. • The Mental Health datasets (to include Mental Health and Learning Disabilities [11 words unchanged] Health Minimum Data Set (MHMDS) – can provide information about dementia treatments. These datasets may contribute to the work to assess the utility of routinely collected healthcare data for identification of dementia outcomes in trials. [1 paragraph unchanged] This dataset can also contribute information about prescriptions of dementia treatments. This dataset will also be a part of the work to assess the utility of routinely collected healthcare data for identification of cardiovascular and other outcomes in trials. The request for this dataset meets the criteria set out in the NHS BSA Direction for data usage as ASCEND will provide intelligence about the long-term safety and effectiveness of medications tested in the trial i.e. aspirin and omega-3 FA. Both these medications are widely used in clinical practice and so information about their long-term safety and effectiveness is relevant to large numbers of patients in the UK and globally. • Demographics data is being requested for the purpose of being able to identify when to censor participants during follow-up. The censor date is needed to calculate the follow-up time for each participant. • Demographics data is being requested for the purpose of being able to identify when to censor participants during follow-up. The censor date is needed to calculate the follow-up time for each participant. The follow-up time is ‘censored’ on this date because the trial team would not get information about any outcomes occurring after this date. The variables Reason_for_removal, and Date_of_removal are required for this. The study team also use fact of death, plus formal and informal date of death data to contribute information about participants who have died. [1 paragraph unchanged] • National Diabetes Audit (NDA) data will help to identify diabetes related outcomes, in particular diabetic retinopathy, macular degeneration and kidney disease. This dataset can also be used to find ways of identifying outcomes in combination with other datasets which can be used in future trials e.g. for the ASCEND PLUS trial. Linkage with the National Diabetes Audit (NDA) data will contribute to the ASCEND Eye sub-study – for which involvement was voluntary - which assesses the effects of taking aspirin and/or omega-3 fish oils on diabetic retinopathy and age-related macular degeneration, a leading cause of vision loss in people over age 60. This project will be part of a team member’s DPhil (being undertaken at the Nuffield Department of Population Health (NDPH) at University of Oxford). The Macular Society has provided a grant to go towards the costs of this work. As part of this sub-study, participants were asked to complete a questionnaire about their vision and additional data processing activity for the eye assessment sub-study are ongoing during the long-term phase of the study. Diabetic retinopathy data will be derived from the NHS Diabetic Eye Screening Programme (DESP) in England and the Diabetic Eye Screening Wales (DESW) (previously the Diabetic Retinopathy Screening Service for Wales). The eye sub-study includes data about eye conditions previously collected during the scheduled treatment phase of the study, some of which were sourced from the following NHS England datasets: the HES Accident and Emergency, HES Outpatient, and HES Admitted Patient Care. Furthermore, since the National Diabetes Audit data is now collecting retinopathy data in some regions, data from the audit will be used to supplement the analyses using data from DESP and the Diabetic Eye Screening Wales (DESW), in particular in assessing the impact of the study treatments on long-term retinopathy outcomes. In addition, data on diabetes control from the National Diabetes Audit will be used to understand the relevance of the results on diabetic retinopathy among people with different levels of blood sugar control. • National Diabetes Audit (NDA) data will help to identify diabetes related outcomes, in particular diabetic retinopathy, macular degeneration and kidney disease. This dataset can also be used to find ways of identifying outcomes in combination with other datasets which can be used in future trials e.g. for the ASCEND PLUS trial. For the main trial analysis, study outcomes reported by participants or their doctors were adjudicated (i.e. confirmed using copies of the medical notes obtained from the participant’s doctors). However it may be necessary to seek further clarification from GPs about particular events reported during the trial follow-up or identified within HES or other routinely collected healthcare data to confirm accuracy. In order to ensure that requests for supporting documents are sent to the correct GP surgery we are requesting an updated GP practice code for ASCEND participants. The level of the Data will be : Work is ongoing to assess the utility of routinely collected healthcare data for identification of cardiovascular and other outcomes in trials. A paper on cardiovascular outcomes has been published (see Outputs section) and work is ongoing to assess other outcomes including major bleeding. This work will primarily use HES and Civil Registrations data but the utility of additional datasets, including DIDS and NHSBSA medicines data in identifying relevant outcomes is planned. Identifiable – necessary for validation and data integrity checks LEGAL BASIS FOR PROCESSING OF DATA The Data will be minimised as follows: The ASCEND study processes and stores data and special category data in accordance with Article 6(1)(e)(UK GDPR) i.e. it is a task in the public interest, and 9(2)(j)(UK GDPR) i.e. the processing is necessary for archiving purposes in the public interest, scientific or historical research purposes. The processing of personal sensitive data is in the public interest as it will potentially provide information to guide doctors’ decisions about the management of their patients. The processing is of medical data about particular consented individuals and will be related to their involvement in the ASCEND trial. • Limited to a study cohort size of 15480 participants The research team have taken the opinion of the North West - Haydock Research Ethics Committee, who have granted approval to the study in its current form. University of Oxford is the research sponsor and controller as the organization responsible for ensuring that the Data will only be processed for the purpose described above. CONTROLLERSHIP AND FUNDING The lawful basis for processing personal data under the UK GDPR is: Article 6(1)(e) - processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller The University of Oxford is the sole Controller and is also the Processor for this study. No other organisations process these data for the purpose described in this Data Sharing Agreement. The lawful basis for processing special category data under the UK GDPR is: Article 9(2)(j) - processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject. A Data Monitoring Committee and a Steering Committee were involved in an advisory capacity and only had access to aggregated outputs. The Data Monitoring Committee was only active during the scheduled treatment phase of the study and is no longer relevant for the long-term follow-up. The processing of personal sensitive data is in the public interest as it will potentially provide information to guide doctors’ decisions about the management of their patients. The processing is of medical data about particular consented individuals and will be related to their involvement in the ASCEND trial The funding for the data linkage costs and costs of undertaking the analyses comes from multiple sources: The funding is provided by British Heart Foundation (BHF) • British Heart Foundation (BHF) Personal Chair grant to Sir Prof Rory Collins (Director of NDPH, University of Oxford) – funding expires in March 2024 The funder will have no ability to suppress or otherwise limit the publication of findings. • Medical Research Council (MRC) grant to the University of Oxford – funding expires in March 2025 The PPIE team at the Nuffield Department of Population Health (NDPH), also known as Oxford Population Health, coordinates the work of three Public Advisory Groups (PAGs). The public contributors were recruited via various methods, such as collaborating with regional and national networks, charities and community groups but also using our communication channels. They come from across the UK with various socioeconomic, ethnic and age backgrounds and were recruited to represent the UK’s diverse population and ensure that they bring their unique lived experiences in each study. • Health Data Research UK (HDRUK) Clinical Trials Programme grant to the University of Oxford – funding expires in March 2028 At a face-to-face meeting of these groups (on 2nd December 2023) these public contributors provided their input on the continued use of data and were supportive of long-term follow-up trials (including ASCEND). Research staff analysing these data in order to produce methodological research (for example developing enhanced methods for reliably identifying trial outcomes using routinely collected datasets) may be funded by relevant trial grants. For example, the grant for the ASCEND PLUS study to the University of Oxford from Novo-Nordisk. Funding for the scheduled treatment part of the trial came from the British Heart Foundation, with support from Bayer (aspirin), Abbott, Mylan and Solvay (fish oils) who provided the study treatments and funding to cover the drug packaging. There are no ongoing obligations to any of these companies and all provided grants had expired by August 2017. Whilst the study has no further obligations to these companies, it is noted that if the study publishes findings which support the beneficial use of certain medicines produced by these companies, they will derive a financial benefit. Alzheimer’s Research UK provided a grant to allow an assessment of cognitive function in the ASCEND trial participants. The grant ran from the 1st August 2016 to 31st July 2018. The Macular Society has provided a grant to go towards the costs of the ASCEND-EYE project. The grant ran from 08-Sep-2015 to 07-Sep-2016.

Processing activities

All of the Data received from NHS England are stored and processed solely by the Nuffield Department of Population Health ASCEND study team, at the University of Oxford. University of Oxford will transfer data to NHS England. The data will consist of identifying details: NHS Number, Date of Birth, Postcode, Gender and a unique ID for the cohort to be linked with NHS England data. The flagged cohort was sent to NHS England by the ASCEND study team (recruitment was completed in 2011) and is currently held by NHS England, under an existing agreement. Regular data updates have been provided to the ASCEND study team with details of any deaths or diagnoses of cancer since 2006. Approval to receive HES data, acquired in 2016, did not require any further participant identifiers to be sent from the ASCEND team as NHS England were able to use the cohort previously provided. NHS England will provide the relevant records from the list of datasets HES APC, HES OP, HES CC, Emergency Care Data Set (ECDS), Demographics, Civil Registration of Death, Cancer Registrations, Mental Health and Learning Disabilities Data Set (MHLDDS), Mental Health Minimum Data Set (MHMDS), Mental Health Services Data Set (MHSDS), National Diabetes Audit (NDA), Diagnostic Imaging Dataset (DIDS) to University of Oxford. The Data will contain directly identifying data items including Date of Birth which are required for validation and data integrity checks. As part of this amended Data Sharing Agreement (Version 9), to ensure that the linkage is up-to-date, the ASCEND study team will send NHS England a list of those study participants who have since withdrawn from the ASCEND study. The Data will not be transferred to any other location Identifiers that NHS England hold are: The Data will be stored on servers at the Nuffield Department of Population Health(NDPH) at the University of Oxford. • Study ID The Data will be accessed onsite at the premises of NDPH. • NHS Number • Postcode • Sex • Date of Birth The ASCEND team will send in the following identifiers: • Study ID • Date of Withdrawal NHS England will link with the following data sets (previously requested in older versions of this DSA), under the ad-hoc dissemination pattern of approx. March 2024 and March 2025: • HES APC • HES OP • HES CC • Emergency Care Data Set (ECDS), which replaced HES A&E in 2019) • Demographics • Civil Registration of Death • Cancer Registrations (plus a third ad-hoc drop in approx. Sept 2024) Additionally, the following new dataset is requested : • Medicine dispensed in Primary Care (NHSBSA data) from 2015 under the ad-hoc dissemination pattern of approx. March 2024 and March 2025. • The study team are also asking for the final Annual Refresh file for HES APC and OP data for the year 2018/2019. Although this data has already been provided in the provisional data format the study team have identified gaps in their data and are requesting the final Annual Refresh to ensure a complete dataset. • The ASCEND team are requesting that NHS England link and extract the new additional datasets to be disseminated once this DSA is signed: • Mental Health and Learning Disabilities Data Set (MHLDDS) from 2014 to 2016 – one-off • Mental Health Minimum Data Set (MHMDS) from 2006 to 2014 – one-off • Mental Health Services Data Set (MHSDS) from 2017 – one-off • National Diabetes Audit (NDA) from 2005 • Diagnostic Imaging Dataset (DIDS) from 2012 These additional datasets are to include historic data and latest available. The data received from NHS England are transferred into the trial database. The ASCEND trial database is stored on a study specific section of a secure computer server within the University of Oxford. The identifiable NHS England data are stored separately on a secure server to which access is granted on a "need to know" basis i.e. the level of access will depend on the staff role. All such access is granted on the instruction of the Information Asset Owner for ASCEND. Access is routinely reviewed and revoked when the team member leaves ASCEND. The data supplied by NHS England are linked back to the study database and it is critical that this is robust. Linking the wrong medical information to a participant could result in inaccuracies in the trial results. NHS England Data will be linked to data collected from participants during the trial, including baseline clinical, genetic and biomarker data (including those received from Scotland and Wales). The study team will also link to equivalent datasets received from Digital Health & Care Wales (DHCW), the Welsh Secure Anonymised Information Linkage dataset (SAIL), the Scottish NHS Central Register (NHSCR) and Public Health Scotland. As part of the main trial, participant-reported medical events were recorded. These medical events allow detailed assessment of both the safety and efficacy of the trial treatments and were coded by the study medical team who required additional evidence (such as information about diagnoses as well as dates and nature of operations). This evidence was largely provided from the datasets held by NHS England, providing details of hospital attendances and admissions, medical diagnoses and, for any participant death, the cause of death information from the death certificate, as well as the participant’s GP. These data allowed the study medical team to identify medical events not previously reported. This robust process ensures the results are reliable and unbiased and so comply with current regulations governing clinical trials. The source of medical events during the long-term follow-up phase of the study is intended to be mainly from NHS registries, although confirmation from participant’s GPs may be requested for selected outcomes. Although the University of Oxford has minimised the supplied personal identifiers in the data requested, the remaining supplied data will still be linked to the existing study database, which does store the identifying data. There is an ongoing need to retain identifying data to enable further data linkages, such as that which will take place under this DSA. In addition, it is Nuffield Department of Population Health (NDPH) policy to retain research data for at least 25-years after the end of the trial as per the 2014 Clinical Trials Regulation (EU) No 536/2014: https://ec.europa.eu/health/system/files/2016-11/reg_2014_536_en_0.pdf. [1 paragraph unchanged] - Remote access will only be from secure locations situated within the territory of use stated within this DSA; The Controller(s) must confirm and provide evidence upon audit by NHS England that access via any remote device complies with the data security obligations within this DSA and the Data Sharing Framework Contract For Remote Access: - Remote access will only be from secure locations situated within the territory of use (as further restricted elsewhere within the DSA if so done) stated within this DSA; [4 paragraphs unchanged] - All remote access is undertaken within the scope of the organisation’s DSPT (or other security arrangements as per this DSA) and complies with the organisation’s remote access policy. Remote processing will be from secure locations within the UK. The data will not leave the UK at any time. [1 paragraph unchanged] The Data will not leave the UK at any time. All personnel accessing the Data have been appropriately trained in data protection and confidentiality. NDPH researchers are experienced in handling confidential and participant sensitive data and have appropriate training in information governance. The NDPH servers at the University of Oxford are protected against unauthorised external access by an appropriate strength firewall. The Data will not be linked with any other data. Access to patient identifiable information is protected by the appropriate authentication procedures (user IDs and passwords). Authentication is only given to personnel with a need to access the required data. NDPH has a Corporate Level Security Policy and Personal Devices Policy that has been fully adopted by management and will apply fully to the long-term follow-up study. The identifying details will be stored in a separate database to the linked dataset used for analysis. All analyses will use the pseudonymised dataset. There will be no requirement and no attempt to reidentify individuals when using the pseudonymised dataset. All information is stored securely by the University of Oxford and is kept confidential. Access to the study computer database is by unique combinations of usernames and passwords and only authorised study personnel can access information about participants. Analysts from the University of Oxford will analyse the Data for the purposes described above. Servers that store data are located in a secure building with authorised swipe card access only. The ASCEND study team shall not provide or otherwise make available the NHS England Data to any third party or allow use of it or them by or on behalf of any third party, in whole or in part, whether by way of sale, resale, loan, transfer, hire or any other form of exploitation. However, there are two exceptions, these are: 1. the situation where there may be a need to write to an individual participant’s GP for further clarification about a medical event or death to support the medical event coding process undertaken by the study medical team. 2. the situation where an MHRA inspector, as described in the previous paragraph, may see an individual record with a medical event or death recorded, which may have been supplied by NHS England. No individuals will be identified in any study reports. No participants will be identifiable in any information provided to funding organisations and pharmaceutical companies and they do not have any influence in the research or the results. All data provided to funding organisations and pharmaceutical companies will be aggregate with small numbers suppressed in line with the HES Analysis Guide. No record-level pseudonymised NHS England or identifiable Data will be shared other than with substantive employees of the Controller and as described in this section of the Data Sharing Agreement. Other than the linkages already described within this Agreement, researchers will not link NHS England Data to other datasets.

Unchanged: Expected output, Expected measurable benefits, Benefits reported.

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

Datasets: + Diagnostic Imaging Data Set (DID); + HES-ID to MPS-ID HES Admitted Patient Care; + HES-ID to MPS-ID HES Outpatients; + Medicines dispensed in Primary Care (NHSBSA data); + Mental Health Minimum Data Set (MHMDS); + Mental Health Services Data Set (MHSDS); + Mental Health and Learning Disabilities Data Set (MHLDDS)

Objective for processing

The aim of ASCEND (A Study of Cardiovascular Events iN Diabetes) is to determine reliably whether low dose aspirin and/or supplementation with omega-3 fatty acids, safely prevents cardiovascular events (such as heart attacks and strokes), and deaths in patients with diabetes, who have not previously been diagnosed with arterial disease. Although aspirin is recommended for people with established arterial disease, since it also causes bleeding, the balance of benefits and possible harms are not clear in this group of people with diabetes. The University of Oxford requires access to NHS England Data for the purpose of the following research project: ASCEND (A Study of Cardiovascular Events iN Diabetes) This question is relevant to many millions of people worldwide with diabetes. According to Diabetes UK, in February 2019, 3.9 million people in the UK had a diagnosis of diabetes and this number is rising each year. The World Health Organisation reports the number of people with diabetes worldwide rose from 108 million in 1980 to 422 million in 2014 and the global prevalence of diabetes among adults over 18 years of age rose from 4.7% in 1980 to 8.5% in 2014. Adults with diabetes are at an increased risk of heart attacks and strokes and it is not clear whether aspirin should be routinely recommended for heart disease and stroke prevention and the ASCEND study was designed to help answer this question. The aim of ASCEND is to determine reliably whether low dose aspirin and/or supplementation with omega-3 fatty acids (FA), safely prevents cardiovascular events (such as heart attacks and strokes), and deaths in patients with diabetes, who have not previously been diagnosed with arterial disease. Aspirin is recommended for people with established arterial disease. However, since it also causes bleeding, the balance of benefits and possible harms were not clear in this group of people with diabetes. The ASCEND study processes data and special category data in accordance with Article 6(1)(e) and 9(2)(j) of the General Data Protection Regulation respectively. The processing of personal sensitive data is in the public interest as it will potentially provide information to guide doctors’ decisions about the management of their patients. The processing is of medical data about particular individuals and was related to their involvement in the randomised part of the trial when they were either allocated active or placebo (dummy) aspirin and active or placebo fish oils. This question is relevant to many millions of people worldwide with diabetes. According to Diabetes UK, in February 2023 4.3 million people in the UK had a diagnosis of diabetes and this number is rising each year. The World Health Organisation reports the number of people with diabetes worldwide rose from 108 million in 1980 to 422 million in 2014 at which time 8.5% of adults aged 18 years or older had diabetes. In 2019, diabetes was the direct cause of 1.5 million deaths and 48% of all deaths due to diabetes occurred before the age of 70. Adults with diabetes are at an increased risk of heart attacks and strokes and, when ASCEND began, it was not clear whether aspirin should be routinely recommended for heart disease and stroke prevention. The ASCEND study was designed to help answer this question. By undertaking post-trial follow-up, the study team plan to maximise the value of the contribution made by the ASCEND participants during the scheduled treatment period. After ASCEND was started, further analyses of other aspirin trials suggested that long-term aspirin might protect against cancer, in particular colorectal cancer, although this was not conclusive. The main objective of the ASCEND post-treatment follow-up, is whether aspirin reduces the future risk of cancer. If it does, then this may take up 20 years to become clear. Additional questions, such as the effect of aspirin on future dementia risk, and on the development of heart failure as well as on heart attacks and strokes, will also be valuable to assess. Other analyses may be undertaken if new related, scientific questions arise. ASCEND randomised 15,480 UK participants with diabetes, but who did not already have cardiovascular disease, between June 2005 and July 2011. Results from the main trial (i.e. the period of time that participants were asked to take study treatment between 2005 and 2017, the ‘scheduled treatment period’) showed that aspirin prevented serious vascular disease in these individuals, but, it caused almost as many major bleeds and there was no effect on cancers. By undertaking long-term follow-up (from 2017 onwards), the ASCEND study team plan to maximise the value of the contribution made by the ASCEND participants during the scheduled treatment period. Participants will not be taking study treatment during the long-term follow-up. This phase of the study is planned to continue until at least 2037 with information continuing to be collected from NHS Digital, and related registries, about causes of death, cancer diagnoses and any relevant hospital admissions, attendances at out-patient clinics or A&E departments, and critical care episodes. These data are used to assess whether any benefits of aspirin and the omega-3 fish oils observed within the trial, continue long-term, or whether new benefits emerge. These analyses will include all 15,480 people within the UK, who were formerly entered “randomized” into the study between June 2005 and July 2011 (unless they withdrew consent). Any analyses involves comparisons among all those previously allocated aspirin compared with those previously allocated placebo (dummy) and similarly between those previously allocated fish oil supplements versus those allocated placebo. Study participants were sent an End of Study Participant Information Leaflet with their final follow-up questionnaire between January and July 2017, which provided information about the long-term follow-up, where to get more information and how to opt out. They were also reminded of this long-term data collection within the final follow-up questionnaire itself. In 2018 participants were sent a letter which included information about the study results, the plans for long-term follow-up (and how to opt-out), and a Privacy Notice. Further detail follows about the purpose of the planned cancer analyses. The ASCEND trial provides one of the first opportunities to prospectively test the hypothesis that aspirin prevents gastrointestinal and overall cancer incidence and death, both during the trial and during the ongoing longer term post-trial follow-up. Aspirin did not reduce the risk of cancer during the study treatment phase of the study but this was too early to give a reliable answer as cancers may emerge later, hence the initial analyses are pre-specified to be undertaken in 2022 and 2027 (5 and 10 years into the post-trial follow-up period). The main research objective for the ASCEND long-term follow-up, is to determine whether aspirin reduces the future risk of cancer. If it does, then this may take up to 20 years to become clear. Additional objectives, such as the effect of aspirin on future dementia risk, and on the development of heart failure as well as on heart attacks and strokes, will also be valuable to assess. For example, in the WOSCOP*s study of pravastatin to prevent cardiovascular disease, a protective effect on heart failure was only identified after 5-10 years since this complication of cardiovascular disease takes years to develop. Other analyses may be undertaken if new related, scientific questions arise. Cancer data already held includes the period of the study as well as from the period before the study started. Participants were eligible to take part if they had a history of cancer more than 5 years previously; if they report a new cancer, it is important to ascertain whether this is a new cancer or the recurrence of one that was diagnosed some years before. The analyses includes assessing the impact of the study treatments on newly diagnosed cancers so it is important to be able to be sure that reported cancers are new. The source of the medical events of interest to these analyses is from the NHS Digital cancer and death registry data and also the Outpatient, and Admitted Patient Care data-sets. *The West of Scotland Coronary Prevention Study (WOSCOPS) was a randomized, placebo-controlled, primary prevention trial of pravastatin in men aged 45 to 64 (mean age of 55 years) with no history of myocardial infarction at randomization. Subjects were enrolled from 1989 to 1991 with the final follow-up visits in 1995. Alzheimer’s Research UK provided a grant to allow an assessment of cognitive function in the ASCEND trial participants around the time of the end of the scheduled treatment period. It is known that there is a link between dementia and diseases of the heart or circulation. Risk factors for circulatory diseases such as smoking, diabetes and high blood pressure are also more common in people who develop dementia. However, it is not known whether treatments such as aspirin, which can help to protect against heart attacks and strokes, might affect memory and brain power and affect the risk of developing dementia. Participants voluntarily took part and either completed a cognitive function test conducted via telephone by a research nurse or an on-line version sent via email. The ASCEND trial provides one of the first opportunities to prospectively test the hypothesis that aspirin prevents gastrointestinal and overall cancer incidence and death, both during the scheduled treatment period and during the ongoing long-term follow-up. Aspirin did not reduce the risk of cancer during the scheduled treatment phase of the study but this was too early to give a reliable answer as cancers may emerge later, hence the main analyses are pre-specified to be undertaken in 2024 and 2029 (5 and 10 years into the end of the scheduled treatment period). Note: It had originally been intended to undertake the 5-year analysis in 2022 and 10-year in 2027. However, a lag in the provision of cancer data means that a complete data set is unlikely to be available until Q1 2024. ASCEND represents a unique opportunity to assess the effect of aspirin, and of omega-3 FA, on cognitive function cost-effectively. A planned meta-analysis between ASCEND and the ongoing 19,000 participant ASPREE study (in addition to combining with data from earlier studies if possible) would have good power to detect a small, but important, effect of aspirin on cognitive function. Cancer data already held includes the period of the study as well as from the period before the study started. Participants were eligible to take part if they had a history of cancer more than 5 years previously. If they report a new cancer, it is important to ascertain whether this is a new cancer or the recurrence of one that was diagnosed some years before. The analyses includes assessing the impact of the study treatments on newly diagnosed cancers so it is important to be able to be sure that reported cancers are new. The source of the medical events of interest to these analyses is from the NHS England Cancer and Civil Registrations of Deaths data and also the Hospital Episode Statistics (HES) Outpatient (OP), and HES Admitted Patient Care (APC) datasets. Further to the on-study treatment analyses of the effects of aspirin on cognitive outcomes, 5-year results from analyses were published in 2022 and there are further analyses planned for 10 years after the end of the scheduled treatment period (i.e. including data up to 31.3.2027). The source of the medical events of interest to the cognitive function analyses is from the following NHS Digital registry data-sets: Admitted Patient Care and the Outpatient data on attendance at memory clinics. The source of the information on heart attacks, strokes and heart failure is from the following NHS Digital Hospital data-sets: Admitted Patient Care, Outpatient, Accident and Emergency (being replaced by the Emergency Care data-set) and Critical Care. ASCEND represents a unique opportunity to assess the effects of aspirin, and of omega-3 FA, on dementia and cognitive impairment cost effectively. Additionally the ASCEND Eye sub-study, for which involvement was voluntary, assesses the effects of taking aspirin and/or omega-3 fish oils on diabetic retinopathy and age-related macular degeneration, a leading cause of vision loss in people over age 60. Data processing activity for the eye assessment sub-study will be on-going during the long-term phase of the study but will not involve linking to NHS Digital controlled data-sets as diabetic retinopathy data will derive from the NHS Diabetic Eye Screening Programme (DESP) in England and the Diabetic Retinopathy Screening Service for Wales. However the eye sub-study includes data about eye conditions previously collected during the main treatment phase of the study, some of which were sourced from the following NHS Digital data-sets : the Accident and Emergency, Outpatient and Admitted Patient Care. Results of the effects of aspirin on cognitive outcomes up to 2-years after the scheduled treatment period (to allow for delays in recording of dementia in hospitalisation datasets) were presented at the American Heart Association in 2021 and published in the European Heart Journal in 2022. This analysis did not demonstrate a clear effect of aspirin on dementia and cognitive impairment and long-term follow-up is needed to fully answer this questions. The University of Oxford is the sole Data Controller and also processes the data for this study. No other organisations process these data for the purpose described in this Agreement. A Data Monitoring Committee and a Steering Committee have been involved in an advisory capacity and have only had access to aggregated outputs. The Data Monitoring Committee was only active during the treatment phase of the study and is no longer relevant for the long-term follow-up. The study is currently funded by the Clinical Trials Service Unit (CTSU) core grants from the British Heart Foundation (BHF), Medical Research Council (MRC) and Cancer Research UK. Separate project grant funding is to be sought for the Long-Term Follow-Up from the BHF. Alzheimer’s Research UK provided a grant to allow an assessment of cognitive function in the ASCEND trial participants around the time of the end of the scheduled treatment period. It is known that there is a link between dementia and diseases of the heart or circulation. Risk factors for circulatory diseases such as smoking, diabetes and high blood pressure are also more common in people who develop dementia. Participants voluntarily took part and either completed a cognitive function test conducted via telephone by a research nurse or an on-line version sent via email. Linkage with long-term dementia risk would allow methodological questions to be answered such as assessing the utility of different cognitive function measurements. Lipoprotein(a) (or Lp(a)) samples from participants’ blood samples will be sent to a laboratory in the United States for analysis. Samples will be sent with only the participant study identifier and no personal data. All data will be returned to the University of Oxford for statistical analysis. During the main trial analysis, study outcomes reported by participants or their doctors were adjudicated (i.e. study clinicians reviewed copies of relevant medical notes obtained from the participant’s GP or other sources and assessed whether the outcome really took place using standard criteria). Comparison with outcomes identified using the routinely collected healthcare datasets from NHS England and other health data providers enables methods for future trials to be developed. An analysis from ASCEND assessing the utility of cardiovascular outcomes identified from hospitalisation and mortality datasets compared with the ‘gold-standard’ adjudicated outcomes has been published (doi:10.1001/jamanetworkopen.2021.39748) but showed that these dataset missed around 15% of heart attacks, 20% of strokes and 60% of transient ischaemic attacks or ‘mini’ strokes. Further analyses assessing other potential sources of data about these outcomes (e.g. from a combination of HES Outpatients, Diagnostic Imaging and initiation of particular therapies using the NHSBSA medicines data) would help to find out the best way of identifying cardiovascular outcomes in routinely collected healthcare datasets. In addition, other outcomes such as major bleeding will be assessed. These results will inform the methods to be used for the long-term ASCEND follow-up and other studies. Thus, the long-term follow-up through linkage with datasets from NHS England and other data providers allows several additional important scientific questions to be answered from the participants’ contribution during the scheduled treatment period. DATA LINKAGE Participants were resident in and recruited from across the UK, and may have moved between nations since recruitment, and therefore the ASCEND study team would like to link all participants (including those resident in Scotland during the scheduled treatment period) to data held across England and Wales. The long-term follow-up is planned to continue until 2037 with information continuing to be collected from NHS England, and equivalent NHS registries in Scotland and Wales, about causes of death, cancer diagnoses, hospital admissions, attendances at out-patient clinics or A&E departments, and critical care episodes. Additionally, information about mental health, strokes, prescriptions, diabetes, and diagnostic imaging along with data on diabetes complications from the National Diabetes Audit will be collected. These data will be used to assess whether any benefits of aspirin and the omega-3 fish oils observed within the trial, continue long-term, or whether new benefits emerge. These analyses will include all 15,480 people within the UK, who were formerly entered (“randomised”) into the study between June 2005 and July 2011 (unless they withdrew consent). The main analyses will involve comparisons among all those previously allocated aspirin compared with those previously allocated placebo (dummy) and similarly between those previously allocated fish oil supplements versus those allocated placebo. The study team request linkage only to those datasets and data fields relevant to the research, for only the consented participants of the study. The study team will also link to equivalent datasets received from Digital Health & Care Wales (DHCW), the Secure Anonymised Information Linkage dataset (SAIL), the Scottish NHS Central Register (NHSCR) and Public Health Scotland (PHS). The study team intend to link data from participant DNA analyses to NHS England Data. Linkage to longer-term outcomes (including those identified from NHS England Data) is of considerable importance in fully understanding response to aspirin, other cardiovascular protective medications such as statins and cardiovascular risk and its consequences as a whole. Genetic data may be able to add insights into our understanding of the effects of statins, determinants and mechanisms of risk factors and disease outcomes, that can contribute to the development of better patient treatments and precision medicine approaches. NHS England Data will not be shared with any organisation doing the genetic work, and all analysis will be done at the University of Oxford. • Civil Registrations of Death – necessary to identify participants who have died, plus cause and date of death. This dataset will also be a part of the work to assess the utility of routinely collected healthcare data for identification of cardiovascular and other outcomes in trials. • Cancer Registration – necessary to identify participants who have had a cancer diagnosis, plus information about the type of cancer, and related dates. • HES Admitted Patient Care, HES Outpatient, Accident and Emergency (replaced by the Emergency Care Data Set), and HES Critical Care – datasets will be used as a source of information about heart attacks, strokes and heart failure. There may also be some information about cancers (from HES APC & OP). These datasets will also be a part of the work to assess the utility of routinely collected healthcare data for identification of cardiovascular and other outcomes in trials. HES APC and OP contain information about attendance at relevant clinics and medical events that will be used in the cognitive function analyses for the dementia research. • The Mental Health datasets (to include Mental Health and Learning Disabilities Data Set (MHLDDS), Mental Health Services Data Set (MHSDS) and Mental Health Minimum Data Set (MHMDS) – can provide information about dementia treatments. These datasets may contribute to the work to assess the utility of routinely collected healthcare data for identification of dementia outcomes in trials. • The Medicines Dispensed in Primary Care (NHSBSA) data set – may be used as a source of information about heart attacks, strokes and heart failure in combination with other datasets (e.g. HES). This dataset can also contribute information about prescriptions of dementia treatments. This dataset will also be a part of the work to assess the utility of routinely collected healthcare data for identification of cardiovascular and other outcomes in trials. The request for this dataset meets the criteria set out in the NHS BSA Direction for data usage as ASCEND will provide intelligence about the long-term safety and effectiveness of medications tested in the trial i.e. aspirin and omega-3 FA. Both these medications are widely used in clinical practice and so information about their long-term safety and effectiveness is relevant to large numbers of patients in the UK and globally. • Demographics data is being requested for the purpose of being able to identify when to censor participants during follow-up. The censor date is needed to calculate the follow-up time for each participant. The follow-up time is ‘censored’ on this date because the trial team would not get information about any outcomes occurring after this date. The variables Reason_for_removal, and Date_of_removal are required for this. The study team also use fact of death, plus formal and informal date of death data to contribute information about participants who have died. • Diagnostic Imaging Dataset: this data set may contribute to identifying or confirming outcomes of heart attacks, strokes and heart failure. It can also be used to find ways of identifying outcomes such as transient ischaemic stroke (TIA or ‘mini stroke’) & stroke in combination with other datasets which can be used in future trials e.g. for the ASCEND PLUS trial. • National Diabetes Audit (NDA) data will help to identify diabetes related outcomes, in particular diabetic retinopathy, macular degeneration and kidney disease. This dataset can also be used to find ways of identifying outcomes in combination with other datasets which can be used in future trials e.g. for the ASCEND PLUS trial. Linkage with the National Diabetes Audit (NDA) data will contribute to the ASCEND Eye sub-study – for which involvement was voluntary - which assesses the effects of taking aspirin and/or omega-3 fish oils on diabetic retinopathy and age-related macular degeneration, a leading cause of vision loss in people over age 60. This project will be part of a team member’s DPhil (being undertaken at the Nuffield Department of Population Health (NDPH) at University of Oxford). The Macular Society has provided a grant to go towards the costs of this work. As part of this sub-study, participants were asked to complete a questionnaire about their vision and additional data processing activity for the eye assessment sub-study are ongoing during the long-term phase of the study. Diabetic retinopathy data will be derived from the NHS Diabetic Eye Screening Programme (DESP) in England and the Diabetic Eye Screening Wales (DESW) (previously the Diabetic Retinopathy Screening Service for Wales). The eye sub-study includes data about eye conditions previously collected during the scheduled treatment phase of the study, some of which were sourced from the following NHS England datasets: the HES Accident and Emergency, HES Outpatient, and HES Admitted Patient Care. Furthermore, since the National Diabetes Audit data is now collecting retinopathy data in some regions, data from the audit will be used to supplement the analyses using data from DESP and the Diabetic Eye Screening Wales (DESW), in particular in assessing the impact of the study treatments on long-term retinopathy outcomes. In addition, data on diabetes control from the National Diabetes Audit will be used to understand the relevance of the results on diabetic retinopathy among people with different levels of blood sugar control. For the main trial analysis, study outcomes reported by participants or their doctors were adjudicated (i.e. confirmed using copies of the medical notes obtained from the participant’s doctors). However it may be necessary to seek further clarification from GPs about particular events reported during the trial follow-up or identified within HES or other routinely collected healthcare data to confirm accuracy. In order to ensure that requests for supporting documents are sent to the correct GP surgery we are requesting an updated GP practice code for ASCEND participants. Work is ongoing to assess the utility of routinely collected healthcare data for identification of cardiovascular and other outcomes in trials. A paper on cardiovascular outcomes has been published (see Outputs section) and work is ongoing to assess other outcomes including major bleeding. This work will primarily use HES and Civil Registrations data but the utility of additional datasets, including DIDS and NHSBSA medicines data in identifying relevant outcomes is planned. LEGAL BASIS FOR PROCESSING OF DATA The ASCEND study processes and stores data and special category data in accordance with Article 6(1)(e)(UK GDPR) i.e. it is a task in the public interest, and 9(2)(j)(UK GDPR) i.e. the processing is necessary for archiving purposes in the public interest, scientific or historical research purposes. The processing of personal sensitive data is in the public interest as it will potentially provide information to guide doctors’ decisions about the management of their patients. The processing is of medical data about particular consented individuals and will be related to their involvement in the ASCEND trial. The research team have taken the opinion of the North West - Haydock Research Ethics Committee, who have granted approval to the study in its current form. CONTROLLERSHIP AND FUNDING The University of Oxford is the sole Controller and is also the Processor for this study. No other organisations process these data for the purpose described in this Data Sharing Agreement. A Data Monitoring Committee and a Steering Committee were involved in an advisory capacity and only had access to aggregated outputs. The Data Monitoring Committee was only active during the scheduled treatment phase of the study and is no longer relevant for the long-term follow-up. The funding for the data linkage costs and costs of undertaking the analyses comes from multiple sources: • British Heart Foundation (BHF) Personal Chair grant to Sir Prof Rory Collins (Director of NDPH, University of Oxford) – funding expires in March 2024 • Medical Research Council (MRC) grant to the University of Oxford – funding expires in March 2025 • Health Data Research UK (HDRUK) Clinical Trials Programme grant to the University of Oxford – funding expires in March 2028 Research staff analysing these data in order to produce methodological research (for example developing enhanced methods for reliably identifying trial outcomes using routinely collected datasets) may be funded by relevant trial grants. For example, the grant for the ASCEND PLUS study to the University of Oxford from Novo-Nordisk. Funding for the scheduled treatment part of the trial came from the British Heart Foundation, with support from Bayer (aspirin), Abbott, Mylan and Solvay (fish oils) who provided the study treatments and funding to cover the drug packaging. There are no ongoing obligations to any of these companies and all provided grants had expired by August 2017. Whilst the study has no further obligations to these companies, it is noted that if the study publishes findings which support the beneficial use of certain medicines produced by these companies, they will derive a financial benefit. Alzheimer’s Research UK provided a grant to allow an assessment of cognitive function in the ASCEND trial participants. The grant ran from the 1st August 2016 to 31st July 2018. The Macular Society has provided a grant to go towards the costs of the ASCEND-EYE project. The grant ran from 08-Sep-2015 to 07-Sep-2016.

Processing activities

All of the Hospital Episode Statistics (HES), mortality and cancer registry data provided by NHS Digital under a previous iteration of this agreement, are stored and processed solely by the Clinical Trial Service Unit ASCEND study team at the University of Oxford. The flagged cohort was sent by the ASCEND study team (linkage was completed in 2011) and is currently held linked to the participants NHS record at NHS Digital, under a previous iteration of this agreement. Regular updates have been provided to the ASCEND study team with details of any deaths or diagnoses of cancer since 2006. Approval to receive HES data, acquired in 2016, did not require any further data files to be sent from the ASCEND team as it was possible for NHS Digital to use the cohort previously provided. All of the Data received from NHS England are stored and processed solely by the Nuffield Department of Population Health ASCEND study team, at the University of Oxford. The data received from NHS Digital were transferred into the trial database by the person registered to receive data. The ASCEND trial database is study specific and is stored on an ASCEND specific computer server. The identifiable NHS Digital data are stored separately on a secure server to which access is granted on a "need to know" basis i.e. the level of access will depend on the staff role. All such access is granted on the instruction of the Information Asset Owner for ASCEND. Access is routinely reviewed and revoked when the team member leaves ASCEND. The data supplied by NHS Digital was linked back to the study database and it is critical that this is robust. Linking the wrong medical information to a participant is potentially significant in terms of the analyses and could be assessed as in breach of Good Clinical Practice during any MHRA inspection. Therefore, the team used the participant’s identifiable details so that multi-part checks can be applied. However, this was reassessed in the light of the main study treatment phase of the study closing, as part of the reason for requesting these identifiers was so that the study team had up-to-date participant contact details. However, given that no further contact with the participants was required plus the need to be mindful of data minimisation principals, the decision was agreed to only receive the study ID and the NHS number and no other personal identifiers. The flagged cohort was sent to NHS England by the ASCEND study team (recruitment was completed in 2011) and is currently held by NHS England, under an existing agreement. Regular data updates have been provided to the ASCEND study team with details of any deaths or diagnoses of cancer since 2006. Approval to receive HES data, acquired in 2016, did not require any further participant identifiers to be sent from the ASCEND team as NHS England were able to use the cohort previously provided. As part of the study, participant-reported medical events were recorded. These medical events allowed detailed assessment of both the safety and efficacy of the trial treatments and were coded by the study medical team who require additional evidence (such as information about diagnoses as well as dates and nature of operations). This evidence was largely provided from the datasets held by NHS Digital, providing details of hospital attendances and admissions, medical diagnoses and, for any participant death, the cause of death information from the death certificate, as well as the participant’s GP. These data also allowed the study medical team to identify medical events not previously reported. This robust process ensured the results were reliable and unbiased and so complying with current regulations governing clinical trials. The source of medical events during the Long-term follow-up phase of the study is intended to be solely from NHS registries. Therefore, there will be no further requirement to request the GP code and if any further contact is required, then the practice code would suffice. As part of this amended Data Sharing Agreement (Version 9), to ensure that the linkage is up-to-date, the ASCEND study team will send NHS England a list of those study participants who have since withdrawn from the ASCEND study. Although the University of Oxford has minimised the supplied personal identifiers in the data requested, the remaining supplied data was linked to the existing study database, which does store the identifying data. The study team was required to do this as the study is still ongoing and has to comply with Good Clinical Practice Guidelines enacted in UK law via the Medicines for Human Use Act. Within the UK, the Medicines and Healthcare Products Regulatory Agency (MHRA) ensure these regulations are enforced via inspections. Identifiers that NHS England hold are: The raw data files are stored securely with access limited to those receiving the files from the NHS registries and those who need to process the data in order to merge into the study-specific database. • Study ID NDPH researchers are experienced in handling confidential and participant sensitive data and have appropriate training in information governance. The NDPH servers are protected against unauthorised external access by an appropriate strength firewall. • NHS Number Access to ASCEND data (including patient identifiable data) is protected by the appropriate authentication procedures (unique user IDs and passwords). Authentication is only granted to personnel with a need to access the required data. Only personnel involved in the long-term follow-up study (processing and analysing data) will have access to the data under this Agreement. NDPH has a Corporate Level Security Policy that has been fully adopted by management and will apply fully to the long-term follow-up study. • Postcode All information is stored securely by the University of Oxford and is kept confidential. No individuals will be identified in any study reports. • Sex Data is stored, accessed and backed up at University of Oxford, Richard Doll Building, Old Road Campus, Roosevelt Drive, Oxford, OX3 7LF and University of Oxford, Big Data Institute Building, Old Road Campus, Roosevelt Drive, Oxford, OX3 7LF. • Date of Birth Servers that store data are located in secure buildings with authorised swipe card access only. The ASCEND team will send in the following identifiers: The ASCEND study team shall not provide or otherwise make available the HES, mortality or cancer registry data to any third party or allow use of it or them by or on behalf of any third party, in whole or in part, whether by way of sale, resale, loan, transfer, hire or any other form of exploitation. However, there are two exceptions, these are; • Study ID • Date of Withdrawal NHS England will link with the following data sets (previously requested in older versions of this DSA), under the ad-hoc dissemination pattern of approx. March 2024 and March 2025: • HES APC • HES OP • HES CC • Emergency Care Data Set (ECDS), which replaced HES A&E in 2019) • Demographics • Civil Registration of Death • Cancer Registrations (plus a third ad-hoc drop in approx. Sept 2024) Additionally, the following new dataset is requested : • Medicine dispensed in Primary Care (NHSBSA data) from 2015 under the ad-hoc dissemination pattern of approx. March 2024 and March 2025. • The study team are also asking for the final Annual Refresh file for HES APC and OP data for the year 2018/2019. Although this data has already been provided in the provisional data format the study team have identified gaps in their data and are requesting the final Annual Refresh to ensure a complete dataset. • The ASCEND team are requesting that NHS England link and extract the new additional datasets to be disseminated once this DSA is signed: • Mental Health and Learning Disabilities Data Set (MHLDDS) from 2014 to 2016 – one-off • Mental Health Minimum Data Set (MHMDS) from 2006 to 2014 – one-off • Mental Health Services Data Set (MHSDS) from 2017 – one-off • National Diabetes Audit (NDA) from 2005 • Diagnostic Imaging Dataset (DIDS) from 2012 These additional datasets are to include historic data and latest available. The data received from NHS England are transferred into the trial database. The ASCEND trial database is stored on a study specific section of a secure computer server within the University of Oxford. The identifiable NHS England data are stored separately on a secure server to which access is granted on a "need to know" basis i.e. the level of access will depend on the staff role. All such access is granted on the instruction of the Information Asset Owner for ASCEND. Access is routinely reviewed and revoked when the team member leaves ASCEND. The data supplied by NHS England are linked back to the study database and it is critical that this is robust. Linking the wrong medical information to a participant could result in inaccuracies in the trial results. NHS England Data will be linked to data collected from participants during the trial, including baseline clinical, genetic and biomarker data (including those received from Scotland and Wales). The study team will also link to equivalent datasets received from Digital Health & Care Wales (DHCW), the Welsh Secure Anonymised Information Linkage dataset (SAIL), the Scottish NHS Central Register (NHSCR) and Public Health Scotland. As part of the main trial, participant-reported medical events were recorded. These medical events allow detailed assessment of both the safety and efficacy of the trial treatments and were coded by the study medical team who required additional evidence (such as information about diagnoses as well as dates and nature of operations). This evidence was largely provided from the datasets held by NHS England, providing details of hospital attendances and admissions, medical diagnoses and, for any participant death, the cause of death information from the death certificate, as well as the participant’s GP. These data allowed the study medical team to identify medical events not previously reported. This robust process ensures the results are reliable and unbiased and so comply with current regulations governing clinical trials. The source of medical events during the long-term follow-up phase of the study is intended to be mainly from NHS registries, although confirmation from participant’s GPs may be requested for selected outcomes. Although the University of Oxford has minimised the supplied personal identifiers in the data requested, the remaining supplied data will still be linked to the existing study database, which does store the identifying data. There is an ongoing need to retain identifying data to enable further data linkages, such as that which will take place under this DSA. In addition, it is Nuffield Department of Population Health (NDPH) policy to retain research data for at least 25-years after the end of the trial as per the 2014 Clinical Trials Regulation (EU) No 536/2014: https://ec.europa.eu/health/system/files/2016-11/reg_2014_536_en_0.pdf. The Data will be accessed by authorised personnel via remote access or on-site. - Remote access will only be from secure locations situated within the territory of use stated within this DSA; - Access controls granting users the minimum level of access required are in place; - Remote access is only via secure connections to protect data; - Multifactor authentication (MFA) is required for remote access; - Device security, including up-to-date software and operating systems, antivirus software, and enabled firewalls are utilised for the remote access; - All remote access is undertaken within the scope of the organisation’s DSPT and complies with the organisation’s remote access policy. Access is restricted to employees or agents of University of Oxford who have authorisation on the instruction of the Information Asset Owner for ASCEND. Only personnel involved in the long-term follow-up study (processing and analysing data) will have access to the data under this Data Sharing Agreement. The Data will not leave the UK at any time. NDPH researchers are experienced in handling confidential and participant sensitive data and have appropriate training in information governance. The NDPH servers at the University of Oxford are protected against unauthorised external access by an appropriate strength firewall. Access to patient identifiable information is protected by the appropriate authentication procedures (user IDs and passwords). Authentication is only given to personnel with a need to access the required data. NDPH has a Corporate Level Security Policy and Personal Devices Policy that has been fully adopted by management and will apply fully to the long-term follow-up study. All information is stored securely by the University of Oxford and is kept confidential. Access to the study computer database is by unique combinations of usernames and passwords and only authorised study personnel can access information about participants. Servers that store data are located in a secure building with authorised swipe card access only. The ASCEND study team shall not provide or otherwise make available the NHS England Data to any third party or allow use of it or them by or on behalf of any third party, in whole or in part, whether by way of sale, resale, loan, transfer, hire or any other form of exploitation. However, there are two exceptions, these are: [1 paragraph unchanged] 2. the situation where an MHRA inspector inspector, as described in the previous paragraph, may see an individual record with a medical event or death recorded, which may have been supplied by NHS Digital. England. No individuals will be identified in any study reports. No participants will be identifiable in any information provided to funding organisations [7 words unchanged] have any influence in the research or the results. All data provided would to funding organisations and pharmaceutical companies will be aggregate with small numbers suppressed in line with the HES Analysis Guide. Data will not be shared with any other organisations. No record-level pseudonymised NHS England or identifiable Data will be shared other than with substantive employees of the Controller and as described in this section of the Data Sharing Agreement. All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract ie: employees, agents and contractors of the Data Recipient who may have access to that data). Other than the linkages already described within this Agreement, researchers will not link NHS England Data to other datasets.

Expected output

The NDPH contributes widely to health policy, particularly in the area of [10 words unchanged] conference participation, lectures to the public and advice to government (including NHS Digital). England). Examples of the impact of the work performed by NDPH up to 2021 is available from: https://results2021.ref.ac.uk/profiles/institutions/10007774 [3 paragraphs unchanged] The outputs from the long-term follow-up phase will be in the form of scientific reports and presentations at National and International Scientific meetings. The ASCEND study team intend to publish results in mid-2024. 2024. The results of the long-term follow-up, particularly on the effects of aspirin [8 words unchanged] analyses on eye outcomes and heart failure, will be directly relevant to 400M over 400 million people with diabetes worldwide. Methodological outputs from this research are intended to inform the design of future trials wishing to collect study outcomes from routinely collected healthcare data. These results will be communicated in peer reviewed journals and disseminated in collaboration with Health Data Research UK and the NIHR/MRC Trials Methodology Research Partnership. [2 paragraphs unchanged] • Journals • Advice to government [2 paragraphs unchanged] • Posters • Posters - displayed at departmental, national and international conferences • Press/media engagement and other public promotion of the research (e.g. via the Nuffield Department of Population Health website (https://www.ndph.ox.ac.uk/), or Twitter account (@oxford_ndph). Population Health website (https://www.ndph.ox.ac.uk/), or X (formerly Twitter) account (@oxford_ndph). [1 paragraph unchanged] The study participants were mailed a summary of the study results shortly after the main results presentation in 2018. The medical consultants representing the various NHS hospital trusts involved, were presented with a summary of the results and attended a specific meeting at which the results were discussed around the same time. This summary was also available via the study-specific web-site along with short videos of the results (https://ascend.medsci.ox.ac.uk/). The effects of aspirin on dementia and cognitive impairment (identified through the trial follow-up procedures or in the HES and Civil Registrations data up to 31 March 2019 (to allow for the expected delay between the diagnosis of dementia and the recording of dementia in HES which is known to be 1.6 years on average) were published in 2022 and showed no effect of aspirin on dementia risk but longer follow-up is needed to reliably assess this question. Analyses of the effects of aspirin on cognitive outcomes are therefore planned for 5 and 10 years after the end of the scheduled treatment period. Heart failure is a major cause of disability and people with diabetes are at increased risk of developing heart failure. The current work in confirming heart failure events identified from the trial follow-up procedures and the linked health records, will be used to assess the impact of aspirin on heart failure during the scheduled treatment period. This will also allow algorithm development to reliably identify heart failure events in the long-term follow-up to allow the assessment of the effects of aspirin on long-term heart failure risk. The study participants were mailed a summary of the study results shortly after the main results presentation in 2018. The medical consultants representing the various NHS hospital trusts involved, were presented with a summary of the results and attended a specific meeting at which the results were discussed around the same time. This summary was also available via the study-specific website along with short videos of the results (https://ascend.medsci.ox.ac.uk/). [2 paragraphs unchanged] Outputs for the ASCEND trial so far includes multiple publications in peer-reviewed [5 words unchanged] conferences. Information about the publications can be found on the ASCEND website (https://ascend.medsci.ox.ac.uk/) and the benefits are described in Section 6iii (Yielded Benefits). (https://ascend.medsci.ox.ac.uk/).

Expected measurable benefits

[1 paragraph unchanged] It had been suggested that low-dose aspirin might protect against cancer, and [19 words unchanged] showed no reduction in any cancers during a mean 7.3 years of follow-up.. follow-up. However, the main focus of the analyses will be after long-term follow-up, [11 words unchanged] in cancer incidence between the arms due to larger numbers of events. [2 paragraphs unchanged] The effects of aspirin on dementia and cognitive impairment (identified through the trial follow-up procedures or in the HES and Civil Registrations data up to 31 March 2019 (to allow for the expected delay between the diagnosis of dementia and the recording of dementia in HES which is known to be 1.6 years on average) were published in 2022 and showed no effect of aspirin on dementia risk but longer follow-up is needed to reliably assess this question. Analyses of the effects of aspirin on cognitive outcomes are therefore planned for 5 and 10 years after the end of the scheduled treatment period. [1 paragraph unchanged] Heart failure is a major cause of disability and people with diabetes are at increased risk of developing heart failure. The current work in confirming heart failure events identified from the trial follow-up procedures and the linked health records, will be used to assess the impact of aspirin on heart failure during the scheduled treatment period. This will also allow algorithm development to reliably identify heart failure events in the long-term follow-up to allow the assessment of the effects of aspirin on long-term heart failure risk.

Benefits reported

[3 paragraphs unchanged] • Post hoc analyses of the ASCEND trial suggest that routinely collected [32 words unchanged] primary prevention cardiovascular trials, without the need for verification by clinical adjudication. These results were published in Heart 2023; 109:1467-1472. [1 paragraph unchanged] Evidence from large trials now suggests that omega-3 FA supplementation may have a dose-related protective effect on coronary events but it has been suggested that the risk of atrial fibrillation (AF) is increased. Analysis of atrial fibrillation identified from ASCEND trial follow-up procedures and linked health records did not show and increase in AF risk of omega-3 FA supplementation. However, systematic reporting of arrhythmia outcomes in existing and future trials is required to clarify whether supplementation also has any adverse effects on nonfatal arrhythmias. • An ASCEND team member undertook exploratory analyses of the effects of aspirin and omega-3 fatty acids on heart failure (HF) outcomes in ASCEND. The results (reported at the European Society of Cardiology in 2022) showed that data linkage with routinely collected data, followed by adjudication of clinical and routinely collected data, allowed the ASCEND trial to identify many additional participants with heart failure compared to patient reported events alone. This finding will allow for more effective assessment of the effects of aspirin and omega-3 on heart failure. Evidence from large trials now suggests that omega-3 FA supplementation may have a dose-related protective effect on coronary events but it has been suggested that the risk of atrial fibrillation (AF) is increased. Analysis of atrial fibrillation identified from ASCEND trial follow-up procedures and linked health records did not show an increase in AF risk of omega-3 FA supplementation. However, systematic reporting of arrhythmia outcomes in existing and future trials is required to clarify whether supplementation also has any adverse effects on nonfatal arrhythmias. [4 paragraphs unchanged] The main results of the ASCEND trial now provide much needed clarity. [55 words unchanged] diabetes, pre-diabetes, and cardiovascular diseases developed in collaboration with the EASD). The results of ASCEND were specifically mentioned in the Feb 2023 update to the NICE guideline Cardiovascular disease: risk assessment and reduction, including lipid modification [NICE guideline CG181]. The World Health Organisation statistics detail 422 million state that the number of people with diabetes rose from 108 million in 1980 to 422 million in 2014, and that between 2000 and 2019, there was a rise of almost 400% since 1980. 3% increase in diabetes mortality rates by age. The prevalence of diabetes in the UK and global population is expected to continue to rise. [3 paragraphs unchanged]

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

Objective for processing

[1 paragraph unchanged] This question is relevant to many millions of people worldwide with diabetes. According to Diabetes UK, in November 2016 3.6 February 2019, 3.9 million people in the UK had a diagnosis of diabetes and this [5 words unchanged] The World Health Organisation reports the number of people with diabetes worldwide has risen rose from 108 million in 1980 to 422 million in 2014 and the global prevalence of diabetes among adults over 18 years of age has risen rose from 4.7% in 1980 to 8.5% in 2014. Adults with diabetes are [18 words unchanged] routinely recommended for heart disease and stroke prevention and the ASCEND study is was designed to help answer this question. The ASCEND study processes data and special category data in accordance with [36 words unchanged] their patients. The processing is of medical data about particular individuals and will be was related to their involvement in the randomised part of the trial when they were either were allocated active or placebo (dummy) aspirin and active or placebo fish oils. [1 paragraph unchanged] The main question to be addressed by objective of the ASCEND post-treatment follow-up, is whether aspirin reduces the future risk of [46 words unchanged] assess. Other analyses may be undertaken if new related, scientific questions arise. This phase of the study is planned to continue until at least [25 words unchanged] at out-patient clinics or A&E departments, and critical care episodes. These data will be are used to assess whether any benefits of aspirin and the omega-3 fish [30 words unchanged] between June 2005 and July 2011 (unless they withdrew consent). Any analyses will involve involves comparisons among all those previously allocated aspirin compared with those previously allocated placebo (dummy) and similarly between those previously allocated fish oil supplements versus those allocated placebo. [1 paragraph unchanged] Cancer data already held includes the period of the study as well [44 words unchanged] the recurrence of one that was diagnosed some years before. The analyses will include includes assessing the impact of the study treatments on newly diagnosed cancers and so it is important to be able to be sure that reported cancers are new. The source of the medical events of interest to these analyses will be is from the NHS Digital cancer and death registry data and also the Outpatient, and Admitted Patient Care data-sets. [2 paragraphs unchanged] Further to the on-study treatment analyses of the effects of aspirin on cognitive outcomes, 5-year results from analyses were published in 2022 and there are further analyses planned for 5 and 10 years after the end of the scheduled treatment period (i.e. including data up to 31.3.2022 and 31.3.2027). The source of the medical events of interest to the cognitive function analyses will be is from the following NHS Digital registry data-sets: Admitted Patient Care and the [7 words unchanged] The source of the information on heart attacks, strokes and heart failure will be is from the following NHS Digital Hospital data-sets: Admitted Patient Care, Outpatient, Accident and Emergency (being replaced by the Emergency Care data-set) and Critical Care. Additionally the ASCEND Eye sub-study, for which involvement was voluntary, will assess assesses the effects of taking aspirin and/or omega-3 fish oils on diabetic retinopathy [59 words unchanged] and the Diabetic Retinopathy Screening Service for Wales. However the eye sub-study will include includes data about eye conditions previously collected during the main treatment phase of [10 words unchanged] NHS Digital data-sets : the Accident and Emergency, Outpatient and Admitted Patient Care.] Care. [1 paragraph unchanged]

Processing activities

All of the Hospital Episode Statistics (HES), mortality and cancer registry data received, provided by NHS Digital under a previous iteration of this agreement, are stored and processed solely by the Clinical Trial Service Unit ASCEND [24 words unchanged] currently held linked to the participants NHS record at NHS Digital, under an existing a previous iteration of this agreement. Regular updates have been provided to the ASCEND study team with [33 words unchanged] it was possible for NHS Digital to use the cohort previously provided. The data received from NHS Digital are were transferred into the trial database by the person registered to receive data. [11 words unchanged] an ASCEND specific computer server. The identifiable NHS Digital data are stored in an encrypted TrueCrypt container separately on a secure server to which access is granted on a "need to know" basis i.e. [31 words unchanged] when the team member leaves ASCEND. The data supplied by NHS Digital is was linked back to the study database and it is critical that this [24 words unchanged] breach of Good Clinical Practice during any MHRA inspection. Therefore, the team use used the participant’s identifiable details so that multi-part checks can be applied. However, this has been was reassessed in the light of the main study treatment phase of the [19 words unchanged] participant contact details. However, given that no further contact with the participants is intended was required plus the need to be mindful of data minimisation principals, the decision has been was agreed to only receive the study ID and the NHS number and no other personal identifiers. As part of the study, participant-reported medical events are were recorded. These medical events allow allowed detailed assessment of both the safety and efficacy of the trial treatments and are were coded by the study medical team who require additional evidence (such as information about diagnoses as well as dates and nature of operations). This evidence is was largely provided from the datasets held by NHS Digital, providing details of [16 words unchanged] from the death certificate, as well as the participant’s GP. These data will also allow allowed the study medical team to identify medical events not previously reported. This robust process will ensure ensured the results are were reliable and unbiased and so complying with current regulations governing clinical trials. [34 words unchanged] if any further contact is required, then the practice code would suffice. Although the University of Oxford has minimised the supplied personal identifiers in the data requested, the remaining supplied data will still be was linked to the existing study database, which does store the identifying data. The study team is was required to do this as the study is still ongoing and has [24 words unchanged] Healthcare Products Regulatory Agency (MHRA) ensure these regulations are enforced via inspections. The ASCEND databases are stored on the WING server on the Linux platform, which is backed up according to the following schedule: The raw data files are stored securely with access limited to those receiving the files from the NHS registries and those who need to process the data in order to merge into the study-specific database. Full backups to tape are taken at the beginning of the month to tape; the January and July tape are taken out of rotation as archive copies. Differential backups are taken in the early hours of Monday to Thursday, with an incremental backup early on Friday. NDPH researchers are experienced in handling confidential and participant sensitive data and have appropriate training in information governance. The NDPH servers are protected against unauthorised external access by an appropriate strength firewall. At the end of the retention period, the ASCEND databases files are deleted from the virtual Linux server as part of the decommissioning process by the study team. Once decommissioned, it is deleted from the hypervisor and the virtual machine data store. The data store is built around an array of arrays, with the files split between many hard disks making retrieval of any meaningful data non-viable in the short term, with the available memory soon overwritten by other virtual machines. Access to ASCEND data (including patient identifiable data) is protected by the appropriate authentication procedures (unique user IDs and passwords). Authentication is only granted to personnel with a need to access the required data. Only personnel involved in the long-term follow-up study (processing and analysing data) will have access to the data under this Agreement. NDPH has a Corporate Level Security Policy that has been fully adopted by management and will apply fully to the long-term follow-up study. As the NHS registry data is stored on a segregated encrypted NAS device, no archive or backup copies are made throughout the lifecycle of the data. The NAS device is mirrored to ensure information integrity, with both devices subject to secure erase (O&O SafeErase) upon expiry. A certificate of Destruction will be issued. All information is stored securely by the University of Oxford and is kept confidential. No individuals will be identified in any study reports. The raw data files will be stored securely with access limited to those receiving the files from the NHS registries and those who need to process the data in order to merge into the study-specific database. [1 paragraph unchanged] Access to data is restricted via user identification provided by Active Directory Services (username and password). Servers that store data are located in secure buildings with authorised swipe card access only. Where required, specific Sensitive and Personal data sets are contained and isolated within a dedicated high compliance NAS array, with additional login (password) requirements. This allows selective destruction of both live data and backups upon request. [2 paragraphs unchanged] 2. the situation where an MHRA inspector, as described in the previous paragraph, inspector may see an individual record with a medical event or death recorded, which may have been supplied by NHS Digital. [3 paragraphs unchanged]

Expected output

The NDPH contributes widely to health policy, particularly in the area of vascular risk prevention. It contributes to debate with academic papers, conference participation, lectures to the public and advice to government (including NHS Digital). Examples of the impact of the work performed by NDPH up to 2021 is available from: https://results2021.ref.ac.uk/profiles/institutions/10007774 [1 paragraph unchanged] The main outputs from the ASCEND research will be so far have been in the form of scientific reports of the results of the trial. The results from the scheduled treatment phase were presented during 2018 at [36 words unchanged] the British Heart Foundation helped with promotion and dissemination of the results. The results are likely to be incorporated into an individual participant data meta-analysis of similar trials run by the Anti-Thrombotic Trialists’ Collaboration, which is coordinated by the University of Oxford. The outputs from the Long-term Follow-up long-term follow-up phase will be in the form of scientific reports and presentations at National and International Scientific meetings. Summaries of key findings will be placed on the The ASCEND study website and disseminated through Oxford University Communications channels. team intend to publish results in mid-2024. All outputs will consist of aggregate level data only with small numbers suppressed in line with HES analysis guide. The results of the long-term follow-up, particularly on the effects of aspirin on long-term cancer and dementia risk, and additional analyses on eye outcomes and heart failure, will be directly relevant to 400M people with diabetes worldwide. The study participants were sent a summary of the study results shortly after the main presentation in 2018 and the medical consultants representing the various NHS hospital trusts involved, will also be presented with a summary of the results and attended a specific meeting at which the results were discussed around the same time. This summary was also be available via the study-specific web-site along with a short video of the results (https://ascend.medsci.ox.ac.uk/). Methodological outputs from this research are intended to inform the design of future trials wishing to collect study outcomes from routinely collected healthcare data. In addition to this, the two pharmaceutical companies provided the medication and matching placebo for the study and some funds to cover the costs of packaging the treatment have an interest in seeing these treatments (aspirin and fish oils) properly evaluated in a large well run randomised trial. They received the same publicly available results. The pharmaceutical companies may wish to use tabulated data from the study to seek approval from the regulatory agencies for the marketing of these treatments to this group. The data provided would be aggregated with small numbers suppressed in line with the HES Analysis Guide. Summaries of key findings will be placed on the study website and disseminated through Oxford University Communications channels. If a submission is made to a regulatory agency, the Clinical Trial Service Unit at Oxford University would provide relevant information in the form of tabulations of numbers (for example: number of participants randomized, experiencing serious or other adverse events) with no individuals being identifiable in any submission. There would be considerable overlap between any tabulations provided to the companies and the published results, however the companies might ask for specific detail that were not considered necessary to publish. No participants will be identifiable in any information provided and the pharmaceutical companies do not have any influence in the research or the results. All data provided to the companies would be aggregate with small numbers suppressed in line with the HES Analysis Guide. The ASCEND study team will share outputs via the listed channels: • Study website (https://ascend.medsci.ox.ac.uk/) • Open lectures and talks • Posters • Press/media engagement and other public promotion of the research (e.g. via the Nuffield Department of Population Health website (https://www.ndph.ox.ac.uk/), or Twitter account (@oxford_ndph). All outputs will consist of aggregate level data only with small numbers suppressed in line with the HES analysis guide. The study participants were mailed a summary of the study results shortly after the main results presentation in 2018. The medical consultants representing the various NHS hospital trusts involved, were presented with a summary of the results and attended a specific meeting at which the results were discussed around the same time. This summary was also available via the study-specific web-site along with short videos of the results (https://ascend.medsci.ox.ac.uk/). In addition to this, the pharmaceutical companies that provided the medication and matching placebo for the study, plus some funds to cover the costs of packaging the treatment, received the same publicly available results. [1 paragraph unchanged] Outputs for the ASCEND trial so far includes multiple publications in peer-reviewed journals and presentations at International conferences. Information about the publications can be found on the ASCEND website (https://ascend.medsci.ox.ac.uk/) and the benefits are described in Section 6iii (Yielded Benefits).

Expected measurable benefits

The data collected from central NHS registries during the Long–Term Follow-up long-term follow-up phase of the study, will be used to assess whether any benefits or harms of aspirin observed within the trial, continue long-term or additional benefits or harms emerge during longer-term follow-up. It had been suggested that low-dose aspirin might protect against cancer cancer, and ASCEND provides one of the first opportunities to test this hypothesis. The analyses based on the study treatment phase showed no reduction in any cancers during the up to 9 a mean 7.3 years of follow-up available so far. follow-up.. However, the main focus of the analyses will be after longer-term post-trial long-term follow-up, when there will be much better power to detect plausible differences in cancer incidence between the arms due to larger numbers of events. Dementia and cognitive impairment present major health care and social burdens which [30 words unchanged] of dementia and a 20% increase in the rate of cognitive decline. Therefore, it is of particular importance to obtain randomized evidence of the effects of therapies for vascular prevention on cognitive decline among people with diabetes. Furthermore, the higher risks of cognitive decline and dementia among people with diabetes make them a potentially powerful population for investigating cognitive effects. Dementia or cognitive impairment has been captured from multiple sources in the ASCEND trial: during routine trial follow-up, participants were asked to record all serious illnesses and hospitalisations (including dementia or cognitive impairment); at final participant follow-up, participants were asked whether they had been referred to a memory clinic; additionally participants may state that they are unable to complete cognitive assessment because of cognitive impairment, or use of dementia drugs may be recorded; at final GP follow-up, dementia or cognitive impairment was specifically asked about. Linked electronic NHS central registry data, provides an additional source of evidence of dementia and indications of cognitive impairment. A study evaluating NHS Digital Hospital Episode Statistics (HES) data as a source of dementia diagnosis, concluded that there was good agreement with evidence from primary care records but a mean delay of 1.6 years between first mention of the condition in hospital admissions, compared to in primary care records. Therefore, diagnoses of dementia-related outcomes from HES data or from death certificates will be included up to 31 March 2019, a date 1.5-2 years beyond when cognitive assessment was sought (March-December 2017). Therefore, it is of particular importance to obtain randomized evidence of the effects of therapies for vascular prevention on cognitive decline among people with diabetes. Furthermore, the higher risks of cognitive decline and dementia among people with diabetes make them a potentially powerful population for investigating cognitive effects. The main ASCEND study includes long-term follow-up of the effects of aspirin on its main outcomes. In parallel, analyses dementia and cognitive impairment (identified through the trial follow-up procedures or in the HES and Civil Registrations data up to 31 March 2019 (to allow for the expected delay between the diagnosis of dementia and the recording of dementia in HES which is known to be 1.6 years on average) were published in 2022 and showed no effect of aspirin on dementia risk but longer follow-up is needed to reliably assess this question. Analyses of the effects of aspirin on cognitive outcomes are therefore planned for 5 and 10 years after the end of the scheduled treatment period (i.e. including data up to 31.3.2022 and 31.3.2027). period. [1 paragraph unchanged] Heart failure is a major cause of disability and people with diabetes are at increased risk of developing heart failure. The current work in confirming heart failure events identified from the trial follow-up procedures and the linked health records, will be used to assess the impact of aspirin on heart failure during the scheduled treatment period. This will also allow algorithm development to reliably identify heart failure events in the long-term follow-up to allow the assessment of the effects of aspirin on long-term heart failure risk.

Benefits reported

The following is a summary of the results from ASCEND so far. • Aspirin use prevented serious vascular events in persons who had diabetes and no evident cardiovascular disease at trial entry, but it also caused major bleeding events. The absolute benefits were largely counterbalanced by the bleeding hazard. • Among patients with diabetes without evidence of cardiovascular disease, there was no significant difference in the risk of serious vascular events between those who were assigned to receive n-3 fatty acid supplementation and those who were assigned to receive placebo. • Post hoc analyses of the ASCEND trial suggest that routinely collected hospital admission and death registry data in the UK could be used as the sole method of follow-up for myocardial infarction, ischemic stroke resulting in hospitalization, vascular death, and arterial revascularization in primary prevention cardiovascular trials, without the need for verification by clinical adjudication. • The Cognitive Function and Dementia analysis (presented at the American Heart Association (AHA) in November 2021, and published in the European Heart Journal in June 2022), found that there was no statistically significant effect on dementia outcomes in participants taking aspirin and that trials with larger numbers of incident dementia cases are needed to assess whether there are any benefits after 5-7 years of aspirin use. It was also determined that in the UK, routine electronic health data provided a cost-effective means of assessing the impact of vascular preventive therapies on dementia. Results on the effects of omega-3 fatty acids were presented at the European Society of Cardiology in 2022. This analysis showed that omega-3 fatty acid supplementation had not detectable effect on dementia or cognitive function. Evidence from large trials now suggests that omega-3 FA supplementation may have a dose-related protective effect on coronary events but it has been suggested that the risk of atrial fibrillation (AF) is increased. Analysis of atrial fibrillation identified from ASCEND trial follow-up procedures and linked health records did not show and increase in AF risk of omega-3 FA supplementation. However, systematic reporting of arrhythmia outcomes in existing and future trials is required to clarify whether supplementation also has any adverse effects on nonfatal arrhythmias. The following points explain why these results are important and what the wider benefits are. 1. Impact on patients [1 paragraph unchanged] The results of ASCEND now provide much needed clarity. The findings are likely to be widely incorporated into future guidelines for the prevention of cardiovascular events in people with diabetes both nationally and internationally. The World Health Organisation statistics detail 422 million people with diabetes in 2014, a rise of almost 400% since 1980. The prevalence of diabetes in the UK and global population is expected to continue to rise. 2. Impact on healthcare providers in the UK and Worldwide The main results of the ASCEND trial now provide much needed clarity. The findings incorporated into updated guidelines for the prevention of cardiovascular disease both nationally and internationally (including the 2022 US Preventive Services Task Force (USPSTF) Aspirin Use to Prevent Cardiovascular Disease and the 2021 ESC Guidelines on cardiovascular disease prevention in clinical practice) and for the management of diabetes (including the 2019 ESC Guidelines on diabetes, pre-diabetes, and cardiovascular diseases developed in collaboration with the EASD). The World Health Organisation statistics detail 422 million people with diabetes in 2014, a rise of almost 400% since 1980. The prevalence of diabetes in the UK and global population is expected to continue to rise. 3. Impact on trial design and cost Recent (2021) ASCEND analyses found that in the UK, routine electronic health data provided a cost-effective means of assessing the impact of vascular preventive therapies on dementia. They also suggest that routinely collected hospital admission and death registry data in the UK could be used as the sole method of follow-up for myocardial infarction, ischemic stroke resulting in hospitalization, vascular death, and arterial revascularization in primary prevention cardiovascular trials, without the need for verification by clinical adjudication. This could have far-reaching implications for future trial design and minimising research costs.

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

Datasets: + National Diabetes Audit · − HES-ID to MPS-ID HES Accident and Emergency; − HES-ID to MPS-ID HES Admitted Patient Care; − HES-ID to MPS-ID HES Outpatients

Changed only in punctuation, spacing or capitalisation: Processing activities.

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

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

Datasets: + HES-ID to MPS-ID HES Accident and Emergency; + HES-ID to MPS-ID HES Admitted Patient Care; + HES-ID to MPS-ID HES Outpatients

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

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

Datasets: + Cancer Registration Data; + Civil Registrations of Death; + Demographics

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