SABRE: Southall and Brent Revisited - S251 participants not cancer notifiable

The University College London requires access to NHS England data for the purpose of the following research project: SABRE: Southall and Brent Revisited

SABRE is a population-based cohort study, in its current 25 year follow-up phase, the study focuses on identifying and understanding the underlying reasons for ethnic group and sex differences in cardiometabolic disease and in physical, psychological and cognitive function in older age.

Specific questions for the 25 year follow-up study are:

1. How large are ethnic /sex differences in cardiac function, cognitive function and hippocampal volumes in older age?

2. To what extent do cardiac function, cognitive function and hippocampal volumes change over a 5 year period in each ethnic group?

3. Which risk factors measured in mid-life and in early old age are most strongly associated with current cardiac and cognitive function and hippocampal volumes and with 5 year changes in these parameters? Can these risk factors explain ethnic differences in cardiac and cognitive function?

4. How large are gender differences in current disorders of cardiac and cognitive function and in their associations with current risk factors?

5. Do ethnic differences in incident cardiometabolic disorders persist into older age?

6. Which risk factors or risk factor profiles measured in mid-life and early old age are most strongly associated with incident cardiometabolic disorders and which best explain ethnic differences in incidence?

The following NHS England data will be accessed:

Hospital Episode Statistics

o Admitted Patient Care – necessary to identify incident cardiometabolic events (in particular coronary heart disease, heart failure, stroke, dementia, diabetes), and other events which may affect physical and cognitive function.

• Civil Registration Mortality – necessary to conduct survival analysis (time to event) and understand cause of death in relation the previously collected clinic data.

• Demographics – necessary to invite participants to future data collections, provide feedback or notify them of study-related findings

The level of the data will be:

• Identifiable – necessary because although the applicant will pseudonymise the data before analysis, the identifiers are retained separately within the Data Safe Haven solely for administration purposes.

The data will be minimised as follows:

• Limited to a study cohort identified by UCL- for this subset the cohort size is 69 (Participants have not given informed consent and previously specifically declined consent for the study to obtain details of their cancer registrations. Type 2 Patient Objections will be applied and no cancer notifications will be released for this subset.)

UCL is the research sponsor and the controller as the organisation responsible for ensuring that the data will only be processed for the purpose described above.

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

Article 6(1)(e) - processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller;

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

Article 9(2)(j) - processing is necessary for scientific and research purposes, subject to appropriate safeguards. The aim of this study - understanding of underlying causes of difference in physical and cognitive function in people of different ethnicities is in the public interest because it will ultimately lead to appropriate preventive strategies and treatments at different stages of life.

The funding is provided by Medical Research Council.

Amazon Web Services (AWS) is a processor acting under the instructions of UCL. AWS’ role is limited to secure back-up of data stored in UCL’s Data Safe Haven.

UCL uses offsite data centre services provided by VIRTUS data centre. VIRTUS does not have access to the data.

Data will be accessed by:

• Undergraduate, Masters or PhD students affiliated with UCL. Any student working with the data held under this Agreement must have completed relevant data protection and confidentiality training and are subject to UCL’s policies on data protection and confidentiality. Any students accessing the data will do so under the supervision of a substantive employee of UCL. UCL would be responsible and liable for any work carried out by students. These students would only work on the data for the purposes described in this Agreement.

The SABRE cohort are a tri-ethnic ageing population. Work has previously been undertaken to invite participants to Advisory Committees. Most recently, UCL has worked in conjunction with the UK Longitudinal Linkage Collaboration (UK LLC) and its PPIE group to draft appropriate privacy notice materials

The expected outputs of the processing are listed below. The outputs will be communicated to relevant recipients through the following dissemination channels:

• Peer reviewed scientific journals, predominantly related to epidemiology, cardiovascular and metabolic disorders, cognitive, physical and psychological function.

• Participant and GP feedback sessions

• Relevant professional conference attendances with abstract presentations: include British Society of Cardiology, European Society of Cardiology, European Association for the Study of Diabetes, British and Irish Hypertension Society, Artery, Physiology Society, American College of Cardiology, Diabetes UK

• Public events – Presentation at the BHF 60th birthday parliamentary reception.

• Posters displayed at numerous conference described above

• Press/media engagement – Podcast on “How the pandemic highlighted BAME inequalities”. Guardian article on “Why are people from BAME groups dying disproportionately of Covid-19?”.

• Public promotion of the research - in order to recruit SABRE participants for the clinic visit, radio interviews and newspaper articles promote the SABRE study.

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

Target dates for production and dissemination of the outputs are as follows:

Complete clinical data preparation and ongoing analyses and write up for publication of HES, mortality and cancer data during 2022-2025, analyses will include associations between phenotypic and questionnaire data collected at previous visits and clinical outcomes (Hospital admissions, cancer registrations and deaths) with particular focus on ethnicity and risk factor associations. However, it is anticipated that further analyses will continue beyond 2025.

Participants were asked to complete questionnaires regarding their health and the impact of the COVID-19 pandemic on them and their families. These data will be linked to hospital admissions and mortality data and these outcome data together with the questionnaire data will be related to earlier waves of data collection in order to study associations and underlying reasons for ethnic differences in health and lifestyle during the pandemic. Target date: 2022-25

end

The research has enabled improved ascertainment of incident coronary heart disease and stroke events and resulted in many publications in high impact journals relating these outcomes to risk factors measured in mid-life (ages 40-70 at baseline).

This is a mechanistic research study in a population-based sample, looking at underlying reasons for disparities in health and function in older age, with a particular focus on ethnic minorities. In an ageing general population and with over 16% of the population of England being from ethnic minority backgrounds (most of South Asian or Black African or African Caribbean origins), the findings impact on a very large number of people. As an example, our findings regarding the extraordinarily high incidence of new onset diabetes in South Asians and African Caribbeans, even in old age (Diabetes Care, 2013), were widely reported in the national press and on national TV/radio and led to a widespread acknowledgement that this is a much greater problem than anticipated. Findings regarding the associations between different levels of obesity and future development of type 2 diabetes were informative in NICE’s determination of BMI levels for prevention of type 2 diabetes in different ethnic minority groups. Many of UCL’s findings are published and widely cited in peer reviewed journals and thus UCL inform future research questions and add to the global understanding of ethnic differences in cardiometabolic disease (globally 18 million deaths were due to cardiovascular disease and 1.5 million due to diabetes in 2019, while numbers living with diabetes continue to rise- 537 million adults in 2021). UCL’s study is not designed to measure impact -but given the heavy burden of diabetes and cardiovascular disease in terms of length and quality of life and its particular effect on people of ethnic minority backgrounds, increased awareness and understanding can only lead to better preventive and therapeutic strategies in both primary and secondary care settings.

Due to their age and pace of implementing intervention, the SABRE study participants are unlikely to directly benefit from this research, but through UCL’s dissemination strategy, UCL hope the research will be used in policy, which can be implemented by third parties and used to benefit successive generations. For example, SABRE are part of the Longitudinal Health and Wellbeing National Core Study (LH&W NCS) and were included in the SAGE reports provided to Government.

This Agreement only permits the Data Controller (University of London) to retain data that was previously supplied under past iterations of this Agreement. This is an interim measure and therefore under no circumstances can the data specified in this Agreement be processed.

The following provides the objectives of the study for which the data was disseminated for

University College London (UCL) requires notifications of mortality and cancer registrations and linked HES data for its study cohort for use in the Medical Research Project: SABRE (Southall And Brent Revisited). This is a population-based cohort study, conducted at University College London, funded by the British Heart Foundation in its current 25 year follow-up phase. It is unique as a long-standing tri-ethnic cohort consisting of people of European descent and first generation migrants of South Asian or African Caribbean descent. This is an academic research study focusing on identifying and understanding the underlying reasons for ethnic group and sex differences in cardiometabolic disease and in physical, psychological and cognitive function in older age.

Specific questions for the 25 year follow-up study are:

1. How large are ethnic /sex differences in cardiac function, cognitive function and hippocampal volumes in older age?

2. To what extent do cardiac function, cognitive function and hippocampal volumes change over a 5 year period in each ethnic group?

3. Which risk factors measured in mid-life and in early old age are most strongly associated with current cardiac and cognitive function and hippocampal volumes and with 5 year changes in these parameters? Can these risk factors explain ethnic differences in cardiac and cognitive function?

4. How large are gender differences in current disorders of cardiac and cognitive function and in their associations with current risk factors?

5. Do ethnic differences in incident cardiometabolic disorders persist into older age?

6. Which risk factors or risk factor profiles measured in mid-life and early old age are most strongly associated with incident cardiometabolic disorders and which best explain ethnic differences in incidence?

The study has previously utilised the List Cleaning service from time to time when in active follow-up in order to ensure that the correct participant addresses are used in order to contact participants. Use of this service has helped the study to avoid trying to contact deceased participants. The List Cleaning outputs were used to update the administration database (held separately from other data within the UCL data safe haven) so that UCL could write to as many participants as possible inviting them to complete questionnaires or come into the UCL clinic for a detailed investigation. Under this Data Sharing Agreement, UCL may retain List Cleaning outputs received previously but is not permitted to make further use of the List Cleaning service.

Linked HES APC data from NHS Digital has previously been disseminated to the SABRE study in order to identify incident cardiometabolic events (in particular coronary heart disease, heart failure, stroke, dementia, diabetes), and other events which may affect physical and cognitive function, which have occurred during the follow-up period. Details of all hospital episodes involving the cohort (not limited to the previously stated conditions) were required to address key study objectives with regard to physical and cognitive function in older age in association with current and mid-life risk factors.

Analysis considers any and all potential contributing factors.

These events will supplement information provided by participant self-report at 20 and 25 years, from primary care medical record review conducted during the 20 year follow-up and from cancer and mortality flagging, together with detailed clinical measurements made at the SABRE clinics at baseline, 20 and 25 year follow-up. The data will be used to analyse risk factors measured in mid- and later life in association with these incident events in order to build on current understanding of causal mechanisms.

Data from 1989 to the present are required because participants underwent detailed examinations at baseline (1989-91) and the aim is to follow this cohort through their experiences since to understand what happened in later life and relate that to the baseline. This will enable UCL to gain as complete as possible a picture of hospital admissions, and hence incident events, over the entire cohort follow-up. Data from the entire study period are crucial for determining age of onset of events, as well as the extent and nature of ill-health from mid to later life, and for relating these to current and mid-life cardiometabolic and other risk factors and how these influence the key study outcomes of physical and cognitive function in older life in each of the three ethnic groups.

This request is to extend the term of the data sharing agreement to complete the complex analyses which will be required to integrate outcomes derived from HES data with mortality and cancer registration data and the very detailed phenotypic data collected during the course of the study period since 1989. In addition, this request is also to continue to receive mortality and cancer registration data, as these are, sadly, key outcomes which can be related to risk factors identified in earlier life. It is also important that notifications of death are received. No raw HES data from NHS Digital are being disseminated under this new agreement.

There are a number of reasons why UCL had not sought to consent all study participants for study health data linkage:

- At the start of the third wave of follow-up (2014-2018) UCL conducted a pilot study to assess whether it was feasible to obtain consent for data linkage using a postal approach to existing participants (UCL sent out a covering letter, participant information, consent form including an item on data linkage, health and lifestyle questionnaire and reply paid return envelope). At that time there were 3400 surviving participants, aged between 65 and 98 years. UCL approached 773 participants who had visited the clinic at the previous phase of follow-up. Of these people, who UCL had considered to be highly motivated participants, UCL received positive responses from 65%, negative responses from 9%, no response from 17%, address incorrect (1%) and a further 10% who requested time to consider or who asked to be contacted again in the future. Of the latter group, although UCL know that they have received the information, very few have been contactable with regard to the study, bringing the total estimated non-response rate in this group to approximately 25%. This non-response rate was unexpectedly high and likely to be related to ill health, particularly cognitive impairment.

- With a 17-25% non-response rate already demonstrated in the pilot study of the highly motivated group, UCL had reasonable grounds for assuming a much higher non-response rate in the rest of the cohort (2070 survivors), with undoubted risk to the scientific integrity of the follow-up study. UCL had consulted with the Confidentiality Advisory Group (CAG) and followed their advice regarding application for section 251 support. With CAG agreement UCL modified the approach to existing participants by removing the item on consent to data linkage from the postal consent form, while adding an item at the start of the questionnaire that enabled participants to opt-out of data linkage even if they chose not to come in to the clinic. 21 people chose to opt out of data linkage in this way. UCL then placed an additional item requesting consent for data linkage in the clinic consent form for responders who had agreed to visit UCL for clinical follow-up.

- Loss to follow-up is an important problem in older cohorts. People who do not respond, or are lost to follow-up in other ways, may be different from those who participate - often they are more unwell than responders. This is likely to lead to bias and to jeopardise the scientific value of the findings and grossly bias public health messages and potential interventions. Given the unique experiences of the first generation migrants in this cohort, which may be vital in informing health service strategies in an increasingly migratory world, UCL believe that it would be a waste of the data and commitment to the study in earlier years if UCL are unable to reach sound scientific conclusions due to loss to follow-up. This is where linkage to health service records can minimise the effects of loss to follow-up due to non-response.

This agreement is for university research, the lawful basis for processing data is GDPR 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’. Also referred to as ‘Public Task’. As the research involves health data, which is included in the definition of special categories of personal data, it requires an additional condition for processing. Based on guidance, for health research this is article 9(2)(j), which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. The aim of this study - understanding of underlying causes of difference in physical and cognitive function in people of different ethnicities is in the public interest, because it will ultimately lead to appropriate preventive strategies and treatments at different stages of life.

This Agreement only permits the Data Controller (University of London) to retain data that was previously supplied under past iterations of this Agreement. This is an interim measure and therefore under no circumstances can the data specified in this Agreement be processed.

The following provides previous and intended future outputs (subject to a future formal agreement) of the study for which the data was disseminated for

Study findings will continue to be published in peer-reviewed scientific journals, predominantly related to epidemiology, cardiovascular and metabolic disorders, cognitive, physical and psychological function, but also including more generic journals such as the BMJ, reflecting the increasing focus on overall health and function in older age. Publications will contain only aggregate level data without local identifiers and with suppression of small numbers in line with HES analysis guide.

Publications to date are listed on the study website: www.sabrestudy.org. All publications since 2008 are open-access. The audience is expected to consist mainly of academic researchers and clinicians.

The study team at UCL will further disseminate findings via participant and GP feedback sessions; newsletters, and the study website.

At the end of the current funding period (extended to December 2020) a report will be submitted to the funders (the British Heart Foundation) summarising findings. This may be published on their website.

Update (May2019) : Since the previous data sharing agreement was approved in 2018, the study has been in active clinical follow-up (ended January 2019) and data preparation is not yet complete. However, a number of new research papers based on this cohort have been submitted for publication or published during the period, some of which utilise earlier cardiovascular outcome data, some examples:

Targets:

Complete clinical data preparation and processing of HES, and mortality data during 2019-20, ready for analysis and write up.

Outputs are expected to include dissemination of key findings to participants and their GPs during 2019-2020.

Relevant professional conference attendances with abstract presentations: include British Society of Cardiology, European Society of Cardiology, European Association for the Study of Diabetes, British and Irish Hypertension Society, Artery. 2019-2020 and ongoing. Note: All data for abstracts/presentations will be at aggregate level with suppression of small numbers in line with HES analysis guide.

Papers to be prepared for submission to peer-reviewed journals during 2019-2020 will include studies of risk factors over 30 years of follow-up in association with:

- ethnicity and all-cause and cardiovascular disease-related mortality,

- ethnicity and acute and chronic cardiovascular disease , including heart failure

- ethnicity and diabetes incidence and complications

- ethnicity and mental health

Earlier publications which used NHS-Digital data:

The research has enabled improved ascertainment of incident coronary heart disease and stroke events and resulted in many publications in high impact journals relating these outcomes to risk factors measured in mid-life (ages 40-70 at baseline). Findings include:

- Diabetes incidence in older British South Asians and African Caribbeans remains at least 2-fold even at age 80

years compared with British Europeans. The ethnic differentials in women were largely explained by midlife truncal obesity and insulin resistance, but the study was unable to explain the ethnic difference in men.

-Obesity cut-points of 24 and 27 kg/m2 in South Asians and African Caribbeans respectively were equivalent to a body mass index of 30kg/m2 in Europeans in terms of diabetes risk; these latter analyses contributed to recent NICE guidelines for prevention of diabetes. Diabetes was also found to be more toxic in terms of stroke risk in the ethnic minorities.

-Widely used tools (Framingham and QRISK2) for estimation of cardiovascular disease risk were found to be less precise in South Asians and African Caribbeans (particularly women), while a selection of 3 metabolic markers measured by NMR spectroscopy was found to be strongly predictive of cardiovascular risk regardless of ethnicity.

- Lack of adherence to four combined health behaviours was associated with a 2 to 3-fold increased risk of incident CVD in Europeans and South Asians. A substantial population impact in the South Asian group indicates important potential for disease prevention in this high-risk group by adherence to healthy behaviours.

- The study has also found marked ethnic differences in associations between blood pressure parameters and

stroke and concluded that undue focus on systolic blood pressure for risk prediction, and current age and

treatment thresholds may be inappropriate for individuals of South Asian ancestry.

- detrimental associations were reported between air pollution (particulate measures) and cardiovascular

disease mortality in both the SABRE and Whitehall cohorts.

-Results highlighted ethnic differences in associations between prediabetes in midlife and later development of coronary heart disease and stroke.

- Excess coronary heart disease incidence in South Asians, with lower incidence in African Caribbeans compared with Europeans continues and elevated risk of stroke was confirmed in both ethnic minority groups. Measured baseline metabolic risk factors could not explain the ethnic group differences. Of particular concern was a much stronger association between diabetes and stroke risk in both ethnic minority groups compared with Europeans with diabetes- an association which is the subject of ongoing study in the SABRE cohort. HES data contributed importantly to the identification of incident coronary and stroke events reported in these analyses.

Selection of manuscripts recently accepted for publication:

-Baseline (1989-1991) vascular risk factors explained the observed ethnic variation in cardiovascular disease recurrence and long-term mortality, with a relative improvement in survival of minority ethnic groups over time.

Vyas MV, Chaturvedi N, Hughes AD, Marmot M, Tillin T. Cardiovascular disease recurrence and long-term mortality in a tri-ethnic British cohort. Heart. 2020 Oct 16;107(12):996–1002. doi: 10.1136/heartjnl-2020-317641. Epub ahead of print. PMID: 33067326; PMCID: PMC8165149.

-Markedly stronger associations were observed between family history and clinical CHD in South Asians, and a similar trend for subclinical CHD. Early preventive and therapeutic interventions are particularly important in South Asians with a family history of CHD.

Wang J, Tillin T, Hughes AD, Chaturvedi N. Associations between family history and coronary artery calcium and coronary heart disease in British Europeans and South Asians. Int J Cardiol. 2020 Feb 1;300:39-42. doi: 10.1016/j.ijcard.2019.07.101. Epub 2019 Aug 1. PMID: 31400886; PMCID: PMC6970219.

- Urine albumin:creatinine ratio may have greater utility than estimated glomerular filtration rates in CVD risk prediction in South Asians.

Eastwood S, Chaturvedi N, Sattar N, Welsh P, I, Hughes A, D, Tillin T: Impact of Kidney Function on Cardiovascular Risk and Mortality: A Comparison of South Asian and European Cohorts. Am J Nephrol 2019;50:425-433. doi: 10.1159/000503873.

-Novel results provide support for ethnic differences in sleep quality and mortality, and may have implications for precision medicine.

Garfield V, Joshi R, Garcia-Hernandez J, Tillin T, Chaturvedi N. The relationship between sleep quality and all-cause, CVD and cancer mortality: the Southall and Brent REvisited study (SABRE). Sleep Med. 2019 Aug;60:230-235. doi: 10.1016/j.sleep.2019.03.012. Epub 2019 Mar 27. PMID: 31182326; PMCID: PMC6626293.

-Results support a detrimental long-term effect for air pollutants on cardiovascular mortality.

Dehbi HM, Blangiardo M, Gulliver J, Fecht D, de Hoogh K, Al-Kanaani Z, Tillin T, Hardy R, Chaturvedi N, Hansell AL. Air pollution and cardiovascular mortality with over 25years follow-up: A combined analysis of two British cohorts. Environ Int. 2017 Feb;99:275-281. doi: 10.1016/j.envint.2016.12.004. Epub 2016 Dec 7.

At the end of the recent funding period (extended to June 2021) a report was submitted to the funders (the British Heart Foundation) summarising findings. Some details are published on their website. (https://www.bhf.org.uk/what-we-do/our-research/research-successes/ethnicity-and-heart-disease )

Brief summary of findings to date: ethnic differentials in diabetes and coronary heart disease incidence continue even into older age and at lower levels of obesity compared with people of European descent. Not only is diabetes more frequent, but it may be more strongly associated with stroke risk in South Asian and African Caribbean populations. Family history of cardiovascular disease and mid-life diastolic as well as systolic blood pressure were more strongly associated with cardiovascular disease in South Asians. Widely used risk scores for cardiovascular disease may be less accurate in ethnic minority groups. Earlier and improved methods of monitoring and preventive strategies are needed for diabetes and its consequences in ethnic minority populations.

This is a mechanistic research study in a population-based sample, looking at underlying reasons for disparities in health and function in older age, with a particular focus on ethnic minorities. In an ageing general population and with over 16% of the population of England being from ethnic minority backgrounds (most of South Asian or Black African or African Caribbean origins), the findings impact on a very large number of people. As an example, our findings regarding the extraordinarily high incidence of new onset diabetes in South Asians and African Caribbeans, even in old age (Diabetes Care, 2013), were widely reported in the national press and on national TV/radio and led to a widespread acknowledgement that this is a much greater problem than anticipated. Findings regarding the associations between different levels of obesity and future development of type 2 diabetes were informative in NICE’s determination of BMI levels for prevention of type 2 diabetes in different ethnic minority groups. Many of UCL’s findings are published and widely cited in peer reviewed journals and thus UCL inform future research questions and add to the global understanding of ethnic differences in cardiometabolic disease (globally 18 million deaths were due to cardiovascular disease and 1.5 million due to diabetes in 2019, while numbers living with diabetes continue to rise- 537 million adults in 2021). UCL’s study is not designed to measure impact -but given the heavy burden of diabetes and cardiovascular disease in terms of length and quality of life and its particular effect on people of ethnic minority backgrounds, increased awareness and understanding can only lead to better preventive and therapeutic strategies in both primary and secondary care settings.

This Agreement only permits the Data Controller (University of London) to retain data that was previously supplied under past iterations of this Agreement. This is an interim measure and therefore under no circumstances can the data specified in this Agreement be processed.

The following provides the objectives of the study for which the data was disseminated for

University College London (UCL) requires notifications of mortality and cancer registrations and linked HES data for its study cohort for use in the Medical Research Project: SABRE (Southall And Brent Revisited). This is a population-based cohort study, conducted at University College London, funded by the British Heart Foundation in its current 25 year follow-up phase. It is unique as a long-standing tri-ethnic cohort consisting of people of European descent and first generation migrants of South Asian or African Caribbean descent. This is an academic research study focusing on identifying and understanding the underlying reasons for ethnic group and sex differences in cardiometabolic disease and in physical, psychological and cognitive function in older age.

Specific questions for the 25 year follow-up study are:

1. How large are ethnic /sex differences in cardiac function, cognitive function and hippocampal volumes in older age?

2. To what extent do cardiac function, cognitive function and hippocampal volumes change over a 5 year period in each ethnic group?

3. Which risk factors measured in mid-life and in early old age are most strongly associated with current cardiac and cognitive function and hippocampal volumes and with 5 year changes in these parameters? Can these risk factors explain ethnic differences in cardiac and cognitive function?

4. How large are gender differences in current disorders of cardiac and cognitive function and in their associations with current risk factors?

5. Do ethnic differences in incident cardiometabolic disorders persist into older age?

6. Which risk factors or risk factor profiles measured in mid-life and early old age are most strongly associated with incident cardiometabolic disorders and which best explain ethnic differences in incidence?

The study has previously utilised the List Cleaning service from time to time when in active follow-up in order to ensure that the correct participant addresses are used in order to contact participants. Use of this service has helped the study to avoid trying to contact deceased participants. The List Cleaning outputs were used to update the administration database (held separately from other data within the UCL data safe haven) so that UCL could write to as many participants as possible inviting them to complete questionnaires or come into the UCL clinic for a detailed investigation. Under this Data Sharing Agreement, UCL may retain List Cleaning outputs received previously but is not permitted to make further use of the List Cleaning service.

Linked HES APC data from NHS Digital has previously been disseminated to the SABRE study in order to identify incident cardiometabolic events (in particular coronary heart disease, heart failure, stroke, dementia, diabetes), and other events which may affect physical and cognitive function, which have occurred during the follow-up period. Details of all hospital episodes involving the cohort (not limited to the previously stated conditions) were required to address key study objectives with regard to physical and cognitive function in older age in association with current and mid-life risk factors.

Analysis considers any and all potential contributing factors.

These events will supplement information provided by participant self-report at 20 and 25 years, from primary care medical record review conducted during the 20 year follow-up and from cancer and mortality flagging, together with detailed clinical measurements made at the SABRE clinics at baseline, 20 and 25 year follow-up. The data will be used to analyse risk factors measured in mid- and later life in association with these incident events in order to build on current understanding of causal mechanisms.

Data from 1989 to the present are required because participants underwent detailed examinations at baseline (1989-91) and the aim is to follow this cohort through their experiences since to understand what happened in later life and relate that to the baseline. This will enable UCL to gain as complete as possible a picture of hospital admissions, and hence incident events, over the entire cohort follow-up. Data from the entire study period are crucial for determining age of onset of events, as well as the extent and nature of ill-health from mid to later life, and for relating these to current and mid-life cardiometabolic and other risk factors and how these influence the key study outcomes of physical and cognitive function in older life in each of the three ethnic groups.

This request is to extend the term of the data sharing agreement to complete the complex analyses which will be required to integrate outcomes derived from HES data with mortality and cancer registration data and the very detailed phenotypic data collected during the course of the study period since 1989. In addition, this request is also to continue to receive mortality and cancer registration data, as these are, sadly, key outcomes which can be related to risk factors identified in earlier life. It is also important that notifications of death are received. No raw HES data from NHS Digital are being disseminated under this new agreement.

There are a number of reasons why UCL had not sought to consent all study participants for study health data linkage:

- At the start of the third wave of follow-up (2014-2018) UCL conducted a pilot study to assess whether it was feasible to obtain consent for data linkage using a postal approach to existing participants (UCL sent out a covering letter, participant information, consent form including an item on data linkage, health and lifestyle questionnaire and reply paid return envelope). At that time there were 3400 surviving participants, aged between 65 and 98 years. UCL approached 773 participants who had visited the clinic at the previous phase of follow-up. Of these people, who UCL had considered to be highly motivated participants, UCL received positive responses from 65%, negative responses from 9%, no response from 17%, address incorrect (1%) and a further 10% who requested time to consider or who asked to be contacted again in the future. Of the latter group, although UCL know that they have received the information, very few have been contactable with regard to the study, bringing the total estimated non-response rate in this group to approximately 25%. This non-response rate was unexpectedly high and likely to be related to ill health, particularly cognitive impairment.

- With a 17-25% non-response rate already demonstrated in the pilot study of the highly motivated group, UCL had reasonable grounds for assuming a much higher non-response rate in the rest of the cohort (2070 survivors), with undoubted risk to the scientific integrity of the follow-up study. UCL had consulted with the Confidentiality Advisory Group (CAG) and followed their advice regarding application for section 251 support. With CAG agreement UCL modified the approach to existing participants by removing the item on consent to data linkage from the postal consent form, while adding an item at the start of the questionnaire that enabled participants to opt-out of data linkage even if they chose not to come in to the clinic. 21 people chose to opt out of data linkage in this way. UCL then placed an additional item requesting consent for data linkage in the clinic consent form for responders who had agreed to visit UCL for clinical follow-up.

- Loss to follow-up is an important problem in older cohorts. People who do not respond, or are lost to follow-up in other ways, may be different from those who participate - often they are more unwell than responders. This is likely to lead to bias and to jeopardise the scientific value of the findings and grossly bias public health messages and potential interventions. Given the unique experiences of the first generation migrants in this cohort, which may be vital in informing health service strategies in an increasingly migratory world, UCL believe that it would be a waste of the data and commitment to the study in earlier years if UCL are unable to reach sound scientific conclusions due to loss to follow-up. This is where linkage to health service records can minimise the effects of loss to follow-up due to non-response.

This agreement is for university research, the lawful basis for processing data is GDPR 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’. Also referred to as ‘Public Task’. As the research involves health data, which is included in the definition of special categories of personal data, it requires an additional condition for processing. Based on guidance, for health research this is article 9(2)(j), which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. The aim of this study - understanding of underlying causes of difference in physical and cognitive function in people of different ethnicities is in the public interest, because it will ultimately lead to appropriate preventive strategies and treatments at different stages of life.

This Agreement only permits the Data Controller (University of London) to retain data that was previously supplied under past iterations of this Agreement. This is an interim measure and therefore under no circumstances can the data specified in this Agreement be processed.

The following provides previous and intended future outputs (subject to a future formal agreement) of the study for which the data was disseminated for

Study findings will continue to be published in peer-reviewed scientific journals, predominantly related to epidemiology, cardiovascular and metabolic disorders, cognitive, physical and psychological function, but also including more generic journals such as the BMJ, reflecting the increasing focus on overall health and function in older age. Publications will contain only aggregate level data without local identifiers and with suppression of small numbers in line with HES analysis guide.

Publications to date are listed on the study website: www.sabrestudy.org. All publications since 2008 are open-access. The audience is expected to consist mainly of academic researchers and clinicians.

The study team at UCL will further disseminate findings via participant and GP feedback sessions; newsletters, and the study website.

At the end of the current funding period (extended to December 2020) a report will be submitted to the funders (the British Heart Foundation) summarising findings. This may be published on their website.

Update (May2019) : Since the previous data sharing agreement was approved in 2018, the study has been in active clinical follow-up (ended January 2019) and data preparation is not yet complete. However, a number of new research papers based on this cohort have been submitted for publication or published during the period, some of which utilise earlier cardiovascular outcome data, some examples:

Targets:

Complete clinical data preparation and processing of HES, and mortality data during 2019-20, ready for analysis and write up.

Outputs are expected to include dissemination of key findings to participants and their GPs during 2019-2020.

Relevant professional conference attendances with abstract presentations: include British Society of Cardiology, European Society of Cardiology, European Association for the Study of Diabetes, British and Irish Hypertension Society, Artery. 2019-2020 and ongoing. Note: All data for abstracts/presentations will be at aggregate level with suppression of small numbers in line with HES analysis guide.

Papers to be prepared for submission to peer-reviewed journals during 2019-2020 will include studies of risk factors over 30 years of follow-up in association with:

- ethnicity and all-cause and cardiovascular disease-related mortality,

- ethnicity and acute and chronic cardiovascular disease , including heart failure

- ethnicity and diabetes incidence and complications

- ethnicity and mental health

Earlier publications which used NHS-Digital data:

The research has enabled improved ascertainment of incident coronary heart disease and stroke events and resulted in many publications in high impact journals relating these outcomes to risk factors measured in mid-life (ages 40-70 at baseline). Findings include:

- Diabetes incidence in older British South Asians and African Caribbeans remains at least 2-fold even at age 80

years compared with British Europeans. The ethnic differentials in women were largely explained by midlife truncal obesity and insulin resistance, but the study was unable to explain the ethnic difference in men.

-Obesity cut-points of 24 and 27 kg/m2 in South Asians and African Caribbeans respectively were equivalent to a body mass index of 30kg/m2 in Europeans in terms of diabetes risk; these latter analyses contributed to recent NICE guidelines for prevention of diabetes. Diabetes was also found to be more toxic in terms of stroke risk in the ethnic minorities.

-Widely used tools (Framingham and QRISK2) for estimation of cardiovascular disease risk were found to be less precise in South Asians and African Caribbeans (particularly women), while a selection of 3 metabolic markers measured by NMR spectroscopy was found to be strongly predictive of cardiovascular risk regardless of ethnicity.

- Lack of adherence to four combined health behaviours was associated with a 2 to 3-fold increased risk of incident CVD in Europeans and South Asians. A substantial population impact in the South Asian group indicates important potential for disease prevention in this high-risk group by adherence to healthy behaviours.

- The study has also found marked ethnic differences in associations between blood pressure parameters and

stroke and concluded that undue focus on systolic blood pressure for risk prediction, and current age and

treatment thresholds may be inappropriate for individuals of South Asian ancestry.

- detrimental associations were reported between air pollution (particulate measures) and cardiovascular

disease mortality in both the SABRE and Whitehall cohorts.

-Results highlighted ethnic differences in associations between prediabetes in midlife and later development of coronary heart disease and stroke.

- Excess coronary heart disease incidence in South Asians, with lower incidence in African Caribbeans compared with Europeans continues and elevated risk of stroke was confirmed in both ethnic minority groups. Measured baseline metabolic risk factors could not explain the ethnic group differences. Of particular concern was a much stronger association between diabetes and stroke risk in both ethnic minority groups compared with Europeans with diabetes- an association which is the subject of ongoing study in the SABRE cohort. HES data contributed importantly to the identification of incident coronary and stroke events reported in these analyses.

Selection of manuscripts recently accepted for publication:

-Baseline (1989-1991) vascular risk factors explained the observed ethnic variation in cardiovascular disease recurrence and long-term mortality, with a relative improvement in survival of minority ethnic groups over time.

Vyas MV, Chaturvedi N, Hughes AD, Marmot M, Tillin T. Cardiovascular disease recurrence and long-term mortality in a tri-ethnic British cohort. Heart. 2020 Oct 16;107(12):996–1002. doi: 10.1136/heartjnl-2020-317641. Epub ahead of print. PMID: 33067326; PMCID: PMC8165149.

-Markedly stronger associations were observed between family history and clinical CHD in South Asians, and a similar trend for subclinical CHD. Early preventive and therapeutic interventions are particularly important in South Asians with a family history of CHD.

Wang J, Tillin T, Hughes AD, Chaturvedi N. Associations between family history and coronary artery calcium and coronary heart disease in British Europeans and South Asians. Int J Cardiol. 2020 Feb 1;300:39-42. doi: 10.1016/j.ijcard.2019.07.101. Epub 2019 Aug 1. PMID: 31400886; PMCID: PMC6970219.

- Urine albumin:creatinine ratio may have greater utility than estimated glomerular filtration rates in CVD risk prediction in South Asians.

Eastwood S, Chaturvedi N, Sattar N, Welsh P, I, Hughes A, D, Tillin T: Impact of Kidney Function on Cardiovascular Risk and Mortality: A Comparison of South Asian and European Cohorts. Am J Nephrol 2019;50:425-433. doi: 10.1159/000503873.

-Novel results provide support for ethnic differences in sleep quality and mortality, and may have implications for precision medicine.

Garfield V, Joshi R, Garcia-Hernandez J, Tillin T, Chaturvedi N. The relationship between sleep quality and all-cause, CVD and cancer mortality: the Southall and Brent REvisited study (SABRE). Sleep Med. 2019 Aug;60:230-235. doi: 10.1016/j.sleep.2019.03.012. Epub 2019 Mar 27. PMID: 31182326; PMCID: PMC6626293.

-Results support a detrimental long-term effect for air pollutants on cardiovascular mortality.

Dehbi HM, Blangiardo M, Gulliver J, Fecht D, de Hoogh K, Al-Kanaani Z, Tillin T, Hardy R, Chaturvedi N, Hansell AL. Air pollution and cardiovascular mortality with over 25years follow-up: A combined analysis of two British cohorts. Environ Int. 2017 Feb;99:275-281. doi: 10.1016/j.envint.2016.12.004. Epub 2016 Dec 7.

At the end of the recent funding period (extended to June 2021) a report was submitted to the funders (the British Heart Foundation) summarising findings. Some details are published on their website. (https://www.bhf.org.uk/what-we-do/our-research/research-successes/ethnicity-and-heart-disease )

Brief summary of findings to date: ethnic differentials in diabetes and coronary heart disease incidence continue even into older age and at lower levels of obesity compared with people of European descent. Not only is diabetes more frequent, but it may be more strongly associated with stroke risk in South Asian and African Caribbean populations. Family history of cardiovascular disease and mid-life diastolic as well as systolic blood pressure were more strongly associated with cardiovascular disease in South Asians. Widely used risk scores for cardiovascular disease may be less accurate in ethnic minority groups. Earlier and improved methods of monitoring and preventive strategies are needed for diabetes and its consequences in ethnic minority populations.

This is a mechanistic research study in a population-based sample, looking at underlying reasons for disparities in health and function in older age, with a particular focus on ethnic minorities. In an ageing general population and with over 16% of the population of England being from ethnic minority backgrounds (most of South Asian or Black African or African Caribbean origins), the findings impact on a very large number of people. As an example, our findings regarding the extraordinarily high incidence of new onset diabetes in South Asians and African Caribbeans, even in old age (Diabetes Care, 2013), were widely reported in the national press and on national TV/radio and led to a widespread acknowledgement that this is a much greater problem than anticipated. Findings regarding the associations between different levels of obesity and future development of type 2 diabetes were informative in NICE’s determination of BMI levels for prevention of type 2 diabetes in different ethnic minority groups. Many of UCL’s findings are published and widely cited in peer reviewed journals and thus UCL inform future research questions and add to the global understanding of ethnic differences in cardiometabolic disease (globally 18 million deaths were due to cardiovascular disease and 1.5 million due to diabetes in 2019, while numbers living with diabetes continue to rise- 537 million adults in 2021). UCL’s study is not designed to measure impact -but given the heavy burden of diabetes and cardiovascular disease in terms of length and quality of life and its particular effect on people of ethnic minority backgrounds, increased awareness and understanding can only lead to better preventive and therapeutic strategies in both primary and secondary care settings.

University College London (UCL) requires notifications of mortality and cancer registrations and linked HES data for its study cohort for use in the Medical Research Project: SABRE (Southall And Brent Revisited). This is a population-based cohort study, conducted at University College London, funded by the British Heart Foundation in its current 25 year follow-up phase. It is unique as a long-standing tri-ethnic cohort consisting of people of European descent and first generation migrants of South Asian or African Caribbean descent. This is an academic research study focusing on identifying and understanding the underlying reasons for ethnic group and sex differences in cardiometabolic disease and in physical, psychological and cognitive function in older age.

Specific questions for the 25 year follow-up study are:

1. How large are ethnic /sex differences in cardiac function, cognitive function and hippocampal volumes in older age?

2. To what extent do cardiac function, cognitive function and hippocampal volumes change over a 5 year period in each ethnic group?

3. Which risk factors measured in mid-life and in early old age are most strongly associated with current cardiac and cognitive function and hippocampal volumes and with 5 year changes in these parameters? Can these risk factors explain ethnic differences in cardiac and cognitive function?

4. How large are gender differences in current disorders of cardiac and cognitive function and in their associations with current risk factors?

5. Do ethnic differences in incident cardiometabolic disorders persist into older age?

6. Which risk factors or risk factor profiles measured in mid-life and early old age are most strongly associated with incident cardiometabolic disorders and which best explain ethnic differences in incidence?

The study has previously utilised the List Cleaning service from time to time when in active follow-up in order to ensure that the correct participant addresses are used in order to contact participants. Use of this service has helped the study to avoid trying to contact deceased participants. The List Cleaning outputs were used to update the administration database (held separately from other data within the UCL data safe haven) so that UCL could write to as many participants as possible inviting them to complete questionnaires or come into the UCL clinic for a detailed investigation. Under this Data Sharing Agreement, UCL may retain List Cleaning outputs received previously but is not permitted to make further use of the List Cleaning service.

Linked HES APC data from NHS Digital has previously been disseminated to the SABRE study in order to identify incident cardiometabolic events (in particular coronary heart disease, heart failure, stroke, dementia, diabetes), and other events which may affect physical and cognitive function, which have occurred during the follow-up period. Details of all hospital episodes involving the cohort (not limited to the previously stated conditions) were required to address key study objectives with regard to physical and cognitive function in older age in association with current and mid-life risk factors.

Analysis considers any and all potential contributing factors.

These events will supplement information provided by participant self-report at 20 and 25 years, from primary care medical record review conducted during the 20 year follow-up and from cancer and mortality flagging, together with detailed clinical measurements made at the SABRE clinics at baseline, 20 and 25 year follow-up. The data will be used to analyse risk factors measured in mid- and later life in association with these incident events in order to build on current understanding of causal mechanisms.

Data from 1989 to the present are required because participants underwent detailed examinations at baseline (1989-91) and the aim is to follow this cohort through their experiences since to understand what happened in later life and relate that to the baseline. This will enable UCL to gain as complete as possible a picture of hospital admissions, and hence incident events, over the entire cohort follow-up. Data from the entire study period are crucial for determining age of onset of events, as well as the extent and nature of ill-health from mid to later life, and for relating these to current and mid-life cardiometabolic and other risk factors and how these influence the key study outcomes of physical and cognitive function in older life in each of the three ethnic groups.

This request is to extend the term of the data sharing agreement to complete the complex analyses which will be required to integrate outcomes derived from HES data with mortality and cancer registration data and the very detailed phenotypic data collected during the course of the study period since 1989. In addition, this request is also to continue to receive mortality and cancer registration data, as these are, sadly, key outcomes which can be related to risk factors identified in earlier life. It is also important that notifications of death are received. No raw HES data from NHS Digital are being disseminated under this new agreement.

There are a number of reasons why UCL had not sought to consent all study participants for study health data linkage:

- At the start of the third wave of follow-up (2014-2018) UCL conducted a pilot study to assess whether it was feasible to obtain consent for data linkage using a postal approach to existing participants (UCL sent out a covering letter, participant information, consent form including an item on data linkage, health and lifestyle questionnaire and reply paid return envelope). At that time there were 3400 surviving participants, aged between 65 and 98 years. UCL approached 773 participants who had visited the clinic at the previous phase of follow-up. Of these people, who UCL had considered to be highly motivated participants, UCL received positive responses from 65%, negative responses from 9%, no response from 17%, address incorrect (1%) and a further 10% who requested time to consider or who asked to be contacted again in the future. Of the latter group, although UCL know that they have received the information, very few have been contactable with regard to the study, bringing the total estimated non-response rate in this group to approximately 25%. This non-response rate was unexpectedly high and likely to be related to ill health, particularly cognitive impairment.

- With a 17-25% non-response rate already demonstrated in the pilot study of the highly motivated group, UCL had reasonable grounds for assuming a much higher non-response rate in the rest of the cohort (2070 survivors), with undoubted risk to the scientific integrity of the follow-up study. UCL had consulted with the Confidentiality Advisory Group (CAG) and followed their advice regarding application for section 251 support. With CAG agreement UCL modified the approach to existing participants by removing the item on consent to data linkage from the postal consent form, while adding an item at the start of the questionnaire that enabled participants to opt-out of data linkage even if they chose not to come in to the clinic. 21 people chose to opt out of data linkage in this way. UCL then placed an additional item requesting consent for data linkage in the clinic consent form for responders who had agreed to visit UCL for clinical follow-up.

- Loss to follow-up is an important problem in older cohorts. People who do not respond, or are lost to follow-up in other ways, may be different from those who participate - often they are more unwell than responders. This is likely to lead to bias and to jeopardise the scientific value of the findings and grossly bias public health messages and potential interventions. Given the unique experiences of the first generation migrants in this cohort, which may be vital in informing health service strategies in an increasingly migratory world, UCL believe that it would be a waste of the data and commitment to the study in earlier years if UCL are unable to reach sound scientific conclusions due to loss to follow-up. This is where linkage to health service records can minimise the effects of loss to follow-up due to non-response.

This agreement is for university research, the lawful basis for processing data is GDPR 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’. Also referred to as ‘Public Task’. As the research involves health data, which is included in the definition of special categories of personal data, it requires an additional condition for processing. Based on guidance, for health research this is article 9(2)(j), which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. The aim of this study - understanding of underlying causes of difference in physical and cognitive function in people of different ethnicities is in the public interest, because it will ultimately lead to appropriate preventive strategies and treatments at different stages of life.

Study findings will continue to be published in peer-reviewed scientific journals, predominantly related to epidemiology, cardiovascular and metabolic disorders, cognitive, physical and psychological function, but also including more generic journals such as the BMJ, reflecting the increasing focus on overall health and function in older age. Publications will contain only aggregate level data without local identifiers and with suppression of small numbers in line with HES analysis guide.

Publications to date are listed on the study website: www.sabrestudy.org. All publications since 2008 are open-access. The audience is expected to consist mainly of academic researchers and clinicians.

The study team at UCL will further disseminate findings via participant and GP feedback sessions; newsletters, and the study website.

At the end of the current funding period (extended to December 2020) a report will be submitted to the funders (the British Heart Foundation) summarising findings. This may be published on their website.

Update (May2019) : Since the previous data sharing agreement was approved in 2018, the study has been in active clinical follow-up (ended January 2019) and data preparation is not yet complete. However, a number of new research papers based on this cohort have been submitted for publication or published during the period, some of which utilise earlier cardiovascular outcome data, some examples:

Targets:

Complete clinical data preparation and processing of HES, and mortality data during 2019-20, ready for analysis and write up.

Outputs are expected to include dissemination of key findings to participants and their GPs during 2019-2020.

Relevant professional conference attendances with abstract presentations: include British Society of Cardiology, European Society of Cardiology, European Association for the Study of Diabetes, British and Irish Hypertension Society, Artery. 2019-2020 and ongoing. Note: All data for abstracts/presentations will be at aggregate level with suppression of small numbers in line with HES analysis guide.

Papers to be prepared for submission to peer-reviewed journals during 2019-2020 will include studies of risk factors over 30 years of follow-up in association with:

- ethnicity and all-cause and cardiovascular disease-related mortality,

- ethnicity and acute and chronic cardiovascular disease , including heart failure

- ethnicity and diabetes incidence and complications

- ethnicity and mental health

The research has also enabled improved ascertainment of incident coronary heart disease and stroke events and resulted in 10 publications in high impact journals relating these outcomes to risk factors measured in mid-life (ages 40-70 at baseline). Findings include:

- Diabetes incidence in older British South Asians and African Caribbeans remains at least 2-fold even at age 80 years compared with British Europeans. The ethnic differentials in women were largely explained by midlife truncal obesity and insulin resistance, but the study was unable to explain the ethnic difference in men. It is anticipated that more recent follow-up may elucidate these mechanisms.

-The study showed that obesity cut-points of 24 and 27 kg/m2 in South Asians and African Caribbeans respectively were equivalent to a body mass index of 30kg/m2 in Europeans in terms of diabetes risk; these latter analyses contributed to recent NICE guidelines for prevention of diabetes. Diabetes was also found to be more ‘toxic’ in terms of stroke risk in the ethnic minorities.

-Widely used tools (Framingham and QRISK2) for estimation of cardiovascular disease risk were found to be less precise in South Asians and African Caribbeans (particularly women), while a selection of 3 metabolic markers measured by NMR spectroscopy was found to be strongly predictive of cardiovascular risk regardless of ethnicity.

- Lack of adherence to four combined health behaviours was associated with a 2 to 3-fold increased risk of incident CVD in Europeans and South Asians. A substantial population impact in the South Asian group indicates important potential for disease prevention in this high-risk group by adherence to healthy behaviours.

- The study has also found marked ethnic differences in associations between blood pressure parameters and stroke and concluded that undue focus on systolic blood pressure for risk prediction, and current age and treatment thresholds may be inappropriate for individuals of South Asian ancestry.

- The study reported detrimental associations between air pollution (particulate measures) and cardiovascular disease mortality in both the SABRE and Whitehall cohorts. The study highlighted ethnic differences in associations between prediabetes in midlife and later development of coronary heart disease and stroke.

- The study has confirmed ongoing excess coronary heart disease incidence in South Asians, with lower incidence in African Caribbeans compared with Europeans and confirmed elevated risk of stroke in both ethnic minority groups. Measured baseline metabolic risk factors could not explain the ethnic group differences. Future work in the cohort will examine whether these ethnic differentials continue into older age and whether newer genetic, epigenetic and metabolomic analyses will add to understanding of the underlying mechanisms. Of particular concern was a much stronger association between diabetes and stroke risk in both ethnic minority groups compared with Europeans with diabetes- an association which is the subject of ongoing study in the SABRE cohort. HES data contributed importantly to the identification of incident coronary and stroke events reported in these analyses.

- The study has also shown ethnic differences in sleep quality which related to cardiovascular disease and cancer mortality, notably difficulty in falling asleep was associated with increased risk of all-cause mortality in South Asians, while early morning waking was associated with increased risk of cardiovascular death in Europeans

-We have also shown that different markers of kidney function predict cardiovascular disease and mortality in Europeans compared with South Asians, suggesting that urinary albumin: creatinine ratio may have greater utility in South Asians and that estimated glomerular filtration rate may be more useful in Europeans

Recently accepted for publication:

Victoria Garfield, Roshni Joshi, Jorge Garcia-Hernandez, Therese Tillin, Nish Chaturvedi

The relationship between sleep quality and all-cause, CVD and cancer mortality: The Southall and Brent REvisited Study (SABRE). Sleep Medicine, in press

Dehbi HM, Blangiardo M, Gulliver J, Fecht D, de Hoogh K, Al-Kanaani Z, Tillin T, Hardy R, Chaturvedi N, Hansell AL Air pollution and cardiovascular mortality with over 25years follow-up: A combined analysis of two British cohorts. Environ Int. 2017 Feb;99:275-281. doi: 10.1016/j.envint.2016.12.004. Epub 2016 Dec 7.

Halonen JI, Dehbi HM, Hansell AL, Gulliver J, Fecht D, Blangiardo M, Kelly FJ, Chaturvedi N, Kivimäki M, Tonne C. Associations of night-time road traffic noise with carotid intima-media thickness and blood pressure: The Whitehall II and SABRE study cohorts. Environ Int. 2017 Jan;98:54-61. doi: 10.1016/j.envint.2016.09.023. Epub 2016 Oct 3.

Recently submitted:

Journal: The Lancet

Title: Age at natural menopause and risk of incident cardiovascular disease: A pooled analysis of 15 studies.

Gita Mishra; Dongshan Zhu; Hsin-Fang Chung; Annette J Dobson; Nirmala Pandeya; Graham G Giles; Fiona Bruinsma; Eric J Brunner; Diana Kuh; Rebecca Hardy; Nancy E Avis; Ellen B Gold; Carol A Derby; Karen A Matthews; Janet E Cade; Darren C Greenwood; Panayotes Demakakos; Daniel E Brown; Lynnette L Sievert; Debra Anderson; Kunihiko Hayashi; Jung Su Lee; Hideki Mizunuma; Therese Tillin; Mette K Simonsen; Hans-Olov Adami; Elisabete Weiderpass;

Journal:

International Journal of Cardiology

Wang J, Tillin T, Hughes AD, Chaturvedi NC. Associations between family history and coronary artery calcification and coronary heart disease in Europeans and South Asians

Journal: Nephrology, Dialysis and Transplantation

Eastwood S, Chaturvedi N, Sattar N, Welsh P, Hughes AD, Tillin T. Associations between markers of kidney function and mortality or incident cardiovascular disease; a comparison of UK South Asian and European groups

University College London (UCL) requires notifications of mortality and cancer registrations and linked HES data for its study cohort for use in the Medical Research Project: SABRE (Southall And Brent Revisited). This is a population-based cohort study, conducted at University College London, funded by the British Heart Foundation in its current 25 year follow-up phase. It is unique as a long-standing tri-ethnic cohort consisting of people of European descent and first generation migrants of South Asian or African Caribbean descent. This is an academic research study focusing on identifying and understanding the underlying reasons for ethnic group and sex differences in cardiometabolic disease and in physical, psychological and cognitive function in older age.

Specific questions for the 25 year follow-up study are:

1. How large are ethnic /sex differences in cardiac function, cognitive function and hippocampal volumes in older age?

2. To what extent do cardiac function, cognitive function and hippocampal volumes change over a 5 year period in each ethnic group?

3. Which risk factors measured in mid-life and in early old age are most strongly associated with current cardiac and cognitive function and hippocampal volumes and with 5 year changes in these parameters? Can these risk factors explain ethnic differences in cardiac and cognitive function?

4. How large are gender differences in current disorders of cardiac and cognitive function and in their associations with current risk factors?

5. Do ethnic differences in incident cardiometabolic disorders persist into older age?

6. Which risk factors or risk factor profiles measured in mid-life and early old age are most strongly associated with incident cardiometabolic disorders and which best explain ethnic differences in incidence?

The study has previously utilised the List Cleaning service from time to time when in active follow-up in order to ensure that the correct participant addresses are used in order to contact participants. Use of this service has helped the study to avoid trying to contact deceased participants. The List Cleaning outputs were used to update the administration database (held separately from other data within the UCL data safe haven) so that UCL could write to as many participants as possible inviting them to complete questionnaires or come into the UCL clinic for a detailed investigation. Under this Data Sharing Agreement, UCL may retain List Cleaning outputs received previously but is not permitted to make further use of the List Cleaning service.

Linked HES APC data from NHS Digital has previously been disseminated to the SABRE study in order to identify incident cardiometabolic events (in particular coronary heart disease, heart failure, stroke, dementia, diabetes), and other events which may affect physical and cognitive function, which have occurred during the follow-up period. Details of all hospital episodes involving the cohort (not limited to the previously stated conditions) were required to address key study objectives with regard to physical and cognitive function in older age in association with current and mid-life risk factors.

Analysis considers any and all potential contributing factors.

These events will supplement information provided by participant self-report at 20 and 25 years, from primary care medical record review conducted during the 20 year follow-up and from cancer and mortality flagging, together with detailed clinical measurements made at the SABRE clinics at baseline, 20 and 25 year follow-up. The data will be used to analyse risk factors measured in mid- and later life in association with these incident events in order to build on current understanding of causal mechanisms.

Data from 1989 to the present are required because participants underwent detailed examinations at baseline (1989-91) and the aim is to follow this cohort through their experiences since to understand what happened in later life and relate that to the baseline. This will enable UCL to gain as complete as possible a picture of hospital admissions, and hence incident events, over the entire cohort follow-up. Data from the entire study period are crucial for determining age of onset of events, as well as the extent and nature of ill-health from mid to later life, and for relating these to current and mid-life cardiometabolic and other risk factors and how these influence the key study outcomes of physical and cognitive function in older life in each of the three ethnic groups.

This request is to extend the term of the data sharing agreement to complete the complex analyses which will be required to integrate outcomes derived from HES data with mortality and cancer registration data and the very detailed phenotypic data collected during the course of the study period since 1989. In addition, this request is also to continue to receive mortality and cancer registration data, as these are, sadly, key outcomes which can be related to risk factors identified in earlier life. It is also important that notifications of death are received. No raw HES data from NHS Digital are being disseminated under this new agreement.

There are a number of reasons why UCL had not sought to consent all study participants for study health data linkage:

- At the start of the third wave of follow-up (2014-2018) UCL conducted a pilot study to assess whether it was feasible to obtain consent for data linkage using a postal approach to existing participants (UCL sent out a covering letter, participant information, consent form including an item on data linkage, health and lifestyle questionnaire and reply paid return envelope). At that time there were 3400 surviving participants, aged between 65 and 98 years. UCL approached 773 participants who had visited the clinic at the previous phase of follow-up. Of these people, who UCL had considered to be highly motivated participants, UCL received positive responses from 65%, negative responses from 9%, no response from 17%, address incorrect (1%) and a further 10% who requested time to consider or who asked to be contacted again in the future. Of the latter group, although UCL know that they have received the information, very few have been contactable with regard to the study, bringing the total estimated non-response rate in this group to approximately 25%. This non-response rate was unexpectedly high and likely to be related to ill health, particularly cognitive impairment.

- With a 17-25% non-response rate already demonstrated in the pilot study of the highly motivated group, UCL had reasonable grounds for assuming a much higher non-response rate in the rest of the cohort (2070 survivors), with undoubted risk to the scientific integrity of the follow-up study. UCL had consulted with the Confidentiality Advisory Group (CAG) and followed their advice regarding application for section 251 support. With CAG agreement UCL modified the approach to existing participants by removing the item on consent to data linkage from the postal consent form, while adding an item at the start of the questionnaire that enabled participants to opt-out of data linkage even if they chose not to come in to the clinic. 21 people chose to opt out of data linkage in this way. UCL then placed an additional item requesting consent for data linkage in the clinic consent form for responders who had agreed to visit UCL for clinical follow-up.

- Loss to follow-up is an important problem in older cohorts. People who do not respond, or are lost to follow-up in other ways, may be different from those who participate - often they are more unwell than responders. This is likely to lead to bias and to jeopardise the scientific value of the findings and grossly bias public health messages and potential interventions. Given the unique experiences of the first generation migrants in this cohort, which may be vital in informing health service strategies in an increasingly migratory world, UCL believe that it would be a waste of the data and commitment to the study in earlier years if UCL are unable to reach sound scientific conclusions due to loss to follow-up. This is where linkage to health service records can minimise the effects of loss to follow-up due to non-response.

This agreement is for university research, the lawful basis for processing data is GDPR 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’. Also referred to as ‘Public Task’. As the research involves health data, which is included in the definition of special categories of personal data, it requires an additional condition for processing. Based on guidance, for health research this is article 9(2)(j), which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. The aim of this study - understanding of underlying causes of difference in physical and cognitive function in people of different ethnicities is in the public interest, because it will ultimately lead to appropriate preventive strategies and treatments at different stages of life.

Study findings will continue to be published in peer-reviewed scientific journals, predominantly related to epidemiology, cardiovascular and metabolic disorders, cognitive, physical and psychological function, but also including more generic journals such as the BMJ, reflecting the increasing focus on overall health and function in older age. Publications will contain only aggregate level data without local identifiers and with suppression of small numbers in line with HES analysis guide.

Publications to date are listed on the study website: www.sabrestudy.org. All publications since 2008 are open-access. The audience is expected to consist mainly of academic researchers and clinicians.

The study team at UCL will further disseminate findings via participant and GP feedback sessions; newsletters, and the study website.

At the end of the current funding period (extended to December 2020) a report will be submitted to the funders (the British Heart Foundation) summarising findings. This may be published on their website.

Update (May2019) : Since the previous data sharing agreement was approved in 2018, the study has been in active clinical follow-up (ended January 2019) and data preparation is not yet complete. However, a number of new research papers based on this cohort have been submitted for publication or published during the period, some of which utilise earlier cardiovascular outcome data, some examples:

Targets:

Complete clinical data preparation and processing of HES, and mortality data during 2019-20, ready for analysis and write up.

Outputs are expected to include dissemination of key findings to participants and their GPs during 2019-2020.

Relevant professional conference attendances with abstract presentations: include British Society of Cardiology, European Society of Cardiology, European Association for the Study of Diabetes, British and Irish Hypertension Society, Artery. 2019-2020 and ongoing. Note: All data for abstracts/presentations will be at aggregate level with suppression of small numbers in line with HES analysis guide.

Papers to be prepared for submission to peer-reviewed journals during 2019-2020 will include studies of risk factors over 30 years of follow-up in association with:

- ethnicity and all-cause and cardiovascular disease-related mortality,

- ethnicity and acute and chronic cardiovascular disease , including heart failure

- ethnicity and diabetes incidence and complications

- ethnicity and mental health

The research has also enabled improved ascertainment of incident coronary heart disease and stroke events and resulted in 10 publications in high impact journals relating these outcomes to risk factors measured in mid-life (ages 40-70 at baseline). Findings include:

- Diabetes incidence in older British South Asians and African Caribbeans remains at least 2-fold even at age 80 years compared with British Europeans. The ethnic differentials in women were largely explained by midlife truncal obesity and insulin resistance, but the study was unable to explain the ethnic difference in men. It is anticipated that more recent follow-up may elucidate these mechanisms.

-The study showed that obesity cut-points of 24 and 27 kg/m2 in South Asians and African Caribbeans respectively were equivalent to a body mass index of 30kg/m2 in Europeans in terms of diabetes risk; these latter analyses contributed to recent NICE guidelines for prevention of diabetes. Diabetes was also found to be more ‘toxic’ in terms of stroke risk in the ethnic minorities.

-Widely used tools (Framingham and QRISK2) for estimation of cardiovascular disease risk were found to be less precise in South Asians and African Caribbeans (particularly women), while a selection of 3 metabolic markers measured by NMR spectroscopy was found to be strongly predictive of cardiovascular risk regardless of ethnicity.

- Lack of adherence to four combined health behaviours was associated with a 2 to 3-fold increased risk of incident CVD in Europeans and South Asians. A substantial population impact in the South Asian group indicates important potential for disease prevention in this high-risk group by adherence to healthy behaviours.

- The study has also found marked ethnic differences in associations between blood pressure parameters and stroke and concluded that undue focus on systolic blood pressure for risk prediction, and current age and treatment thresholds may be inappropriate for individuals of South Asian ancestry.

- The study reported detrimental associations between air pollution (particulate measures) and cardiovascular disease mortality in both the SABRE and Whitehall cohorts. The study highlighted ethnic differences in associations between prediabetes in midlife and later development of coronary heart disease and stroke.

- The study has confirmed ongoing excess coronary heart disease incidence in South Asians, with lower incidence in African Caribbeans compared with Europeans and confirmed elevated risk of stroke in both ethnic minority groups. Measured baseline metabolic risk factors could not explain the ethnic group differences. Future work in the cohort will examine whether these ethnic differentials continue into older age and whether newer genetic, epigenetic and metabolomic analyses will add to understanding of the underlying mechanisms. Of particular concern was a much stronger association between diabetes and stroke risk in both ethnic minority groups compared with Europeans with diabetes- an association which is the subject of ongoing study in the SABRE cohort. HES data contributed importantly to the identification of incident coronary and stroke events reported in these analyses.

- The study has also shown ethnic differences in sleep quality which related to cardiovascular disease and cancer mortality, notably difficulty in falling asleep was associated with increased risk of all-cause mortality in South Asians, while early morning waking was associated with increased risk of cardiovascular death in Europeans

-We have also shown that different markers of kidney function predict cardiovascular disease and mortality in Europeans compared with South Asians, suggesting that urinary albumin: creatinine ratio may have greater utility in South Asians and that estimated glomerular filtration rate may be more useful in Europeans

Recently accepted for publication:

Victoria Garfield, Roshni Joshi, Jorge Garcia-Hernandez, Therese Tillin, Nish Chaturvedi

The relationship between sleep quality and all-cause, CVD and cancer mortality: The Southall and Brent REvisited Study (SABRE). Sleep Medicine, in press

Dehbi HM, Blangiardo M, Gulliver J, Fecht D, de Hoogh K, Al-Kanaani Z, Tillin T, Hardy R, Chaturvedi N, Hansell AL Air pollution and cardiovascular mortality with over 25years follow-up: A combined analysis of two British cohorts. Environ Int. 2017 Feb;99:275-281. doi: 10.1016/j.envint.2016.12.004. Epub 2016 Dec 7.

Halonen JI, Dehbi HM, Hansell AL, Gulliver J, Fecht D, Blangiardo M, Kelly FJ, Chaturvedi N, Kivimäki M, Tonne C. Associations of night-time road traffic noise with carotid intima-media thickness and blood pressure: The Whitehall II and SABRE study cohorts. Environ Int. 2017 Jan;98:54-61. doi: 10.1016/j.envint.2016.09.023. Epub 2016 Oct 3.

Recently submitted:

Journal: The Lancet

Title: Age at natural menopause and risk of incident cardiovascular disease: A pooled analysis of 15 studies.

Gita Mishra; Dongshan Zhu; Hsin-Fang Chung; Annette J Dobson; Nirmala Pandeya; Graham G Giles; Fiona Bruinsma; Eric J Brunner; Diana Kuh; Rebecca Hardy; Nancy E Avis; Ellen B Gold; Carol A Derby; Karen A Matthews; Janet E Cade; Darren C Greenwood; Panayotes Demakakos; Daniel E Brown; Lynnette L Sievert; Debra Anderson; Kunihiko Hayashi; Jung Su Lee; Hideki Mizunuma; Therese Tillin; Mette K Simonsen; Hans-Olov Adami; Elisabete Weiderpass;

Journal:

International Journal of Cardiology

Wang J, Tillin T, Hughes AD, Chaturvedi NC. Associations between family history and coronary artery calcification and coronary heart disease in Europeans and South Asians

Journal: Nephrology, Dialysis and Transplantation

Eastwood S, Chaturvedi N, Sattar N, Welsh P, Hughes AD, Tillin T. Associations between markers of kidney function and mortality or incident cardiovascular disease; a comparison of UK South Asian and European groups