A population-based retrospective cohort study into the factors associated with Emergency Department attendance by people with dementia in the last year of life.

King’s College London requires data from the Hospital Episode Statistics Data (HES) Accident and Emergency Set (A+E), HES Admitted Patient Care (APC) data set and Civil Registration (Deaths) - Secondary Care Cut data for the study “A population-based retrospective cohort study into the factors associated with Emergency Department attendance by people with dementia at the end of life”. In this study, the Emergency Department (ED) is defined as a consultant-led accident and emergency service with full resuscitation facilities operating all day, every day (1).

Data will be used for analysis of a single observational study that constitutes the quantitative strand of a mixed methods study as part of a standalone PhD. This PhD is one of three funded by the Alzheimer's Society and aims to understand why people with dementia attend the Emergency Department (ED) at the end of life. The PhD Clinical Training Fellow is a substantive employee of King’s College London and is supervised by two senior clinical academics who are also substantive employees of King’s College London. The other two PhDs are not focused on ED attendance or end of life care for people with dementia and are not the focus of this application.

An estimated 850,000 people have dementia in the UK (2). The disease has an uncertain trajectory and for many, in the last year of life, there are multiple hospitalisations and Emergency Department (ED) attendances (3,4). Hospital transfer at the end of life is widely considered to be a marker or poor-quality care (5). It can be distressing and disorientating for the patient with dementia (6) and places additional strain on the emergency and acute care services with sharp increases in hospital costs in the final few months of life (7).

A recent systematic review of international literature has found high strength evidence of individual, clinical and environmental risk factors increasing ED attendance by people with dementia towards the end of life (8). However, the majority of the reviewed studies were conducted in the USA. Furthermore, whilst much of the literature focuses on ‘burdensome’ transitions to hospital, comparatively few focus on ED attendance, fewer are conducted in the UK, and none that are on a population-based scale. One small-scale study within four London boroughs showed increasing number of ED attendances with closer proximity to death. The study also identified differential predictors of ED attendance, including various socio-economic variations (4). Expanding on these findings, and contributing to the international literature in the field, this study aims to capture population-based variance to increase generalisability and practical application.

On the basis of the literature, this study aims to identify the factors associated with ED attendance by people with dementia in the last year of life. The objectives of the study are:

1) To examine the frequency of ED attendance and re-attendance by people with dementia in the last year of life;

2) To ascertain the reasons for ED attendance among people with dementia in the last year of life;

3) To identify the predictors of ED attendance by people with dementia in the last year of life.

To meet these objectives, King's College London requires pseudonymised data from HES A&E, HES APC and Civil Registration (Deaths) - Secondary Care Cut databases for linkage. Details of how access to this data will meet these objectives are as follows:

Objective 1: Access to HES A&E to obtain the attendance category, disposal and department type to enable review of frequency of ED attendance, stratified by outcome of ED attendance and number of attendances for the same incident. Linked to data from the Civil Registration (Deaths) - Secondary Care Cut, this data will provide insight into the proportion and frequency of ED attendances, at different intervals within the last year of life (e.g. last twelve, six and three months, and last 30, seven and three days of life).

Objective 2: Access to HES A&E to obtain the attendance category, department type, arrival mode, incident location type, source of referral for A&E and A&E diagnosis to enable review of the types and frequency of precipitants to ED attendance. A&E investigations and clinical treatments will provide insight into the acuity and severity of the incident that has led to ED attendance. Access to HES APC to obtain the diagnosis and episode duration and type if the patient was admitted into hospital from the ED. Linked to data from the Civil Registration (Deaths) - Secondary Care Cut, this data will provide detail on the main reasons for ED attendance in the last year of life (i.e. ambulatory care sensitive conditions, carer strain).

Objective 3: Access to HES A&E to obtain the arrival mode, arrival date, arrival time, IMD decile group, IMD Index of Multiple Deprivation, Lower Super Output Area, CCG of GP Practice, county of residence, rural/urban indicator, A&E diagnosis, age at activity date, age on arrival, carer support indicator, ethnic category, month and year of birth, patient sex and postcode district. Access to HES APC will provide data on diagnoses, comorbidities, ambulatory case sensitive conditions, carer support and marital status. Linked to data from the Civil Registration (Deaths) - Secondary Care Cut, the data will provide valuable information of the predisposing factors associated with ED attendance and repeat attendance in the last year of life.

In order to improve the care and management of patients, it is important to understand the factors associated with ED attendance at the end of life. It is proposed that by analysing the frequency of, reasons for, and factors associated with ED attendance at the end of life, high-risk patients can be profiled and modifiable risk factors identified to direct future targeted policy and service interventions. It will also generate further research avenues into the effectiveness of these interventions and to address any gaps identified in the evidence base as a result of the findings of this study.

The study will require data about adult decedents with a diagnosis of dementia (of any sub-type and any severity) who died between 01 April 2018 and 31 March 2019. Linked HES A&E, HES APC and Civil Registration (Deaths) - Secondary Care Cut data will allow assessment of ED attendance in the last year of life, to correspond with the definition of end of life (9). Based on previous literature guiding this study, factors associated with end of life ED attendances will be grouped into categories and subject to mixed effects regression models to measure ED attendance counts and occurrences.

Data are requested from HES years 2016 to 2019 to identify people who had an ED attendance 12-36 months before death.

Only record-level data, with pseudonymised HES IDs, can yield the necessary information to draw conclusions about patterns of ED attendances, therefore no less obtrusive methods are available. Date of death has been requested. Acknowledging the sensitivity status of this field, there are no other means of measuring the number of ED attendances within the last twelve, six and three months, and last 30, seven and three days of life. Month and year of birth and age at death alone will be insufficient to accurately determine days, weeks and months in the last year of life. There is therefore no less obtrusive method available to determine this time-frame. However, it will be integral to the study in order to replicate findings of a previous local study (4) and to identify factors associated with the frequency of ED attendance compared to proximity of death.

King's College London is the only organisation involved in this study, being the sole Data Controller and Data Processor. No other organisations will process the data for this purpose. All researchers involved in the study are employed by King's College London as research or academic staff. Under this Agreement, King's College London requires pseudonymised data until the expiry date of this agreement. At this expiry date, all pseudonymised record level data will be destroyed with a Data Destruction Certificate submitted to NHS Digital as evidence.

Data will be presented at aggregate level with small number suppression applied in line with the HES analysis guide on all published results - so that no individuals are identifiable. The outputs will help clinicians and policy makers ensure the future provision of best care.

NHS Digital will extract data on patients as per the coding provided, securely link across all datasets requested and send to King’s College London a pseudonymised dataset for analysis.

References

(1) NHS Data Model and Dictionary (2020). Accident and emergency department type. Available from: https://www.datadictionary.nhs.uk/attributes/accident_and_emergency_department_type.html

(2) NHS England. (no date). Dementia. Available from: https://www.england.nhs.uk/mental-health/dementia/

(3) Leniz, J., Higginson, I. J., Stewart, R., & Sleeman, K. E. (2019). Understanding which people with dementia are at risk of inappropriate care and avoidable transitions to hospital near the end-of-life: a retrospective cohort study. Age and Ageing, 48(5), 672-679.

(4) Sleeman, K. E., Perera, G., Stewart, R., & Higginson, I. J. (2018). Predictors of emergency department attendance by people with dementia in their last year of life: Retrospective cohort study using linked clinical and administrative data. Alzheimer's and Dementia, 2018. 14(1): p. 20-27.

(5) Gozalo, P., Teno, J. M., Mitchell, S. L., Skinner, J., Bynum, J., Tyler, D., & Mor, V. (2011). End-of-life transitions among nursing home residents with cognitive issues. New England Journal of Medicine, 365(13), 1212-1221.

(6) Clevenger CK, Chu TA, Yang Z, Hepburn KW. Clinical Care of Persons with Dementia in the

Emergency Department: A Review of the Literature and Agenda for Research. Journal of the

American Geriatrics Society. 2012;60(9):1742-8.

(7) Georghiou, T., Davies, S., Davies, A., & Bardsley, M. (2012). Understanding patterns of health and social care at the end of life. London: Nuffield Trust.

(8) Thoms, L., Evans, C., Leniz Martelli, J., Yorganic, E., Cripps, R., & Sleeman, K. (unpublished). Factors associated with Emergency Department attendance by people with dementia near the end of life: A systematic review.

(9) NICE (2017). End of life care for adults. Quality Standard (QS13). Available from:

https://www.nice.org.uk/guidance/qs13/resources/end-of-life-care-for-adults-pdf-2098483631557.

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)”

NHS Digital will create the cohort using specified filters (ICD 10 Codes relating to dementia - ICD-10 codes F00*-03* and G30), extract relevant records and pseudonymise the data.

There will be a single flow of linked and pseudonymised record level data from NHS Digital to King’s College London, with pseudonymised ID and no identifying data. There will be no further flow of data. King's College London will be the sole organisation involved in processing the data.

Researchers at King's College London will manage the storage, cleaning, analysis and interpretation of the data. The data will not be linked with any other record-level data or be matched with publicly available data. There will be no requirement or attempt to re-identify individuals from the data.

In accordance with the Data Sharing Framework Contract between NHS Digital and King's College London, data will only be accessed by individuals within the Cicely Saunders Institute, King's College London, who have authorisation from NHS Digital to access the data for the purpose described, all of whom are substantive employees of King's College London and trained in data protection and confidentiality.

King's College London will store the data on a secure server at the Cicely Saunders Institute, King's College London. Following review of King's College London remote working policies, authorisation of remote data access from NHS Digital was granted on the 23-06-2020, should the COVID-19 social restrictions prohibit travel and entrance to the Cicely Saunders Institute building. On this basis, data securely stored on the server will only be accessible from within the Cicely Saunders Institute building (located at Denmark Hill campus, King's College London) or remotely using authorised departmental laptops only by researchers named on the project.

Data analysis will be conducted by role-based access, limited to researchers working in the study team, on the departmental computer or laptop. The data will not be made available to any third parties.

Results will be presented at an aggregate level in research outputs, with small cell counts suppressed (n<10). All data will remain anonymous. No record level data falling under this agreement will be shared with any third-party.

For all outputs, results will be presented as aggregate data with small numbers suppressed, in line with HES analysis guide. The PhD Clinical Training Fellow, with two senior clinical academic supervisors, will be responsible for dissemination of findings to local and regional policy makers, Clinical Commissioning Group (CCG) leads and clinicians to ensure results have appropriate reach and impact. In doing so, strategic, service and clinical end of life decisions will be better informed to improve the quality of end of life care for people with dementia.

Dissemination strategy:

Within 18 months of receiving access to the data, the following will be produced:

1. Peer-reviewed publication to open-access, high impact journals to reach as wide an audience as possible

2. Peer to peer dissemination of findings through international conferences, such as Alzheimer's Association International Conference and European Association for Palliative Care Congress. Although dates for 2021 are not yet available, based on this year’s virtual events, anticipated dates will be 26th-30th July and 6th-8th October, respectively.

3. A policy brief to summarise the results of the study and to respond to any relevant calls for information from MPs, healthcare committees, task forces or special interest groups. Any individuals or groups with an interest in end of life care policy and dementia will be actively sought for dissemination of findings. These groups include, although are not limited to, the Fix Dementia Care campaign (Alzheimer’s Society), In My Own Bed Please (independent group of researchers and physicians), John’s Campaign (carers of people with dementia), Find Your 1% campaign (Dying Matters, Hospice UK) and the End of Life Care Campaign (national charity coalition).

The purpose of these publications, presentations and policy brief is to disseminate the project’s findings to the scientific, clinical and policy-maker communities and provoke discussion on how the results might shape services and the end of life care experience for people with dementia.

The PhD Clinical Training Fellow has drawn on the established Patient and Public Involvement (PPI) infrastructure at the Cicely Saunders Institute (CSI). This includes presentation at a virtual PPI workshop and use of the CSI Public Involvement Forum, designed to facilitate discussions and feedback between PPI representatives and researchers. The PhD Clinical Training Fellow also meets bi-annually with two volunteers from the Alzheimer’s Society Research Network, who, with their lived experience, monitor the progression of the research project.

Consultation has to date focused on the importance of the research aim and questions for people affected by dementia (please see uploaded evidence). Continued PPI will be integral to the study, as volunteers will assist in identifying and developing themes from data and advising on and helping to develop and implement dissemination plans, including identifying wider dissemination groups. The aims of this involvement will be to ensure the outcomes measured are of relevance to patients and carers, providing interpretations of findings beyond a researcher perspective, based on their lived experience, and supporting effective dissemination of the study results to appropriate groups. The PhD Clinical Training Fellow will continue consultation with PPI members by email and using the CSI Public Involvement forum and will continue to have bi-annual meetings with Alzheimer’s Society Research Network volunteers. The next meeting is scheduled for December 2020.

To support PPI throughout the study, PPI members are invited to receive training, including that available at the local Biomedical Research Centre and Clinical Research Network, and will be asked to reflect on their learning needs throughout the project. PPI members have been and will continue to be reimbursed for out-of-pocket expenses and receive an hourly fee in recognition of their time and contribution, in accordance with NIHR Involve guidance.

Communication strategy:

Within 18 months of receiving access to the data, information about the project will be shared with interested groups and more broadly to members of the public through different channels including, although not limited to, the Cicely Saunders Institute webpage (https://www.kcl.ac.uk/cicelysaunders/newsevents/news) and YouTube channel (https://www.youtube.com/user/CSIKCL), social media platforms (i.e. @CSI_KCL, @ThomsLEA; linkedin.com/in/lesleythoms), and the Alzheimer’s Society, who is funding the study and with whom the PhD Clinical Training Fellow and supervisors have direct contact.

The dissemination of results plans to benefit the provision of healthcare and adult social care by providing evidence to inform change in clinical practice, policy and service design. This study has the opportunity to affect every person dying with dementia, as well as family, carers and healthcare professionals involved in providing their end of life care. It is hoped the outputs will provide robust evidence to demonstrate unmet need among this vulnerable population and the factors associated with poorer quality of end of life care. Wide dissemination through open-access resources and diverse communication channels is therefore considered to be in the public interest; not only for its relevance to a large population of people affected by dementia but also for the potential impact on health and social care provision and service design.

The benefits of the dissemination plan include targeting and reaching the most relevant individuals. Between August 2021 and April 2022, papers and abstracts will be sent to journals and conferences in palliative care and psychiatry. In September 2021, policy briefs and results summaries will be shared with relevant charities and third sector groups. The outputs from this project will hopefully show, for the first time as far is known, what factors are associated with ED attendance at the end of life for people with dementia across England, demonstrating whether there are differences and inequalities between patient profiles. This evidence is needed for change in the provision of care for this patient group.

The benefits will be measured in several ways. First, the reach of the results will be measured by summarising and evaluating the dissemination of the work, including readership of the journal in which the results are published, number of downloads, online readers and retweets. The anticipated benefits of increased awareness among clinicians and carers will also be measured by proxy through discussion between partners and collaboration between palliative care and other relevant specialities such as geriatrics and old age psychiatry.

The PhD Clinical Training Fellow and senior clinical academic supervisors will be responsible for disseminating results to maximise impact. By sharing the predisposing and mediating factors of ED attendance identified from analysis, it is hoped that policy makers, commissioners, and clinicians will be better informed to make strategic, service and clinical end of life decisions, enhancing the quality of end of life care for people with dementia.

It is hoped it will be possible to measure direct benefit for patients when changes are made to reflect the increased awareness of the need for considered ED attendance at the end of life for people with dementia, although this is likely to take time. It is hoped these changes will include fewer ED attendances at the end of life by people with dementia. In future years, King's College London will be able to perform time-trend analysis to explore if outcomes have improved over time.

Yielded Benefits is not a requirement for new applications.