Investigation of the association between different forms of healthcare support for care home residents and both hospital admissions and place of death

This agreement from the University of Manchester (UoM) relates to a National Institute for Health Research School for Social Care Research (NIHR SSCR) funded study entitled ‘Effective Healthcare Support to Care Homes’. By this the UoM mean the support that visiting healthcare professionals such as GPs, district nurses, geriatricians, pharmacists, palliative care staff and specialist care home support teams provide to meet the physical healthcare needs of older, long-term care home residents. The research is being undertaken by staff from Social Care and Society at Manchester (SCS) formerly the Personal Social Services Research Unit (PSSRU) Manchester branch. For the purpose of this agreement data will only be accessed by substantive employees of UoM and record level data will only be stored and processed at the UoM.

In England, over 375,000 older people live in care homes. These individuals have multiple and complex needs, many of which stem from progressive, chronic conditions including neuro-degenerative, musculoskeletal and cardio-respiratory disease. Visual and hearing deficits are common; about 40% of residents are depressed; and the majority of residents have dementia. Median life expectancy in care homes is just 15 months.

Surveys suggest that the healthcare support that care homes receive varies across the country and that the majority of care homes do not have access to all the services and support that they require. It can thus be difficult for staff to obtain timely and appropriate diagnosis and treatment for their residents. In consequence, care home residents have high rates of potentially avoidable admissions to hospital, with significant implications for both NHS costs and residents’ well-being. At present, around a third of residents with unplanned admissions to hospital die during their stay. A report by the National Audit Office suggests that many of these people could be supported at their care home if appropriate services were available.

International evidence indicates that specialist assessment and management of care home residents (including enhanced primary care services) can improve resident outcomes and save money. Nevertheless, although a range of new models of care home support are emerging, little is known about the most effective ways to structure and deliver this input. This includes evidence on the precise organisational forms, staff mix and clinical processes associated with reducing avoidable hospital admissions and enabling residents to remain in the care home at the end of life. Against this background, this study will investigate the association between different forms of healthcare support provided for long-term care home residents and both unplanned (emergency) hospital admissions and place of death.

A 2017/18 survey of care homes in Greater Manchester undertaken by the PSSRU has already collected information on the healthcare support that these facilities receive and subsequent analysis has identified seven different ways (Care Home Support Models 1 to 7) of classifying this. This has enabled the research team to categorise the care homes into subgroups of homes receiving similar forms of support. For example, one model groups the homes according to the mix of professionals that provide support, another according to the frequency of the input they receive and a third according to the nature of the support that they receive and whether this includes care home staff training as well as advice on individual care home residents. This strand of the research did not involve any NHS Digital data.

The data under this Agreement is limited to HES Admitted Patient Care data (6% of the available HES fields have been requested) and Civil Registration (deaths) data restricted to Greater Manchester care home postcodes with data provided on patients aged 75+. HES APC data are restricted to admissions occurring between 01/04/2017 and 31/03/2018. Civil Registration (deaths) data are restricted to records of patients who died between 01/04/2017 and 31/03/2018. No individual patient identifying information is provided. The data is pseudonymised at record level.

The restrictions imposed on the data ensure the data relate to the same cohort and time period as the data on which the Care Home Support Models were predicated (i.e. from the 2017/18 survey of care homes in Greater Manchester referenced above).

The information from NHS Digital will be used to explore the extent to which these different models of support are associated with the unplanned hospital admission and place of death of care home residents (see processing activities below for further details). This data will enable the researchers to move beyond a simple description of the patterns of service delivery to answer the following research question: ‘What are the outcomes of different arrangements for healthcare support to care homes?’ There are no alternative, less intrusive, ways of achieving this purpose.

The justification of processing this data comes under Article 6 (1)(e); processing is necessary for UoM to perform the task in the public interest as stated in the benefits section. In addition, the data is being processed under Article 9(2)(j); archiving, research and statistics. As stated above, the process is necessary for research and only the minimal amount of data has been processed. Further to this UoM have addressed the safeguards set out in Article 89(1) of the GDPR and in section 19 of the DPA 2018.

The UoM is the sole organisation involved. UoM is the sole Data Controller and is the only organisation which processes the data.

Only substantive employees of the UoM who have been appropriately trained in data protection and confidentiality will have access to the data, and only for the purposes described. The data will not be shared with third parties or linked to any other datasets.

Below is a summary of the dataflow and processes;

1. The UoM sent NHS Digital a spreadsheet containing the names and postcodes of the 418 care homes in Greater Manchester along with the first line of their addresses. NHS Digital ran that information through Personal Demographic Service (PDS) and extracted the NHS numbers and used these to identify people at each of these addresses. They then added the seven variables classifying the external healthcare support received by their care home (Care Home Support Models 1 to 7) and eight further variables relating to the care home type (Care Home Type 1 to 8) and appended these to the data. NHS Digital did not disseminate PDS data, they used it purely to identify any individual living at the care home addresses over the age of 75. No individual care home resident identifying information was provided.

2. NHS Digital linked the requested HES data and Mortality data to the provided postcodes, Care Home Support Models 1 to 7 and Care Home Type variables 1 to 8.

3. NHS Digital removed the postcodes and first line of address from the linked data before delivering this to the UoM. No individual patient identifying information was provided. The data was pseudonymised at record level. There will be no attempt to re-identify individuals.

The linked HES and Mortality data will be entered into two statistical software suites [Statistical Package for the Social Sciences (SPSS) and Statistical Analysis System (SAS)] and a range of descriptive, bivariate and multivariate statistics as well as generalised linear models will be employed. The analyses will profile the full sample of admissions from care homes in Greater Manchester and will explore differences in unplanned hospital admissions and place of death between residents of care homes receiving different models of support, controlling for care home size.

All outputs will only contain aggregate level data and small numbers will be suppressed in line with Hospital Episode Statistics (HES) analysis guidance.

Data is stored in the UoM Data Safe Haven (Joule House), which has been built following ISO27001 standards and NHS Digital security requirements. The Data Safe Haven is located on a separate premise (Joule House) to that where the data is accessed and analysed (Crawford House).

The workstations used for accessing the University’s Data Safe Haven (DSH) environment do not directly access the data. Access is via Virtual Desktop Infrastructure (VDI) technology to ensure the data is only processed within, and never leaves the DSH virtual environment. Due to this ‘painted screen’ approach, no remnants of the data are ever stored on the client device through mechanisms such as temp files or browser caches. Access is restricted to on-site computers only. No mobile devices will be used to access the DSH environment

NOTE: There is currently an exception allowing staff working from home during the Covid 19 pandemic to access the DSH remotely. This has been agreed with NHS Digital for projects that use the UoM Data Safe Haven, with the following approach being taken:

Method 1: Via University of Manchester provided laptops

Researchers using a UoM provided laptop logon to the UoM VPN, and it is from there that they authenticate via 2FA to the Data Safe Haven. Therefore, access to the DSH is via University infrastructure and within the UoM estate (on-site). The only change is the physical environment from which the researcher will be working – their homes rather than their offices.

Method 2

Researchers remote VPN onto their University of Manchester staff desktop (workstation), and it is from that workstation that they authenticate via 2FA to the Data Safe Haven. Therefore, access to the Data Safe Haven is via University infrastructure and within the UoM estate (on-site). The only change is the physical environment from which the researcher will be working – their homes rather than their offices.

In both methods, the laptop and the workstations used for accessing the DSH service do not directly access the data. Access is via Virtual Desktop Infrastructure (VDI) technology to ensure the data is only processed within, and never leaves the DSH virtual environment. Due to this ‘painted screen’ approach, no remnants of the data are ever stored on the client device through mechanisms such as temp files or browser caches.

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).

The production of outputs utilising the requested data will begin soon after data receipt and will continue for up to five years. In all cases, all outputs will be at aggregated level and small numbers will be suppressed in line with NHS Digital guidelines. Specific outputs are described below.

A final report on the wider study (including brief details of the methodology, findings and conclusions of this analysis) was submitted to the study’s funders (the NIHR SSCR) in the spring of 2019. Following peer review, a publicly available version of this report will be posted on the websites of the funder (https://www.sscr.nihr.ac.uk/) and the University (https://sites.manchester.ac.uk/sscr/).

Given the nature of the final report (a short summary document) further analysis of the data will be reported in two academic papers detailing the association between the different models of care home support and the hospital admission and place of death of care home residents respectively. The main audience for these papers will comprise commissioners, managers and providers (locally, nationally and internationally) as well as other academics. These papers will thus be submitted to peer-reviewed journals such as Age and Ageing, Primary Healthcare Research and Development and the Journal of the American Medical Directors Association (JAMDA) and will be available via open access where possible. Target publication date: Spring/Summer 2021.

Early findings have been presented at a number of local care home owner, manager and staff forums and the final results will be presented at a national conference such as the SSCRs Annual Conference. Target presentation date: Spring 2021.

The applicants will also produce a short, evidence-based briefing document on the study’s findings, available free of charge as a web-based document. This will form part of the SCS at Manchester Expert Briefing series which is designed to inform service development, both nationally and locally. As such it is expected to be of interest to personnel who commission and deliver healthcare support to care homes, as well as care home providers. Target publication date: Spring 2021.

In addition, a SCS Research and Policy Update summary document detailing the key research findings will be distributed to clinical commissioning groups and local authorities in England, CQC, major care home providers and umbrella organisations for care homes through national mailing lists. Target publication date: Summer 2021.

All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

The study is designed to increase understanding of the extent to which different models of care home support are associated with better resident outcomes. In particular, it will explore differences in the unplanned admission to hospital and place of death of residents of care homes receiving different forms of support, controlling for care home size. As such it will add to the limited existing knowledge base about models of support for care homes, and will yield information with direct implications for:

i/ Health and social care planners and commissioners seeking to promote the delivery of timely and appropriate healthcare services for care home residents and the more efficient use of acute hospital beds.

ii/ Managers and senior staff in provider organisations, enabling them to identify gaps in access to specialist care and negotiate improvements.

iii/ Front line care home staff, care home managers and members of healthcare support services by identifying those care home residents at greatest risk of inappropriate hospitalisation and facilitating the planning of anticipatory care.

iv/ Care home residents, who may achieve increased well-being through the provision of more appropriate health care.

In addition, to shape the influence of this study upon the way care services are delivered, further subsequent research funding will be sought to enable the research team to explore changes in service arrangements and care home resident outcomes over the next five years. These findings will also identify locations where care outcomes for residents are particularly good, highlighting where more in-depth case studies can be undertaken to further identify key service components associated with better quality and safety. These are the elements needed for commissioners to arrange better care for care home residents.

The anticipated dates of the publication of the analysis and plans for its wider dissemination are given in the previous section. The fact that Manchester is likely to have a more diverse range of models than some other places is designed as a strength of the study. Furthermore, the findings are likely to be of additional interest locally in the context of aims to reduce the number of hospital admissions by 60,000 per year. The research findings will contribute to that debate.

Presentations on the extent to which different models of care home support are associated with better resident outcomes were made to local care home owner, manager and staff forums in 2019 with a view to informing service planning. The forthcoming outputs will help to further achieve this goal.