Understanding the role of adult community health services in avoiding hospital admissions

The University of Manchester requires extracts of Hospital Episode Statistics (HES) and the Community Services Data Set (CSDS) linked to date of death for the purpose of undertaking the research project “Understanding the role of adult community health services in avoiding hospital admissions”. The aim of this research project is to examine how rising demand and limited investment in adult community health services has influenced their provision and affected hospital admissions, how this varies by ethnicity, deprivation, and rurality, and how to develop the economic case for investment.

This data can be lawfully processed for this purpose as it is a task carried out in the public interest (GDPR Article 6 (1) (e)). The University of Manchester are a public authority conducting research that will influence the provision of healthcare services in the community, directly benefiting patients who use these services. The research project was funded as the result of a call put out by the National Institute for Health and Care Research (NIHR) requesting research on this area, because of a need for evidence on how best to organise and deliver community health services and better understand their role in avoiding hospital admissions. Community health services play a crucial role in keeping people well, treating and managing acute illness and long-term conditions, and supporting people to live independently in their own homes. Despite their vital role, community health services are relatively under-researched. Lack of data and evidence has hindered much needed improvements in these services. Community health services account for approximately £10bn of the annual NHS budget, but little is known about the care delivered, its quality and outcomes, or value for money (Charles et al., 2018). It is in the public interest to ensure that the NHS budget is well spent, and the evidence produced by this research will inform future spending decisions. Community services deliver 100 million patient contacts annually (Charles et al., 2018), representing 100 million opportunities for evidence to improve patient care each year. Producing this evidence, which this project will do, is in the public interest as it will be used to inform improvements in community health services delivery which will in turn improve outcomes for patients treated by these services. The University of Manchester project team engaged with members of the public and people with lived experience of community health services (themselves or as carers) when preparing the research project application and protocol, to ensure that the evidence produced will be of value to the public. Members of the public with whom the University of Manchester project team spoke confirmed the need for this evidence, voicing just how important it was to improve community health services, and helped to design the research project to ensure that the research will meet their evidence needs.

Processing of the data is necessary for scientific research purposes and will be used to undertake the research project “Understanding the role of adult community health services in avoiding hospital admissions” (GDPR Article 9 (2) (j)). This research project was funded by the NIHR on behalf of the Department of Health and Social Care (DHSC). The NIHR put out a call specifically requesting research on this topic because of an identified need for research evidence to improve the quality, accessibility and organisation of community health and social care services in order to benefit patients. This research project was funded under this commissioned call. Processing of this data is necessary in order to meet this evidence need.

The research will produce much-needed evidence on how to organise and deliver community health services to best meet population needs, and on how community provision impacts hospital use. Dissemination of this research does not pose a risk of potential harm to the public. Dissemination will be of societal benefit.

The data requested are from HES which contains information on hospital utilisation, and the CSDS which contains information on community health services utilisation, linked to date of death information. The University of Manchester project team will examine how use of community health services impacts use of hospital services at the patient-level and at the Local Authority level.

The data will be used for this research project only (NIHR134436 - Understanding the role of adult community health services in avoiding hospital admissions).

The original data subjects are all individuals in the HES dataset aged 18 and over from 01/04/2010 to 31/03/2020 and all individuals in the CSDS aged 18 and over from 01/10/2017 to 31/03/2020. Amendment request extends this to additional data subjects using hospital and/or community health services in a post-Covid-pandemic period: The data subjects for the amendment are all individual in the HES dataset aged 18 and over from 01/04/2022 to 31/03/2023 and all individuals in the CSDS aged 18 and over from 01/04/2022 to 31/03/2023.

The data will be processed for use in three of the five work packages (WPs) in the overall research project. These are:

• WP2: Workforce and supply - the aim is to examine how supply of adult community health services has responded to rising demand (uses CDSD only)

• WP4: Impact on hospital use - the aim is to examine the impact of adult community health services on hospital utilisation (uses linked HES and CSDS)

• WP5: Economic case for investment - the aim is to examine the costs and benefits of adult community health services (uses linked HES and CSDS)

The remaining two WPs of the project, listed below, will not access the data requested here but are described for completeness:

• WP1: Demand - the aim is to examine how the drivers of demand for community health services have changed over time in each local authority in England (uses only publicly available data on population estimates and national population surveys)

• WP3: Commissioning and planning - the aim is to understand how commissioners make decisions about investment in community services, and how providers allocate their resources (uses qualitative interview data that will be collected during the project)

It is not possible to make deductions about the impact of community services on patients who receive those services based on analysis of aggregate data alone. Individual-level pseudonymised data is therefore required in order to achieve the project’s aim. All data will be pseudonymised and linked using common pseudo-IDs applied by NHS Digital.

Data on patient ethnicity and lower-layer super output area (LSOA) of residence are required in order to examine whether there are inequalities in healthcare provision and outcomes by ethnicity or deprivation. Data on patient age, sex, primary and secondary diagnosis are required in order to correctly risk-adjust for differences in patient population composition between providers in the analyses. Date of death is required for the purpose of undertaking the research project “Understanding the role of adult community health services in avoiding hospital admissions”. In particular, date of death is required to undertake WP4 of the project. WP4 aims to examine the impact of adult community health services on hospital utilisation. Data on date of death is required in order to correctly adjust for survival rates following initial admission when analysing the impact of community health services on hospital readmissions, as per the recommendations of Laudicella et al. (2013). Similarly, date of death is also required in order to correctly adjust for survival rates following an initial contact with community health services when analysing the pattern of utilisation of community health services following an initial referral. All other requested fields are required to analyse the patterns of hospital and community health service activity in detail to provide evidence on the relationship between the two.

It is not possible to achieve the project's purpose with fewer data sets. Date of death is requested from Civil Registration (Deaths) - Secondary Care Cut. The request has been minimised by requesting only one field from this data set, which contains the minimum amount of information necessary to undertake the research project. This file will be linkable to HES Admitted Patient Care, HES Accident & Emergency, HES outpatient, and CSDS data via pseudo-IDs. The level of intrusiveness has been minimised by requesting only pseudonymised and ‘general’ data in these data sets. These pseudo-IDs will be specific to the project, and only NHS Digital will hold the personal information upon which these pseudo-IDs are based. No high risk or identifiable data are requested in the data sets to which the mortality data will link. The University of Manchester confirm that it is not possible to achieve the purpose in a less intrusive way.

The original HES data covering 10 financial years is required for the time-series analyses in WP4. Any initial cross-sectional associations between community health services supply and hospital utilisation detected may be biased by factors which cannot be measured in the available data. Longitudinal analysis is therefore required which focuses on changes over time, and so data covering a 10 year period is required to produce unbiased estimates. The amended HES data covering the 2022/23 financial year is required for WP4 and WP5 to see how the cross-sectional relationships detected before the Covid-19 pandemic have changed. This evidence is needed to examine how evidence from health services research undertaken prior to the pandemic is relevant for future healthcare provision decisions.

The CSDS is available from October 2017 onwards. CSDS data covering the period from this start point until the same end date as the HES data is therefore required (31/03/2020 for the original request, CSDS covering the period 01/04/2022 to 31/03/2023 for the amendment).

The research is hoped to inform national service organisation and policy making. National HES and CSDS is therefore required in order to produce projections for each service provider and local authority in England, and to achieve the overall project aim. It would not be possible to implement the approved study design without access to national data. For example, the aggregate analysis in WP4 will examine whether the mismatch between population demands and community health service supply is associated with hospital activity at the local authority level. National HES data is therefore required to measure total hospital use in each local authority in England, and investigate whether this is associated with total community health services activity within each local authority in England as measured by the national CSDS.

National data on community health services activity and hospital utilisation is therefore required to achieve the overall project aim. The University of Manchester confirm that there are no alternative, less intrusive ways of achieving the purpose.

The University of Manchester is the sole data controller. The Principal Investigator leading the research project is a substantive employee of the University of Manchester.

The University of Manchester is the sole organisation processing the data. No other organisations will process the data.

Four other organisations are involved in the wider research project (London School of Economics and Political Science, London School of Hygiene and Tropical Medicine, Derbyshire Community Health Services Trust, Derbyshire Healthcare NHS Foundation Trust), but will have no access to the data. All of the research will be undertaken by the University of Manchester and its substantive employees. The roles of individuals employed by the four other organisations in the project are purely advisory. In particular, two members of the project team are employed by the London School of Economics and Political Science. These two members will provide advice on how the University of Manchester project team can best consider the interface between health and social care in the research. They will provide advice on publicly available data capturing the availability and use of social care services in each local authority in England, which the University of Manchester project team will account for in their analyses. One member of the team is employed by London School of Hygiene and Tropical Medicine. They are a qualitative researcher, and are involved in the qualitative work package of the project (WP3: commissioning and planning). They are not providing input into the quantitative work packages which use the data requested here. One member of the project team is employed at Derbyshire Community Health Services Trust. They are providing advice on the needs of evidence users, specifically focusing on how the University of Manchester project team can best communicate the project results to people who make decisions about how to plan and organise community health services. Finally, one member of the project team is employed by Derbyshire Healthcare NHS Foundation Trust. They will provide advice on the dissemination strategy and pathways to impact. All of these individuals employed by these four organisations will provide advice throughout the project, but all final decisions will be made by the project’s Principal Investigator who is employed by the University of Manchester. As such University of Manchester are the sole controller who makes decisions about the means and purpose for which the personal data are being processed.

The data requested will not be used or accessible to other work packages within the project beyond those mentioned in this data request (work packages 2, 4, and 5). The aggregated findings (aggregated to Local Authority District level) will be used to inform work package 3, but the data will not be accessed or processed by this work package. Specifically, work package 3 will use qualitative research methods, interviewing commissioners and providers of community health services in four areas of England. The outputs of work package 2 will be used to inform which areas of England these interviews are conducted in.

The data is restricted to only adults (aged 18 and over) who have used hospital and/or community health services. Data has been further minimised by only selecting data fields required to undertake the proposed analysis.

References:

Charles A, Ham C, Baird B, Alderwick H, Bennet L. (2018). Community health services explained. London: The King’s Fund.

Available at: https://www.kingsfund.org.uk/publications/community-health-services-explained

Laudicella et al. (2013). Hospital readmission rates: signal of failure or success? Journal of Health Economics, 32(5), 909-921.

No data will flow into NHS Digital.

Pseudonymised (CSDS/HES/Civil Registrations deaths) data will flow from NHS Digital to the University of Manchester. The data will be stored in a Common Internet File System (CIFS) share accessible only from a specified workstation, via a mapped network drive. This specified workstation will be a University of Manchester managed device, complying with the University’s Technical Security Standard. Access control is managed via Active Directory groups, and drives are mapped via the network login script for individual user accounts with relevant group membership. Only members of the research team working on the research project in question will be added to the group Active Directory group (NIHR134436 – Understanding the role of adult community health services in avoiding hospital admissions). Access will therefore be restricted to individuals working on the research project in question, all of whom will be substantive employees of the University of Manchester. Network drives are automatically disconnected when users log off. Together, this means that the data flowing from NHS Digital is stored securely on a share that can only be accessed by pre-authorised members of the project team.

The CIFS share infrastructure is hosted within an ISO27001 certified Equinix data centre located 4 miles from the University of Manchester campus. The data centre is managed by Equinix. However, access to University of Manchester infrastructure within the data centre is physically secured and only accessible by University of Manchester data centre staff. Equinix have no access to the data or infrastructure. All servers, storage and network infrastructure used by the University of Manchester within the data centre is owned by the University of Manchester, operated and maintained by the University of Manchester IT staff and securely caged off. The University of Manchester lease physically secure hosting space from Equinix. No shared services are consumed. Therefore, Equinix is responsible for the physical hosting environment and utilities (standard co-location service, power and cooling).

Data derived from the pseudonymised (CSDS/HES/Civil Registrations deaths) data from NHS Digital is made and stored in a store visible to The University of Manchester’s high performance computing processing platform, where this store is specific to the research project in question and only accessible by the research team that has permission to use the data. For example, derived versions may be aggregated to the hospital level, or contain minimised information on only certain subgroups relevant for a specific part of the analysis. This derived data is stored in the store visible to the high performance computing platform only for the period during which the member of the research project permitted to access the data is logged on to the high performance computing platform to conduct the analysis for which they are responsible. This transfer is secure and encrypted. The high performance computing platform is designed specifically for interactive computationally-intensive work, such as use of applications such as Stata (the statistical analysis software package which will be used by the project research team to analyse the requested data). Access to the high performance computing processing platform is protected by strict project level access controls and multi-factor authentication. Security is by a Unix group Access Control List, restricting access. Only members of the research team working on the research project in question will be added to the group Access Control List (NIHR134436 – Understanding the role of adult community health services in avoiding hospital admissions). Access will therefore be restricted to individuals working on the research project in question, all of whom will be substantive employees of the University of Manchester. Access is via a secure encrypted connection. Account credentials are unique to each member of staff and only the account owner knows the password.

Once processed, the same secure channel will be used to transfer data back to the CIFS where it will be stored. All transfers are encrypted.

This means that extracts from the original pseudonymised NHS Digital data transfer will be temporarily visible to approved members of the research team on a secure platform designed specifically to run computationally intensive statistical analyses. These extracts will be minimised versions of the original data which contain only the information needed to undertake the specific part of the analysis being conducted by the researcher at that time. The approved member of the research team will access the data extract on the secure platform, and undertake the necessary statistical analysis within this secure platform. When they have finished conducting this analysis, the data extract they were temporarily viewing and any further data or information which they have generated will be securely transferred back to the secure share where the original data from NHS Digital is stored. All transfers are secure and fully encrypted. Strict access controls are in place, so only pre-approved members of the research team working on the research project in question will be able to access the data. Members of the research project will have to go through multiple levels of authentication each time they access the data, including individual-specific usernames and passwords, and further authentication through another trusted device (fingerprint or security code entered into an app on the individual’s pre-registered mobile device).

Back-up disaster recovery storage is hosted at Reynolds House within an ISO27001 certified data centre located less than 1 mile from the University of Manchester campus. Reynolds house conforms to the same security standards as the primary storage location (Joule House). The Reynolds House data centre is managed by Ascendas Reit. However, access to University of Manchester infrastructure within the data centre is physically secured and only accessible by University of Manchester data centre staff. Ascendas Reit have no access to the data or infrastructure. All servers, storage and network infrastructure used by the University of Manchester within the data centre is owned by the University of Manchester, operated and maintained by the University of Manchester IT staff and securely caged off. The University of Manchester lease physically secure hosting space from Ascendas Reit. No shared services are consumed. Therefore, Ascendas Reit is responsible for the physical hosting environment and utilities (standard co-location service, power and cooling).

University of Manchester staff are not based on site, however, if required, will be in attendance at Reynolds House (Ascendas Reit). Physical access to the data centre is strictly limited to University of Manchester data centre staff and a limited number of authorised University of Manchester IT services staff. The data centre is protected by physical and electronic access security systems, swipe card access in and out of the data centres and CCTV coverage. The data centre is locked down out of hours and access is discouraged but can be arranged by prior agreement with the data centre manager. The data centre manager is a substantive employee of the University of Manchester.

HES, CSDS and Civil Registrations (deaths) will be linked using common pseudo-IDs applied by NHS Digital, as described above. These pseudo-IDs will be specific to the project, and only NHS Digital will hold the personal information upon which these pseudo-IDs are based (NHS number). The University of Manchester will not have access to the original IDs. Linkage with publicly available area-level information (Local Authority District) will be made in order to control for area-level characteristics in the multivariable regression analyses (for example the average age and sex composition of the population in the Local Authority District within which an individual resides). No individual-level linkage will be undertaken, beyond the HES-CSDS-mortality linkage described above. There will be no attempt to re-identify individuals.

Data processing will only be carried out by substantive employees of The University of Manchester, who is both the sole data processor and the data controller. All individuals processing the data will be appropriately and regularly trained in data protection and confidentiality, and no access to the data will be granted until this training has been confirmed.

No data will be shared with third parties.

University of Manchester researchers granted access to the high performance computing platform share will be able to conduct the analyses proposed above. This will involve loading the data into Stata (a statistical analysis software package) and conducting multivariable regression analyses and sample summary measures. No individual-level output will be made available in any dissemination of the results. Results and findings will be of aggregated volumes and percentage shares across multiple strata.

A description of the planned outputs is provided below. All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

WP2: Workforce and supply

Final report of work package 2 findings will be produced in summer of 2023, reporting estimates of how supply is meeting population demands in each local authority in England and identification of areas experiencing potential issues in patient access. This will highlight areas with potential issues in patient access, unwarranted variation in care provision, and workforce pressures. These findings will be shared with NHS England and Improvement to inform their service planning.

A journal article based on the analysis will be written up for publication, and submitted to a peer-reviewed journal in autumn of 2023. This article will focus on the degree of variation in supply and demand misalignment across the country.

WP4: Impact on hospital use

Final report of work package findings will be produced in May 2024, reporting estimates of the magnitude and mechanisms by which community health service provision affects hospital utilisation. Together these two sets of analyses conducted in this work package will provide a detailed understanding of whether community health service provision does impact individuals’ hospital use, if so how, and whether these impacts are seen at a system level. These findings will be shared with NHS England and Improvement to inform their service planning.

Two journal articles based on this analysis will be written up for publication, and submitted to a peer-reviewed journal in July 2024. The first journal article will report the results of the aggregate-level analysis, demonstrating the impact of community services provision on hospital utilisation at the level of the local system. The second journal article will report the results of the person-level analysis, demonstrating the impact of community services on the hospital use of patients who have contact with those community services.

WP5: Economic case for investment

Report of work package findings will be produced in February 2025, demonstrating the costs and benefits of adult community health services. This will contain evidence on community health service activity configurations to meet current and future population demands. These findings will be shared with NHS England and Improvement to inform their service planning.

A journal article based on the analysis will be written up for publication, and submitted to a peer-reviewed journal in March 2025.

Overall project report

A report summarising the entire research project and its findings will be submitted to the NIHR in May 2025, and published in the NIHR journals library following external peer review.

Wider dissemination

In addition to the above written outputs, the University of Manchester project team will enact their dissemination plan to ensure that the research findings reach the relevant stakeholders outside of academia. The University of Manchester project team will engage provider and commissioner audiences through Integrated Care Systems (ICS) leadership. The University of Manchester project team will also disseminate through additional existing networks such as the NHS R&D Forum, NIHR Applied Research Collaborations (ARCs), Academic Health Science Networks (AHSNs), and the Community Network hosted by the NHS Confederation and NHS Providers which acts as the national voice for community providers. At policy level, the University of Manchester project team we will use their networks to seek engagement of key NHS England and Improvement. The University of Manchester project team are already in regular contact with the community transformation team at NHS England and Improvement, who have expressed great interest in the research. The University of Manchester project team will also disseminate through their established regular contacts with Department for Health and Social Care (DHSC).

The University of Manchester will use press releases to publicise key findings and engage with communications teams to localise the implications for communities and staff groups. Interactive stakeholder engagement events such as webinars will be hosted at regular points throughout the project to avoid overreliance on written outputs. For example, the University of Manchester project team will host a national webinar on key findings with a facilitated expert panel for Q&A and break-out rooms to consider potential system responses to the findings.

The University of Manchester project team will apply to host an organised session at the Health Services Research UK conference to raise awareness in the research community and gain feedback. Interim findings will be presented at conferences throughout the lifetime of the project in order to gain timely engagement and feedback.

The University of Manchester project team will engage with our patient and public involvement (PPI) panel to develop ways of disseminating our research results that are accessible to patients and members of the public. These will include infographics and short lay summaries of the findings of each work package and the overall project findings. These outputs will be published at the same time as the journal articles are submitted for publication (WP2 summary September 2023, WP4 July 2024, WP5 March 2025, overall project findings May 2025).

Community health services account for approximately £10bn of the annual NHS budget and 20% of total NHS workforce (Charles et al., 2018), but little is known about the care delivered, its quality and outcomes, or value for money (Charles et al., 2018; Gershlick and Firth, 2017). A historic lack of data has prevented the monitoring, research, and improvement seen in other sectors. Community health services deliver 100 million patient contacts annually (Charles et al., 2018), representing 100 million opportunities for evidence to improve patient care each year. This research project will fill this evidence gap, resulting in improved patient care.

This research project will aim to produce much-needed evidence on how to organise and deliver community health services to best meet population needs, and on how community provision impacts hospital use. The evidence produced will facilitate improvements in the organisation and provision of community health services, which will ultimately result in improved access and outcomes for patients.

It is hoped that the short- to medium-term impacts within 5 years of study end will include:

• Improved commissioning and provision decisions which better match community health service provision to population demands, achieving better patient outcomes

• Evidence-based workforce planning

• Efficiency improvements in the use of limited healthcare resources across community and secondary care

• Increased use of administrative data in community services to identify and target inequalities in care provision, monitor service quality and support evidence-based decision making

The data requested will not be used for any part of a PhD/ post graduate research study.

References:

Charles A, Ham C, Baird D, Alderwick H, Bennet L. (2018). Community health services explained. London: The King’s Fund. Available at: https://www.kingsfund.org.uk/publications/community-health-services-explained

Gershlick B, Firth Z. (2017) Briefing: Provision of community care: who, what, how much?. London: The Health Foundation. Available at: https://www.health.org.uk/publications/provision-of-community-care-who-what-how-much

This research project has produced the first evidence on the link between the provision of adult community health services and hospital use in England, both at a national level and also at the level of Local Authority District. The research produced to date examines this relationship between 2010 and 2020, before the Covid-19 pandemic changed the way in which healthcare services are organised and delivered.

We have shared this research with key stakeholders in NHS England, as well as stakeholders managing healthcare systems at a local level (for example, Greater Manchester Integrated Care Partnership). These stakeholders are using the research findings to inform their decisions regarding the organisation and provision of adult community health services, meaning that we are on track to achieve the expected measurable benefits listed in the original application within the original target date of 5 years of the study end date.