Protocol for VIVALDI social care: Pilot study to reduce infections, outbreaks and antimicrobial resistance in care homes for older adults

Patient and Public Involvement

This proposal has been informed by extensive engagement with care home residents, their relatives, care home staff, and providers including presentations at Care conferences and events held by the CQC, press releases, magazine articles, newsletters, webinars, and podcasts. Interactions have strongly informed the study design and specifically decisions to 1) seek section 251 support to access data without consent from residents and 2) focus the pilot study on care providers that use DSCR (due to the prohibitive workload associated with extracting information from paper-based systems).

Since October 2022, UCL, Care England and the OS have established two working groups (Engagement & Communications; Data Sharing & Information Governance) with representation from care providers, care home staff, and relatives of people who live in care homes. These groups considered how to engage and communicate with care homes, and issues related to data sharing and information governance.

The project team have visited care homes across England to discuss the project and the problem of infections in care homes. In addition, the study team have partnered with researchers from the DACHA study to develop VIVALDI Social Care ‘activity packs’, which are delivered by activity providers in participating care homes and provide a non-onerous mechanism to engage residents in the study and seek their views. A new collaboration has been established with the charity ‘Care Rights UK’ (https://www.carerightsuk.org/) who represent the views of residents and relatives. The study has been discussed with eight members of this organisation and they have also had input into the design of information leaflets and posters. A webinar for members of this organisation was also held to gather views regarding the use of data from residents without consent. The Engagement & Communications workstream have worked with an animator to co-produce an animation to accessibly explain the concept of data sharing within the VIVALDI Social Care study.

The care home working groups, and Care Rights UK will continue to play a key role in the delivery of this project within the Stakeholder Oversight & Governance Group, Engagement Collective, and Data Access Committee. They will work with the research team to interpret results and advise on dissemination of findings to people who live and work in care homes.

Full details of our activities to date and those that will take place during the study are provided in the Appendix. Video recordings from engagement events are available to watch online: https://theoutstandingsociety.co.uk/case-study/vivaldi-care-home-study/.

Study design

Research database, prospective open cohort study.

Settings

Care Quality Commission (CQC) registered care homes that provide nursing or residential care for older people (aged >=65 years) in England and use DSCR.

Participants

Residents in participating care homes.

Inclusion criteria

Residents who have not opted out of data collection.

Exclusion criteria

Residents of care homes that primarily provide care to adults <65 years, although residents who are younger than 65 years who reside in care homes that primarily provide care to older adults will be eligible for inclusion.

Timelines

Study recruitment will be contingent on appropriate ethical approvals and is anticipated to commence in January 2024. Data collection will commence following the initial onboarding period (approximately two months) and will continue for a maximum of twelve months. The database will be completed and locked in March 2025 and securely stored for ten years. Data analysis will be undertaken during the study period, where appropriate, or after data collection has ended.

Aims & objectives

Aim: To test the feasibility of establishing sentinel surveillance and a research database to facilitate large-scale observational research and future interventional studies to reduce infection, outbreaks, and AMR in care homes for older adults in England.

Primary objective

To establish a care home network, underpinned by linked data to enable sentinel surveillance and research on infection and AMR in older adult residents of 500–1500 care homes in England.

Secondary objectives

Outcomes

Primary outcomes

The establishment of sentinel surveillance and a research database which enables observational and future interventional studies on infection in older adult residents of care homes in England.

Secondary outcomes

Consent

A key feature of the project is its goal to be inclusive, so every resident has the opportunity to take part, to minimise bias and so that findings are relevant for all older adults in care homes. It is also important that the database is accurate, up-to-date, and complete, for studies which aim to measure the burden of infection and related outcomes. This is because infections spread between people, and short-stay residents (who often get excluded from research and studies that are based on routinely collected data) are more likely to bring infections into care homes having recently been in hospital.

An estimated 70% of care home residents live with cognitive impairment¹⁶ and may not be able to provide informed consent to study participation. Whilst it is possible to identify consultees to support the consent process in those who lack capacity; this is time-consuming and requires dedicated staff. Severe staff shortages across the care sector and lack of embedded research infrastructure mean it is not feasible for staff to seek consent from every care home resident. Consequently, to develop an inclusive study and for data to be generalisable, the study has sought Section 251 support from Health Research Authority Confidentiality Advisory Group (HRA CAG) to collect data from all residents without consent, with the option for residents to opt-out of data sharing. In large care provider organisations, the initial decision to take part will be taken by their senior management team and home managers will then be asked to confirm if their care home is willing to take part. In homes that are not part of larger organisations, the decision to take part will be made by the care home manager.

Care home engagement and recruitment

To recruit a wide range of different types of care providers and minimise selection bias, researchers from UCL have been working with Care England and the Outstanding Society since October 2022, to raise awareness of the VIVALDI Social Care Project through presentations at care conferences and events held by the CQC, press releases, magazine articles, newsletters, webinars, organisations’ websites and podcasts (https://theoutstandingsociety.co.uk/case-study/vivaldi-care-home-study/). As of May 2023, over 800 care providers in England have expressed an interest in taking part in the study.

When the study begins, the research team will distribute multi-modal study materials to care homes that provide information at three levels: 1) basic information about the study including how to opt out of data collection, 2) detailed information about the study, and 3) very detailed information including study design, explanation of the data flows and data controllership. These materials have been co-developed with the study’s Engagement Working Group which includes relatives, care home staff, and providers, and piloted with a subset of care homes prior to study start.

Data collection & management

Engagement work with the care sector has made it clear that it is not feasible to extract data from paper-based systems. Following extensive discussions with care providers and the Digital Care Hub (formerly Digital Social Care), a sector led organisation which provides free support to care providers on technology, data protection, and cyber security (www.digitalcarehub.co.uk), it has been agreed to start by working with providers that use DSCR that have been assured by the NHS Transformation Directorate in NHS England (NHSE)¹⁷.

To maintain an accurate record of which residents are in a care home at any point in time, data for each resident will be extracted directly from DSCR rather than from the care homes. This will include residents’ NHS numbers, the date of data collection and the unique care home identifier allocated by the Care Quality Commission (CQC ID) at regular intervals (residents who have opted out of data collection will be excluded). This will be achieved by establishing a permanent daily data feed from the DSCR supplier to a secure environment within NHSE, allowing the dataset to be updated in near real-time. Although this may impact on the generalisability of the sample for the pilot study, the Government has strongly signalled its expectation that all care providers will move to electronic care records (80% of CQC registered providers to have DSCR by March 2024)¹⁸. Currently over 60% of providers use DSCR¹⁹ suggesting it will become increasingly feasible to recruit a diverse set of care homes that use DSCR over time, if the study is continued beyond the pilot. To reduce the administrative burden associated with participating, a set of template agreements (data sharing agreements, data privacy impact assessment, privacy notice) will be developed in partnership with Digital Care Hub and key stakeholders for use by participating providers.

The linked data platform will be established in a secure data environment within NHSE, such as NHS Foundry²⁰, with NHSE acting as data processors on behalf of the project. NHSE will first pseudonymise NHS numbers from residents (provided by the DSCR suppliers) using an automated algorithm and then allocate each record a common pseudo-identifier, based on NHS number. This can link records from residents to routinely collected datasets already held in this secure data environment (hospital admissions, vaccinations, emergency admissions, microbiology and virology results, antimicrobial prescribing, mortality), Figure 1, Table 1. Set-up and curation of the linked data platform in the secure data environment will be led by a software engineer who oversees the VIVALDI data platform.

Figure 1. Data flow diagram.

Diagram illustrating flow of data from providers to the study dataset. Red arrows represent confidential data and grey arrows demonstrate movement of non-confidential idea that has been pseudonymised / de-personalised. Research database will be created in secure data platform such as NHS Foundry or the Federated Data Platform (FDP). AGEM Arden & GEM; UDAL Unified Data Access Layer; SGSS Second Generation Surveillance System; UKHSA UK Health Security Agency; CQC Care Quality Commission; NHS National Health Service; UCL University College London; NHSE FDP Federated Data Platform. *Data sharing following approval from the relevant care provider.

Pseudo-identifiers will be removed before the de-personalised linked data extracts are transferred securely to UCL and stored in the UCL Data Safe Haven (DSH) (https://www.ucl.ac.uk/isd/services/file-storage-sharing/data-safe-haven-dsh). The DSH has been certified to the ISO27001 Information Security standard and conforms to NHS Digital's Data Security and Protection Toolkit. It is built using a walled garden approach, where the data is stored, processed, and managed within the security of the system, avoiding the complexity of assured endpoint encryption. A file transfer mechanism enables information to be transferred into the walled garden simply and securely. Only named researchers within the research team have access to the research database within the DSH. All researchers seeking to use the DSH are required to complete regular training in information governance.

All residents will be given the opportunity to opt-out of data sharing by speaking to any member of care staff. They will also be able to notify the care home in writing that they do not wish to participate. The decision to opt out can also be made on behalf of a resident in accordance with their wishes, for example by their next of kin and/or a consultee, in line with the Mental Capacity Act 2005 and National Institute of Clinical Excellence (NICE) guidelines^21,22. Lists of individuals who have completed either the NHS opt out²³ or the project specific opt out will be regularly updated and shared with DSCR suppliers who will remove these individuals from the data feed with NHSE. Information materials (posters, leaflets, information sheets, animation) will be distributed widely through participating care homes with details of how to opt-out.

Data analysis & statistical plan

The sample size reflects the desire to generate findings that are generalisable across the care sector. In parallel it also acknowledges that there are considerable technical, logistical and governance challenges associated with setting up this study (collection of individual-level data from care homes at scale has not previously been attempted in England).

As there are approximately 11,000 care homes for older adults in England¹, if the target sample size (500–1500 care homes) is achieved, the sample would represent approximately 5–14% of all care homes. However, there is considerable diversity across the care sector, so it is very unlikely to be representative of all the different types of care homes. If the study is successful, and there is a desire to establish a long-term surveillance and linked data infrastructure in care homes, the intention is to work with the CQC and the UKHSA to understand which types of care homes are not included and to try and address these gaps by targeted recruitment. Given the complexity of the data flows and governance, the relatively short timescales, and the absence of a care home registry from which to sample, it is not feasible to develop a sampling frame and attempt to recruit a representative group of care homes in this pilot study.

As the goal of this project is to explore the feasibility of establishing sentinel surveillance and a research database enabling observational and future interventional studies, a Statistical Analysis Plan (SAP) has not been developed. All projects using the research database will be expected to develop a SAP before data analysis commences which are specific to the research question and will detail the sample selection, outcomes, exposures, how confounders will be accounted for, and how bias will be minimised.

Study oversight

Oversight for the study is outlined in Figure 2. The project Steering Committee will provide oversight of the study on behalf of the Sponsor (UCL) and Funder (UKHSA) and will be chaired by an experienced academic in social care policy who is not involved in study management. The Stakeholder Oversight and Governance Group will act on behalf of residents and families to oversee how data are used in the study and study outputs. This group will include representation from residents, relatives, staff, and providers and experts in information governance from care providers, UCL, UKHSA, and NHSE. It will meet two to three times per year. It will provide reassurance to care providers, residents, relatives, and staff that they maintain control of how their data will be used in the project and is a critical part of building trust between care home stakeholders and UKHSA to enable future studies and continued data sharing.

Figure 2. Study oversight diagram.

Flow diagram representing the relationships between groups overseeing the project. Colouring of boxes and arrow demonstrates the role of each group: green represents oversight, orange is operational, and blue is advisory. DAC Data Access Committee; SOGG Stakeholder Oversight and Governance Group

In parallel, a Data Access Committee will specifically oversee the use of the de-personalised, pseudonymised research database in observational / modelling research studies. This group will include representation from residents, relatives, staff, academics, and providers and will be responsible for reviewing and approving proposals from researchers who wish to use the linked dataset.

The Adult Social Care Engagement Collective will have an advisory role feeding directly into the oversight and operational groups, Figure 2. They will ensure the views of residents and relatives are represented throughout the project. This will consist of representatives from the care sector including providers, care managers and more junior staff; residents and their relatives and third sector organisations specialising in data and the rights of people accessing care.

Ethics

Ethical approval to set up and use the anonymised research database for observational / modelling studies on infection and AMR in care home residents has been obtained from the South-West - Frenchay Research Ethics Committee (23/SW/0105) on 5^th December 2023 however all applications to use the database must first obtain approval from the study Data Access Committee. Use of the database in future interventional studies would require additional ethical approvals.

Approval to collect data from residents without informed consent for the purposes of sentinel surveillance and research on AMR has been obtained from the Health Research Authority following review by the Confidentiality Advisory Group (CAG) under Section 251 of the NHS Act 2006. Two applications have been approved enabling use of the data for research (23/CAG/0134) and surveillance (23/CAG/0135).

Dissemination plans

Dissemination of outputs to key stakeholders will use the engagement workstreams outlined above. Results of analyses conducted using the VIVALDI Database will be published in peer-reviewed academic journals in a timely manner and posted on pre-print sites in cases where results are required to rapidly inform policy decisions. Results will also be presented to policymakers, academics, clinicians, and providers at national care conferences, academic infection conferences, and regular meetings with policymakers within UKHSA, Department of Health & Social Care (DHSC) & NHSE. Dashboards and surveillance reports will be accessible to key stakeholders including care providers and public health officials within UKHSA and policymakers in DHSC.

Study outcomes will be shared with participants in an end of study report. This will be developed in collaboration with the engagement workstream, to ensure that the results are clear and easy to understand for the care home population, and to determine the best format for this e.g., video. This will be published on the study website and this link will be widely disseminated so that family members and the wider care population are able to access it.