NHS Digital Data Release Register - reformatted

Cambridge Centre for Health Services Research

Project 1 — DARS-NIC-243359-X4T5M

Opt outs honoured: No - data flow is not identifiable (Does not include the flow of confidential data)

Sensitive: Non Sensitive

When: 2020/05 — 2020/05.

Repeats: One-Off

Legal basis: Health and Social Care Act 2012 – s261(1) and s261(2)(b)(ii)

Categories: Anonymised - ICO code compliant

Datasets:

  • Hospital Episode Statistics Admitted Patient Care
  • Hospital Episode Statistics Accident and Emergency
  • Hospital Episode Statistics Outpatients

Objectives:

At a time of unprecedented financial constraint for health and care, and policy interest in ‘transformation’ of health service delivery, it is inevitable that local organisations will develop plans to pilot and implement new approaches to care. The evidence base for such innovations may be emerging, partial, contested, or plain unknown, and will often be advocated by its planners as having significant potential to impact on indicators such as levels of emergency or avoidable admissions to hospital, and reductions in health care utilisation. Indeed, a service innovation can appear to gain significant traction when lacking a robust (or only having an emerging and tentative) research evidence base, yet being vaunted as ‘what is needed’ to improve the NHS, and can lead to the diversion of management and practitioner time and attention away from more effective and evidence-based activities. Clearly, initiatives should not be assumed automatically to work and maximum learning about what works, why, and in what contexts should be obtained from pilot work before moving to scale. Innovations and developments in health and care services therefore need rigorous, timely and useful evaluation research to be undertaken, so that the risks of such initiatives being assumed ‘to work’ or rolled out in the face of contradictory evidence can be minimised. The BRACE (BRACE stands for Birmingham RAnd [Europe Community Interest Company] and CambridgE) Rapid Evaluation Centre, funded by a 5 year*, 2 million pound grant from National Institute for Health Research (NIHR) HS&DR reference: 16/138/31 (March 2018-March 2023), is a collaboration between the Health Services Management Centre at the University of Birmingham, the independent research consultancy RAND Europe Community Interest Company (CIC), and the Department of Public Health and Primary Care at the University of Cambridge. BRACE will carry out rapid evaluations of promising innovations in the organisation and delivery and heath and care services. The full grant application for this NIHR funding HS&DR reference: 16/138/31 has been uploaded as a supporting document with this application. This work is being carried out under a contract signed between the Secretary of State for Health and the University of Birmingham. RAND Europe Community Interest Company and the University of Cambridge have signed a collaboration agreement with the University of Birmingham. These contracts have been uploaded as supporting documents. The overall timescale of the BRACE rapid evaluation centre is 5 years, and it runs from March 2018 to March 2023. Within BRACE several rapid evaluations will be carried out. The study timescale, design and purpose all contribute to defining a 'Rapid Evaluation' as as follows: • Timescale: e.g. 12 months or less. • Design: Rapid start: getting evaluation project under way quickly; Rapid completion: short timescale from design to dissemination; Rapid cycle: longer evaluation, with early and/or ongoing reporting, on-going learning and feedback of findings. • Purpose: e.g. real-time or ‘alongside’ evaluation, to support innovation development and implementation RAND Europe Community Interest Company (CIC), in collaboration with the University of Birmingham and the University of Cambridge are requesting record-level pseudonymised data from the following four data sets: • Emergency Care Data Set (2019/20 to 2021/22) • HES Admitted Patient Care (2013/14 to 2021/22) • HES A&E (2013/14 up to last data drop of 2019/20 M12) • HES Outpatient (2013/14 to 2021/22) To support the quantitative evaluation work of BRACE these data will be used to allow the team to undertake rapid evaluative studies of emerging service innovations, and to provide high quality evidence about their performance and potential, including when service development should not be pursued. The work of BRACE is guided three overarching principles: 1) Responsiveness. Scoping, designing undertaking and disseminating evaluation research in a manner that is timely and appropriately rapid, pushing at the boundaries of typical research timescales and approaches, and enabling innovation in evaluative practice. 2) Relevance. Working closely with patients, managers, clinicians and health care professionals, and others from health and care in the identification, prioritisation, design, delivery and dissemination of evaluation research, doing this in a truly co-produced and iterative manner. 3) Rigour. Ensuring that all evaluation undertaken by the team is theoretically and methodologically sound, producing highly credible and timely evidence to support planning, action and practice. The data sets supplied through this agreement will be crucial to delivering the work of BRACE. Over the five year period* the BRACE team will identify approximately 11 innovations to evaluate. NIHR, NHS England, Department of Health and Social Care (DHSC) and the BRACE team’s stakeholder advisory board (which includes clinicians, managers, and patient and public representatives) and Health and Social Care Panel will have advisory input into identifying potential innovations. There will be a focus on rapid evaluation of these schemes with some projects aiming to complete evaluation within 6 to 9 months. Others may have a slightly longer timeframe, but the period from inception to analysis will necessarily be short, to allow for timely formative feedback to the areas undertaking innovations. The BRACE team have a remit to produce timely findings of national relevance and immediate use to decision makers. All evaluations will conform to the following scope: - The subject of the evaluation will be a health or care provider (or group of health and/or care providers) which has/have adopted a new way of working which differs to those that have been widely adopted by the health system – i.e. an ‘innovation’. This may be use of new technology or a financial and/or service model redesign. - The HES data would potentially be used to calculate relevant metrics that those implementing the technology would be seeking to affect – for example – short term readmissions to hospital for the targeted groups in those areas. HES data would not necessarily be used in all of the evaluations. It would only be used where it would add value to the evaluation. - The evaluations will attempt to determine whether there were any detectable changes in the metrics over time that might be associated with the introduction of the innovation. - This would be done using a variety of techniques such as comparing metrics for the health care provider(s) pre and post introduction of the innovation or comparing the metrics of the health care provider(s) which adopted the innovation with health care provider(s) where the innovation is not in use. - Standard hospital activity costing methods would be used to calculate the impact in terms of costs or savings to commissioners and hospitals and to model the impact of the innovation if adopted in larger areas (e.g. nationally). - Alongside this quantitative analysis, local qualitative information gathering and analysis would be carried out to understand how and why the innovations might have had their particular impact in those local health economies. The funders, NIHR, explicitly set up this programme to promote much more rapid analysis of service innovations, that might more quickly benefit health care services nationally. RAND Europe Community Interest Company (CIC), the University of Cambridge and the University of Birmingham, as joint data controllers, will be responsible for ensuring that all uses of the data will be within the agreed scope above. The process for selecting research projects to be undertaken by the BRACE team – including the innovations to be evaluated and the health and care services selected - is underpinned by close collaboration with the funder (NIHR) and health and care system stakeholders, including patients and the public. The team undertakes regular horizon scanning activities to identify service innovations which could be potential candidates for evaluation, using the following means: - Consultation with the NIHR, BRACE steering group and Health and Care Panel (a panel of 50 people with representation from all key stakeholder groups including clinicians, health and care service managers and leaders, policymakers, patients and the public, national organisations and arms-length bodies). - Consultation with wider health and care system stakeholders – including the national Academic Health Science Network, with whom BRACE has established links to support and inform horizon scanning activities - Desk-based research including reviews of reports, websites, social media, peer-reviewed research etc. - Inviting people to share details of innovations which would merit evaluation via an interactive function on the BRACE website. The resultant list of potential projects is then assessed using pre-determined criteria, which were developed in consultation with the BRACE steering group and agreed by the NIHR. These criteria include considerations of strategic importance, feasibility and scope. The short-listed topics are then discussed and ranked at a prioritisation workshop with Health and Care Panel members. The workshops are facilitated by an experienced workshop facilitator (who also advises the James Lind Alliance[a]) and are based on an adapted version of the approach employed by the James Lind Alliance Priority Setting Partnerships. The top ranked topics are put forward to the NIHR for consideration. BRACE's Patient Voice and NHS Management co-applicants [b] will provide oversight of the topic identification and prioritisation process. [a] The James Lind Alliance (JLA) is a non-profit making initiative established in 2004. It brings patients, carers and clinicians together in Priority Setting Partnerships (PSPs) to identify and prioritise the Top 10 unanswered questions or evidence uncertainties that they agree are the most important. The aim of this is to make sure that health research funders are aware of the issues that matter most to the people who need to use the research in their everyday lives. James Lind Alliance is funded by NIHR. The format of the workshops used in BRACE are adapted from the James Lind Alliance (JLA) model of consensus development. [b] The use of the word ‘co-applicant’ here refers to the original application to NIHR for funding for this work including patient and NHS management contributions to ensure that the research carried out covers issues most relevant to the people who need to use the research in their everyday lives. However, BRACE's Patient Voice and NHS Management co-applicants do not have responsibility for the final selection of topics nor carrying out the research at any stage. Once a project has been reviewed by the funder (to advise the BRACE team as to whether the project is within the scope of the research grant) the project enters a scoping phase. A topic specification form (TSF) is filled in at this stage, which includes a description of the rationale for undertaking the research, the research question, the proposed approach and timescale. The TSF is internally reviewed by the BRACE Director and Deputy Director, and is then submitted to the NIHR which makes a decision about whether to proceed to evaluation. If NIHR approval is given, a detailed research protocol is then developed by the project team (i.e. RAND Europe Community Interest Company (CIC), the University of Cambridge and the University of Birmingham). This provides a more detailed justification for and overview of the evaluation, the core research questions, and the design and methods proposed to answer those questions including plans for data processing and analysis. The protocol will also include references to the evaluation team’s approach to Patient and Public Involvement (PPI), ethical issues, data and information management, risk management and dissemination of findings. It is at this stage, which involves only RAND Europe Community Interest Company (CIC), the University of Cambridge and the University of Birmingham that the decision about the use of HES data in the evaluation will be made). All protocols are reviewed by the BRACE Director and by three independent peer reviewers, including at least one patient/member of the public. Once the protocol has been signed off by the funder, the team then proceeds to prepare applications for all necessary ethical and research governance approvals. For each project, BRACE seeks advice directly from the Health Research Authority to clarify the status of the project (as ‘research’ or ‘service evaluation’) and determine what level and type of approvals will be required. In developing the evaluation protocols for each rapid evaluation within BRACE the BRACE team will ensure that any use of these HES data conform to the scope above. In addition BRACE will confirm that: • Projects have clearly defined objectives and analysis plans; • In each case, the use of the HES data is necessary and proportionate to the purpose of the project – this will include consideration of the necessity for use of each individual HES data set; the number of years of data; the sizes of any cohorts or control cohorts derived from the data, and the inclusion and exclusion criteria (such as presence of specific diagnostic or procedure codes); • Appropriate safeguards are in place to protect confidentiality; minimise risks of re-identification and use of excessive data beyond necessity The individual or individuals who will carry out the project will be bound by an analysis plan detailing what data is permitted for use in the project and how it shall be processed. DATA CONTROLLERS AND PROCESSORS The University of Birmingham, RAND Europe Community Interest Company (CIC) and the University of Cambridge are joint data controllers for the data under this Data Sharing Agreement. The funder (NIHR), the Stakeholder Advisory Board (including BRACE's Patient Voice and NHS Management co-applicants) and the Health and Social Care panel offer advice on what service evaluation projects might be undertaken. Additionally, the funder may have a role in determining that a given project is not within scope of the research grant. However, decisions in respect of use of the data under this Agreement are taken jointly by University of Birmingham, RAND and the University of Cambridge without obligation to the funder or any other body. The University of Birmingham, RAND and the University of Cambridge will have joint responsibility for determining the purposes for processing the data under this Agreement to support the delivery of the projects and for determining the manner in which the data will be processed. With reference to the Information Commissioner’s published guidance on data controllership, RAND Europe Community Interest Company, the University of Birmingham and the University of Cambridge can confirm that these three organisations are jointly responsible for directing the research project. Specifically they decide: - to collect the data under this Agreement and the legal basis for doing so; - which items of data to collect, i.e. the content of the data; - the purposes the data are to be used for; - which individuals to collect data about; - whether to disclose the data - the data will not be disclosed to any third party; - whether subject access and other individuals’ rights apply; and - how long to retain the data There are two projects currently ongoing within BRACE and these projects do not need to use HES data, however these are provided as examples of innovations which will be investigated and which may potentially use HES data. These are as follows: An evaluation of Primary Care Networks An evaluation of Vertical Integration between primary and secondary care The current evaluation of Vertical Integration does not require the use of HES data, however, a second phase of the Vertical Integration evaluation is being scoped and it is expected that this will be the first BRACE evaluation to use HES data under this agreement. The nature of this NIHR programme is that further projects are not currently specified in this Agreement but will be identified at relatively short notice over the next several years. This Agreement permits RAND Europe Community Interest Company, the University of Birmingham and the University of Cambridge to use subsets of the data for the purposes of the projects specified above and for the purposes of the projects to be identified for the purpose of this NIHR programme which must be compliant with the description of such projects given above. RAND Europe Community Interest Company, the University of Cambridge and the University of Birmingham will send NHS Digital details of each BRACE projects progress as part of an annual update. Some of the innovations RAND Europe Community Interest Company, the University of Cambridge and the University of Birmingham will look at will inevitably need to be analysed with reference to long term historic data and so RAND Europe Community Interest Company, the University of Cambridge and the University of Birmingham will require inpatient, outpatient and A&E data for all of England from 20013/4 to 2022/23. RAND Europe Community Interest Company (CIC) is a not for profit policy research organisation based in Cambridge (registered with Companies House company number 0272801), with a mission to improve policy and decision making through evidence based research and analysis. RAND Europe Community Interest Company (CIC) has extensive experience working with sensitive data across health, education and criminal justice fields. RAND Europe Community Interest Company (CIC)’s offices in Cambridge and IT infrastructure provide an appropriately secure location for data storage and processing. For the purposes of explaining the corporate structure, RAND Europe Community Interest Company (CIC), together with RAND Europe (EU) AISBL, are the two legal entities which comprise RAND Europe, a research unit of the RAND Corporation. The work of BRACE is work is being carried out under a contract signed between the Secretary of State for Health and the University of Birmingham. RAND Europe Community Interest Company and the University of Cambridge have signed a collaboration agreement with the University of Birmingham. RAND Europe Community Interest Company and the University of Cambridge have signed a collaboration agreement with the University of Birmingham [SD8, SD8.1 and SD8.2] and any reference to RAND in this application refers solely to RAND Europe Community Interest Company (CIC).and any reference to RAND in this application refers solely to RAND Europe Community Interest Company (CIC). Data processing is restricted on site at the RAND Europe Community Interest Company (CIC) Offices. Staff from National Voices, the final collaborator in the BRACE evaluation centre provide a strategic voice on person centred care, the engagement of citizens in decisions about health, and the role of the voluntary and community sector to BRACE but will not be part of the team carrying out any analysis of HES data, or specifying how data will be used, and so are not named on the application. For RAND Europe Community Interest Company (CIC), the lawful basis for processing data is GDPR article 6(1)(f): Processing is necessary for the purposes of the legitimate interests pursued by the controller or by a third party except where such interests are overridden by the interests or fundamental rights and freedoms of the data subject which require protection of personal data, in particular where the data subject is a child. The Legitimate Interests Assessment (LIA) was carried out by RAND Europe Community Interest Company (CIC) using the ICO template. RAND Europe Community Interest Company (CIC) concluded that they can rely on legitimate interests for this processing. The legal basis under which the University of Birmingham and the University of Cambridge are applying to process the data is GDPR Article 6 (1) (e) [Public Task]. The legal basis under which RAND Europe Community Interest Company (CIC) are applying to process the data is GDPR Article 6 (1) (f) [Legitimate Interest]. 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. *Funding is for a five year period and this is the duration of the project. This Agreement is for 3 years and therefore a renewal will be required should data be needed for the final 2 years of the project.

Expected Benefits:

The impact of the evaluations carried out by the team will occur at multiple levels: local implementation of research findings, material that assists STP-wide or regional scaling-up of programmes that work, and national impact through providing evidence to inform policy and practice about which service innovations in the NHS are most promising or cost-effective. In particular, as the study team learned from experience, frontline staff and managers want not just evidence of effectiveness, but also details about how changes were made, including enablers and barriers to implementation and how the latter have been (or might be) overcome and this evidence will be provided. Outputs will also include the kind of facilitating information needed to support the decision-making of management teams around the NHS (e.g. information about costs, resources needed, prototype workflows, planning guides). The key benefit to health and/or social care of this research is the implementation into practice or policy of the findings of evaluations of service innovations. This will provide high quality evidence about their performance and potential, including when a service development should not be pursued further. These benefits are expected to occur across all BRACE evaluations, and for benefits to occur within the lifetime of BRACE funding (before 2023).

Outputs:

The core BRACE evaluation activities are: 1) finding and selecting promising services to be studied; 2) designing projects asking: how were these services put into practice; did they achieve their goals; and could they be introduced across the NHS; and 3) sharing the findings in a way that will both interest and inform people, providing the kind of information that front-line NHS and social care staff and managers need to help them change and improve services. Therefore, for each evaluation outputs include: At regular intervals – e.g. mid and end points in each evaluation depending on duration – formative feedback and learning sessions will be held. These will be comprehensive in ensuring a forum for shared learning and reflection on the evaluation evidence gained (e.g. between teams or services within sites, or across different sites) and also formative in terms of informing future action and facilitating scaleability, sustainability, the harnessing of identified opportunities and addressing barriers and challenges. The study team plan to produce an ‘emerging findings’ series, where they will publish slide sets of interim findings, with suitable disclaimers and caveats, prior to final reporting and with a process of rapid (e.g. 2-week) peer review. These slide sets will be made available through the BRACE website, and will be revised and updated as necessary once the project was completed, and/or subsequent insights from ongoing work were gained; The main dissemination channels for the finished work are likely to include: • Formal final reports published in the NIHR Journals Library. • Papers published in high quality, peer-reviewed, academic journals. • Concise summaries of findings, highlighting information relevant for decision makers, provided at the conclusion of each evaluation, targeted at appropriate audiences. Such a summary would likely include: i) syntheses across the portfolio of work, drawing together evidence on key outcomes and cross-cutting themes; and ii) short practical guides focusing on key implementation topics, summarising the learning from specific projects about the ‘nuts and bolts’ of how to plan, make and sustain service transformation; • Slide sets with creative commons licences that can be used in training and briefing (building on the emerging findings slide sets). These are especially useful, for example, in stimulating discussions at board and directorate level in healthcare organisations and for resourcing local training and activities; • The three core collaborating organisations in the Centre will actively promote the findings of each project through web distribution (including via the BRACE website), and a press release, a one-page policy brief/blog and a social media campaign if appropriate (Twitter, Facebook, ResearchGate), as well as widespread distribution of the finished reports as appropriate; • Raising awareness of the work of the team and publicising outputs through extensive networks and activities, including in the many national and regional leadership development and MSc/Diploma programmes that HSMC delivers for NHS and social care staff and via its online bulletins; BRACE proposes to capitalise on existing NHS channels such as providing content for The CCG Bulletin distributed by NHS England: https://www.england.nhs.uk/email-bulletins/ccg-bulletin/# As a matter of principle, BRACE will aim to work within rather than duplicate existing channels of communication. The members of the combined teams, especially the PI and co-applicants, collectively have numerous opportunities every year to speak at conferences and participate in roundtables, webinars and the like, where it would be appropriate to present the findings of the evaluations to NHS, social care and international audiences. • Patient and Public Involvement (PPI) members and the Health and Care Panel will be asked to assist in the production of summaries for non-technical audiences. The study team will also work through National Voices’ extensive networks within the patient charity and voluntary sector in this dissemination. Patients, managers and clinicians involved in individual projects will be encouraged and supported to disseminate through their own networks and activities. Across all of the above, BRACE seeks to combine more traditional (e.g. written) reporting formats with interactive media and visual content such as infographics. Final outputs from each evaluation are expected to be made publicly available within 12-24 months of the end of the project, or sooner.

Processing:

Data will flow from NHS Digital to RAND Europe Community Interest Company (CIC) by secure transfer. The data accessed through this agreement will be managed by RAND Europe Community Interest Company (CIC), all data processing will also be carried out on site at RAND Europe Community Interest Company (CIC). Data will only be shared with University of Birmingham or University of Cambridge colleagues or any other third-parties in aggregated form with small numbers suppressed in line with the HES Analysis Guide). University of Cambridge and University of Birmingham colleagues may access the data on site at RAND Europe Community Interest Company (CIC) offices in Cambridge. DATA ACCESS Data will only be accessed by named individuals from RAND Europe Community Interest Company (CIC), the University of Cambridge and the University of Birmingham. Individuals will sign a project-specific user agreement before they are allowed to access the data. This will be drafted to reflect the terms of the DSA and ensure that staff working with the data understand their obligations and the restrictions on permissible use (in line with RAND Europe Community Interest Company (CIC)’s practice on other HES applications), and in line with the analysis protocol. Access to the data will only be provided to staff on a need to know basis, and only permitted for the purposes of the BRACE evaluations. Only employees of either RAND Europe Community Interest Company (CIC), the University of Cambridge or the University of Birmingham will be allowed access. DATA PROCESSING Below is a representative example of an analysis framework for evaluating an innovation occurring in GP Practices using the data provided by NHS Digital. Raw data sheets provided by NHS Digital to RAND Europe Community Interest Company stored in the secure data server will be uploaded into a SQL database. The relevant data will be extracted from the data sets by calling the SQL database using Stata 15.0, which is installed in the secure area. The record level data will be processed into a cleaned analysis data set with the required outcomes, which it is expected to be A&E attendance, outpatients attendance and emergency admissions. Then the intervention and controls will be identified, and analysis will be performed on the basis of this identification. The analysis could be based on a before-and-after difference-in-differences analysis allowing for baseline and pre-intervention trends and controlling for practice size, other practice characteristics, deprivation, and population age/sex/ethnicity. It is anticipated that data covering a period of up to three years prior to and up to two years post introduction of the an innovation at each location could be used. The plan is to use Stata or SAS for the statistical analyses, both of which are available in the secure data hosting area, where the HES data are stored. In these models random effects with an unstructured co-variance matrix will be included for each year so that the underlying outcome level (and associated clustering/over dispersion) in each practice is incorporated and may change each year. An interaction term between year (following intervention) and intervention group will allow to assess the effect of the introduction of the innovation in each of the two years following intervention. Where possible, a second set of models will model intervention effects with data split up into month long periods of time. These models will be similar but further include a spline approach that models both sudden and gradual changes using interactions between the intervention group and (1) a post intervention indicator and (2) a linear year variable that begins at the intervention, respectively). When analysing secondary care usage a mixed effects Poisson regression will be used to model the count of admissions. For each practice data will be aggregated into 5-year age by gender groups and the rate of admissions will be modelled using freely available data on the age and gender profiles of the practice population. In doing so the applicant will be able to adjust for patient level age and gender and will further adjust for the same practice level variables as listed above, and estimate costs with standard health economics tools including the Diagnosis-Related Group (DRG) grouper. Given the nature of the Rapid Evaluation Service, data will be prepared by NHS Digital and held. Data will then be downloaded once the data controller has confirmed the requirement for the data for the specific project. DATA LINKAGE There will be no requirement nor attempt to re-identify individuals from the data. There may be a requirement to link publicly available data, using the GP-practice identifier (within HES and considered non-sensitive and publicly available), to identify intervention and control groups. No linkage will be done at the record-level; linkage will only be carried out using aggregate data with small numbers suppressed in line with the HES analysis guide. DATA MINIMISATION National data are required as BRACE is a national evaluation centre and will be carrying out up to 11 quantitative evaluations using the data sets provided by NHS Digital. It will not be possible to filter the data, as although the final evaluation protocols have not been agreed, as BRACE is a national evaluation centre and will be considering innovations across geographical regions and settings. Where specific organisations or patient groups are not part of the intervention cohort they will nonetheless contribute person-level data to the control groups used in the analysis. In addition, only the minimum number of fields required are being requested here and no identifiers or sensitive items are requested. 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).