GPs-in-EDs (Phase 2) NIHR HS&DR 15/145/04

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Objective for processing

The research described in this Agreement is an integral part of a wider study funded by the National Institute for Health Research (NIHR). The overall study is divided into separate work packages and Cardiff University, as the contact organisation with the NIHR for this study, is leading on Work Packages relating to the existing academic and practitioner literature, the intervention taxonomy, and the qualitative research. BACKGROUND AND PURPOSE Swansea University is the lead organisation for study Work Packages involving quantitative research, including the work described in this Agreement. Cardiff University has been included as a joint data controller because of its involvement in the earlier phases in this study. Cardiff University jointly scoped the need for this project and determined why this project is required. Cardiff University will not receive or process any data under this Agreement but may provide advice as part of the study management team. Only Swansea University will process the data under this Agreement. The research study to which this Data Sharing Agreement pertains forms an integral part of a wider study funded by the National Institute for Health and Care Research (NIHR). The NIHR-funded study is divided into separate work packages: Cardiff University, the contact organisation with the NIHR for this study, are leading on work packages relating to the existing academic and practitioner literature, the intervention taxonomy (intervention taxonomy seeks to help researchers conceptualize required elements of intervention protocols and to enhance both the internal and external validity of intervention research) and the qualitative research. Swansea University are the lead organisation for study work packages involving quantitative research. It is the work package led by Swansea University that this Data Sharing Agreement relates to. The NIHR is funding this project. There is a steering committee which provides advice to each delegated work package responsible teams. As such the NIHR is not a data controller as it only provides advice as requested by the joint data controllers Cardiff University and Swansea University. The wider study was set up in response to the urgent need for evidence to understand how best to manage workload and demand in the emergency care system to safely achieve the highest standards of clinical and operational care, whilst ensuring that the experience of patients is acceptable and enhanced where possible. A joint report by the Royal Colleges and NHS Confederation suggests that every Emergency Department (ED) should have a co-located primary care facility. However, the evidence base to support different service models of General Practitioners (GPs) working within EDs (GP-ED models) is weak, and may paradoxically increase overall attendances due to provider-induced demand, with only marginal savings per patient. Previous work from the Primary Care Foundation has identified three main GP-ED models: GPs working geographically adjacent to EDs; GPs serving in a triage and screening capacity; and GPs fully integrated into ED service provision. The effectiveness of each model is unclear in terms of its impact on service provision, patient experience and the effectiveness and safety of clinical care delivery. Lawful basis under GDPR Cardiff University and Swansea University are undertaking a realist evaluation to understand how the contextual differences between the three main GP-ED models (including patient, staff and organisational factors) influence how the services function in different settings (mechanisms) to generate variations in intended and unintended outcomes (effectiveness and safety of care provided; patient and staff experience and wider system implications - i.e. resource consequences, training needs, economic impact). GDPR Article 6 (1) (e) states that processing is lawful if it is necessary for the performance of a task carried out in the public interest. This analysis is based around the ability of the different GP models to achieve the key outcome domains described in an effective practice framework: i.e. addressing greatest health needs first; only doing what is needed; reducing inappropriate variation; and co-production. Swansea University has determined that the use of data under this Agreement to undertake an assessment of the various methods of deploying General Practitioners (GPs) working within Emergency Departments (EDs) is in the public interest from both clinical and economic perspectives. Furthermore, the processing proposed is necessary to provide the data required to make this assessment. DATA SUMMARY GDPR Article 9 (2) (j) states that processing of personal data is lawful if it is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes and proportionate to the aim pursued, respect the essence of the right to data protection. Swansea University wish to retain the record-level Hospital Episode Statistics (HES) Accident & Emergency (A&E) and HES Admitted Patient Care (APC) data previously disseminated under this Agreement to strengthen the evidence base on different service models of GPs-in-EDs. The data is retained only for specified study sites (twelve EDs in England). The data subjects are all individuals who appear in HES A&E and/or HES APC between 2010/2011-2018/2019 in these 12 EDs. Swansea University has determined that the processing outlined in this Agreement will enable assessment of various GPs-in-EDs models currently in use at Emergency Departments. This assessment is both in the public interest (as explained above) and, by defining clear objectives for the statistical analysis of the data requested, the processing is proportionate to the study’s research aims. This window (2010/2011 – 2018/2019) has been chosen to provide a time frame before the first intervention of GPs-in-EDs and after the last intervention. Analysis of this period hopes to facilitate a better understanding of the effect of using the GPs-in-EDs model. Swansea University is requesting HES Accident & Emergency and HES Admitted Patient Care data at record level in order to strengthen the evidence base on different service models of GPs-in-EDs. This study seeks to understand the impact of such services on patient health outcomes; their experience and safety, and the health services resource use and related cost of delivering these models. The pseudonymised record-level HES A&E and HES APC data that Swansea University holds provides information on admission date and time, gender, age, the date and time of a health event, visit status (first attendance or re-attendance), attendance category, primary diagnosis, investigations, treatment and disposition codes, and mainly non-sensitive information relating to subsequent admission, episodes and discharge. Swansea University do hold one data item which is categorised as sensitive - this field (referrer code – person who referred the patient) may facilitate, combined with wider analysis of the aforementioned HES A&E and HES APC fields, a more complete understanding of patient assessment and treatment following presentation at study site EDs. The overall aim of the study is to determine the clinical and cost effectiveness of GPs-in-EDs models and to understand the ways in which service design and setting (context) generate variations in outcomes. Swansea University process this record level data to calculate various ED-specific estimates which can be used as outcome measures to compare different GPs-in-EDs models. These measures, in the form of appropriate summary statistics, include the average patient age and gender split; the total number of ED attendances; admissions; the proportion of re-attendances; average waiting times, and the proportion of patients admitted as an inpatient to the same hospital. The study is intended to evaluate the ability of different GPs-in-EDs models to achieve the key outcomes described in an effective practice framework (addressing greatest health needs first; only doing what is needed; reducing inappropriate variation) and the resources needed to deliver these outcomes. The 12 ED sites are listed below, along with the Site Provider codes which, to the best of the study team’s knowledge, have been in use over the study period: Swansea University is requesting pseudonymised patient level data only for specified study sites which include twelve Emergency Departments in England. The data subjects are all individuals who have created HES episodes at A&E, possibly followed by admission, during the period of time requested and minimised to the providers listed below. Swansea University is requesting pseudonymised patient level data providing information on admission date and time, gender, age, health event date and time, visit status (first attendance or re-attendance), attendance category, primary diagnosis, investigations, treatment and disposition codes, and mainly non-sensitive information relating to subsequent admission, episodes and discharge. Swansea University is requesting one data item which is categorised as sensitive; this field will facilitate, through analysis of HES A&E and HES APC fields, a more complete understanding of patient assessment and treatment following presentation at study site EDs. All other data requested is pseudonymised, personal, record level, special category data. Swansea University will use this record level data to calculate various ED-specific estimates to be used as outcome measures to compare different GPs-in-EDs models. These measures, in the form of appropriate summary statistics, include the average patient age and gender split; the total number of ED attendances; admissions; proportion of re-attendances; average waiting times, and the proportion admitted as an inpatient to the same hospital. The study team has only requested HES A&E and HES APC data fields required to execute the study's Statistical and Health Economics Analysis Plan (SHEAP) as approved by the independent Study Scrutiny Committee. The study team have also limited the request to the work package window between 2010/11 and 2018/19. This window has been chosen to provide a time frame before the first intervention of GPs-in-EDs and after the last intervention. The time before and after is to be able to analyse the effect of using the GPs-in-EDs model, from prior and post intervention data. At the data extraction stage, the study team will only request data relating to presentations to and attendances at Emergency Departments (EDs) at twelve English study sites. The sites are listed below, along with the Site Provider codes which, to the best of the study team’s knowledge, were in use over the study period. The twelve sites are as listed below; this is the final list of study Emergency Departments in England. The twelve sites and codes are: [12 paragraphs unchanged] Swansea University has have determined that there is no alternative, less intrusive method of obtaining the various site-level characteristics required to achieve the study’s objectives. DATA CONTROLLERSHIP Cardiff University and Swansea University are joint data controllers for the purposes of this Data Sharing Agreement. Although the quantitative work package for which the NHS Digital data required is being used is led by Swansea University, Cardiff University have been included as a joint data controller because they have responsibility in determining the purposes and means of the processing of NHS Digital data. Cardiff University have not received or processed any of the NHS Digital disseminated under this Agreement; they may only provide advice as part of the study management team. Swansea University alone process the NHS Digital data. There are a number of co-investigators involved in the study (based at the University of Lincoln, University of Warwick, Imperial College London, Oxford School of Emergency Medicine and Macquarie University, Australia; also two public contributors). None of these individuals have or have had decision-making responsibility for the data and have not or will not have access to the data in any form other than aggregated data with small numbers suppressed in line with the HES analysis guide. The NIHR is funding this project. There is a steering committee which provides advice to the teams overseeing each work package but they have no responsibility in determining the purposes and means of the processing of NHS Digital data. As such the NIHR is not considered to be a data controller. LEGAL BASIS FOR PROCESSING Swansea University have a lawful basis for processing personal data and special categories of personal data under the following provisions of the General Data Protection Regulation: - Article 6 (1) (e) states that ‘processing is lawful if it is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller’. - Article 9 (2) (j) states that the processing of personal data is lawful if it is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes and proportionate to the aim pursued, respect the essence of the right to data protection. Swansea University, per Schedule 1 of the Freedom of Information Act 2000, are considered to be a public authority. Swansea University possess a Royal Charter that confers power on the university to “to advance learning and knowledge by teaching and research”. The data processing is considered to be in the public interest as this research study, through an assessment of various GPs-in-EDs models currently in use at EDs, may shape the provision of GP services in EDs at all Emergency Departments in the UK. PATIENT AND PUBLIC INVOLVEMENT (PPI) Cardiff University will invite up to 40 people, to include clinical leads, commissioners, policy makers, GPs and ED clinicians, nurse practitioners, patients & public contributors including parents of children that may use the service, in roughly equal proportions. We will recruit purposively via contacts (individual and organisational) of co-applicant group and from Survey (Online/postal survey of all 200 Type 1 EDs in England & Wales for medical directors and commissioners) responses. Cardiff University will ensure the workshop format enables all attendees to contribute their different perspectives and experiences and for views to be heard and discussed. The public contributor co-applicants will be actively involved in designing the workshop and will hold a pre-meeting for patient and public delegates to assist their participation in order to make effective contributions. PPI has been extensive over the course of this research study. Two public contributors with experience of using primary and secondary health services were recruited from the Public Involvement Community supported by Health and Care Research Wales. Their responsibilities included but were not limited to: - Forming part of the research management group taking part in all discussions about strategic and implementation decisions. The research management group convened every 1-2 months. - Forming part of the study steering committee alongside methodological, clinical and policy specialists providing oversight and advice to the research team. The steering committee met every 6 months. - Involvement in annual Public Involvement Team Meetings. The Public Involvement lead coordinated these meetings and attended with the two public contributor co‐applicants. At the meetings, the team planned and reviewed the role and operation of public involvement in the study, identified and accessed support and training, including any issues causing distress or difficulty for individuals, and identified opportunities to extend public involvement in the study. Briefing sessions were identified and provided, but no formal training was requested by public contributors. Members co‐authored a Role Description document, which they updated annually and reported to the research management group. The study team have received a significant amount of positive feedback in relation to their PPI engagement. Research co-applicants considered the public involvement to have exceeded their expectations and reported positive feelings including pride and self-worth arising from the experience. The funder also described the public involvement as “a fine example of embedding public and patient involvement (PPI) throughout the course of the research” after receiving routine study reports. A paper was published Health Expectations: An International Journal of Public Participation in Health Care and Health Policy on the study’s PPI engagement - found here - https://pubmed.ncbi.nlm.nih.gov/35894169/.

Processing activities

Record level, pseudonymised HES A&E and HES APC data - minimised as described above - will be extracted from the HES A&E and HES APC datasets and securely disseminated to Swansea University. Record-level, pseudonymised HES A&E and HES APC data have already been extracted and securely disseminated to Swansea University via NHS Digital’s Secure Electronic File Transfer (SEFT) service. No further data is being disseminated under this Agreement. The data from NHS Digital will be is combined with the corresponding equivalent SAIL (record level Secure Anonymised Information Linkage (SAIL) data (record-level, pseudonymised data for the sole Welsh A&E site - a study control). There will be no linkage of this combined data and this combination will control) to allow for analysis using a study ‘control’ site. Further There will be no further linkage of this combined data, however further manipulation of the combined data will be is required to generate summary characteristics for the extended control group. These summary characteristics will then be reported. Swansea University will, University, in accordance with the study's Statistical and Health Economics Analysis Plan (SHEAP), then calculate various Emergency Department specific ED-specific estimates to which will be used as outcome measures to compare different GPs-in-EDs models. These measures, [10 words unchanged] patient age and gender split, the total number of ED attendances, admissions, the proportion of re-attendances, average waiting times, and the proportion admitted as an inpatient to the same hospital. For continuous outcomes, appropriate summary statistics will be analysed. There is no requirement to re-identify individuals, and no re-identification has been or will be attempted. There is no requirement to re-identify individuals, and no re-identification will be attempted. Data processing will only be performed by substantive employees of Swansea University, the sole data controller and sole data processor. These employees will possess appropriate, certified and SAIL-specified training in data protection and confidentiality, such as that provided by the Medical Research Council. All NHS Digital data is stored, processed and accessed at Swansea University, with secure access to study researchers provided solely via the SAIL Gateway. Data processing is only performed by substantive employees of Swansea University, the joint data controller who solely processes the NHS Digital data disseminated. Study researchers are accredited SAIL Gateway users who have completed appropriate and certificated training on Research Data and Confidentiality. Researchers will only be able to access the data through SAIL’s secure two-factor authentication processes. These processes involve the use of a personalised ‘Yubikey’ device (a hardware authentication device used to protect access to access to computers, networks and online services), issued only to accredited SAIL Gateway users. All NHS Digital data will be stored, processed and accessed at Swansea University, with secure access to study researchers solely via the SAIL Gateway. Study researchers will be accredited Sail Gateway users, will have completed appropriate and certificated training on Research Data and Confidentiality, and will only be able to access the data through SAIL’s secure two-factor authentication processes. These processes involve the use of a personalised ‘Yubikey’ device, issued only to accredited SAIL Gateway users.

Expected output

To meet a condition of the research funding, Swansea University will produce have produced a comprehensive study report containing appropriate summaries of the study data (including HES data), data) and analysis of that data. This has been supplied to the study funder (NIHR). The deadline for this report is February 2021. The report will anticipated to also form the basis of academic articles to be submitted to high-quality peer-reviewed journals, conference presentations, and other outputs, as specified in the study's dissemination strategy. In addition to fulfilling the requirements to submit a definitive study report to its funder (NIHR, the National Institute for Health Research), the study team will seek simultaneously to publish the main study findings in appropriate high-impact academic journals. Given that these findings concern the provision of GP services, and hence will be of widespread interest, it is envisaged that target journals will include The Lancet, the British Medical Journal, the British Journal of General Practice, and the Annals of Emergency Medicine - all with an international readership. It is also the intention of the study team to present study findings at appropriate National and International conferences, such as those annually hosted by the Health Technology Assessment international (HTAi). All data summaries will comply with SAIL’s disclosure and dissemination policies (including small number suppression) and procedures (including vetting for inadvertent disclosure) and only summary data outputs approved for release from SAIL will be further disseminated. The study team will summarise findings for dissemination to NHS organisations. The Universities anticipate the study will provide the evidence base for unscheduled care delivery in Emergency Departments in England and Wales, evaluating different models of GPs in or alongside EDs for their effectiveness and safety, good use of resources and sustainability, suitable for local contexts. The Universities will explore strategies for dissemination with the Stakeholder group and other senior advisors, consisting of senior policy and operational representatives of both nations and from Royal Colleges (Emergency Medicine, General Practitioners). It is likely that this will involve presentations at leading national conferences, as well as securing invitations to smaller seminars and local meetings to a variety of professional and lay audiences. In addition to fulfilling the requirements to submit a definitive study report to its funder, the study team hope to simultaneously publish the main study findings in appropriate high-impact academic journals. Given that these findings concern the provision of GP services, and hence are expected to be of widespread interest, it is envisaged that target journals will include The Lancet, the British Medical Journal, the British Journal of General Practice, and the Annals of Emergency Medicine - all with an international readership. It is also the intention of the study team to present study findings at appropriate National and International conferences, such as those annually hosted by the Health Technology Assessment international (HTAi). Timing of these communications will be in agreement with the funder, the NIHR. The study team is obliged by the terms of its funding to liaise with the NIHR in order to maximise impact by synchronising the publication of major outputs. The current target date for submission of the funder’s report is February 2021 but the report’s publication timetable is dependent on the NIHR’s reviewing processes. Similarly, conference presentations on study results are also embargoed by the funder until publication of its report. The study team will further liaise with the funder on this matter, again with the view of maximising impact. The study team also hope to summarise findings for dissemination to NHS organisations. Cardiff University and Swansea University anticipate the study will provide the evidence base for unscheduled care delivery in Emergency Departments in England and Wales, the evaluation of different models of GPs in or alongside EDs for their effectiveness and safety, good use of resources and sustainability, and suitability for local contexts. The Universities will explore strategies for dissemination with the stakeholder group and other senior advisors, consisting of senior policy and operational representatives of both nations and from Royal Colleges (Emergency Medicine, General Practitioners). It is likely that this will involve presentations at leading national conferences, as well as to smaller seminars and local meetings with a variety of professional and lay audiences. The timing of these communications will be in agreement with the funder, the NIHR. The study team is obliged by the terms of its funding to liaise with the NIHR in order to maximise impact by synchronising the publication of major outputs. Although the funder’s report has now been submitted to the NIHR, the report’s publication timetable is dependent on the NIHR’s reviewing processes. Similarly, conference presentations on study results are also embargoed by the funder until the publication of its report. The study team will further liaise with the funder on this matter, again with the view of maximising impact.

Expected measurable benefits

Increasing demand on the United Kingdom (UK) National Health Service, urgent and NHS, particularly emergency services services, is testing the capacity of the healthcare system and there are concerns about the consequent ability to deliver good quality care. One of the recommendations produced in a joint report produced by the Royal Colleges of Emergency Medicine, Paediatrics and Child Health, Physicians and Surgeons to address these pressures concerns is that every Emergency Department (ED) should have to implement a co-located primary care facility. facility in every ED in the UK. However, the evidence base to support different service models of General Practitioners [23 words unchanged] the health services resource use and related cost of delivering these outcomes is important. This research is intended to evaluate could be important in evaluating the ability of GPs in the ED setting and different GP models to achieve key outcomes in an effective practice framework: namely, addressing the greatest health needs first; only doing what is needed; needed and reducing inappropriate variation; and variation, co-production and the resources needed to deliver these outcomes. It is anticipated that the research findings (due February 2021) will be of immediate interest to policy makers, service commissioners and providers. The study’s findings have the potential to shape the provision of GP services at all EDs in the UK. The potential impact is, therefore, considerable, effectively extending to the entire general population. It is difficult, in advance of undertaking the cost-effectiveness analyses proposed in this study, to quantify with any precision potential costs and savings but, should the study’s findings lead to savings in the NHS-wide delivery of such services, then these are likely to be considerable. The study’s findings have the potential to shape the provision of GP services in EDs at all Emergency Departments in the UK. The potential impact is, therefore, considerable, effectively extending to the entire general population. It is difficult, in advance of undertaking the cost-effectiveness analyses proposed in this study, to quantify with any precision potential costs and savings but, should the study’s findings lead to savings in the NHS-wide delivery of such services, then these are likely to be considerable. It is anticipated that the research findings (now due end of August 2022) will be of immediate interest to policymakers, service commissioners and providers. Any benefits arising from the study will be to the NHS. The NHS includes in England, including NHS Digital, the NIHR (the study funder), NHS Digital, as well as and the study sites. The study team will only benefit from cost or efficiency savings as members of the public, public - i.e. in the same way as other members of the public using the NHS. [1 paragraph unchanged]

Benefits reported

Yielded Benefits is not a requirement for new applications. Initial work has informed qualitative findings from the study, however, the full synthesis will require a complete analysis of the study data. Over and above COVID-19 delays and the implementation of effective work from home policies within Swansea University, this study has suffered from an unusually high turnover of key analysts, including an original study applicant, the departure of the Trials Unit’s statistician and the extensive reorganisation and reassignment of Health Economics researchers and as such, no benefits have been yielded.