Frequent Users of the Emergency Department: Improving and Standardising Services-a mixed methods study

BACKGROUND: Approximately 2.5% Emergency Department (ED) users account for 10% of total attendances. Many frequent users suffer from mental health and social problems. Services for frequent users have developed recently in a piecemeal fashion and evidence of efficacy of interventions is weak. In previous work we found 80/170 Emergency Departments with frequent user services with 4 different service types. These are: frequent user services based in ED and run by ED staff: frequent user services based in ED and run by mental health staff (liaison mental health teams); frequent user services with in-reach to ED from mental health services; and community based services for frequent users based outside the ED. AIMS: Our aims are to: a) describe current patterns and costs of frequent use of urgent and emergency care (UEC); b) describe the services for frequent users; c) identify predictors of high-cost patterns of attendance, and the impact of frequent user services on attendance; d) identify which interventions appear to work for which types of frequent users and why, and test these findings by in-depth case-studies of 4 different service types; and e) disseminate an implementation framework for frequent user services to improve planning and optimise care. METHODS: Workstream 1: This is a cross-sectional mapping of the current extent of services for frequent users of urgent and emergency care networks in England, and a mixed methods study to characterise 20 representative frequent user services (5 each from 4 different types). From this we will develop early ideas about how interventions may work for different subgroups of frequent users (i.e. safely reduce urgent and emergency care use if appropriate plus/minus provision of additional help and support). Workstream 2: This is a large data study (years 2016/17 to 2020/21) using two complementary datasets. The CUREd dataset links the urgent and emergency care network (ED, 111 and 999) data for Yorkshire and Humber Region. Hospital Episode Statistics data will be linked to ED attendance for the whole of England. The data from CUREd are more finely grained than HES data, whereas HES data provide a national perspective as opposed to a regional perspective capture by CUREd. Together they complement each other. We will: a) identify patterns of frequent use and sub-groups of frequent users; b) examine how frequent users access the whole urgent and emergency care network including attendance at multiple sites; c) study healthcare costs of frequent users to understand where costs are generated and the potential for reduction d) conduct an interrupted time series analysis of the impact on ED frequent users of the introduction of frequent user services. We will also examine the impact of the COVID-19 pandemic on frequent use of the ED. Workstream 3: This involves a realist synthesis to identify and test programme theories about how interventions for subgroups of frequent users produce outcomes. The synthesis will include a realist literature review, and will incorporate the early ideas from workstream 1, 4 in-depth exemplar case studies (one each from the different types of frequent user service) and data emerging from Workstream 2. The synthesis will run iteratively throughout the project to ensure that work across the project is informed by and understood within a testable explanatory framework of cause and effect. Conceptually, it will draw on relevant theoretical models and take a whole systems approach across the urgent and emergency care network at micro, meso and macro levels. Workstream 4: We will develop an implementation framework of 'ideal' models of service delivery tailored for the 4 different types of frequent user service, with a focus on specific interventions for particular subgroups of frequent use. TIMELINE: 0-18months: We will employ staff, set up our programme oversight committee, set-up our PPI reference group, obtain relevant ethics and research governance permissions, begin the analyses using the CUREd dataset, obtain the relevant HES data from NHS digital, start the realist review, conduct the mapping project followed by 20 structured interviews involving frequent user services (5 each from each of the 4 different types). 18-30months: We will carry out 4 in-depth case-site studies, carry out and complete the analyses using the HES data, conduct the main health economic analyses, complete the analyses using the CUREd data, complete the realist synthesis. 31-36months: We will develop the implementation framework with case-site involvement and PPI, disseminate our findings to clinical services via a cascading series of webinars and write up the main results of the programme. DISSEMINATION We will use several different routes to disseminate our findings to relevant stakeholders including top down via NHS England and the medical and nursing Royal Colleges and bottom up via cascading webinars for clinical services plus patient groups. ANTICIPATED IMPACT: We intend to improve outcomes for frequent users of urgent and emergency care networks by optimising services for frequent users, so they provide a safe and appropriate therapeutic response whilst helping to reduce urgent and emergency care use. We will: provide an up-to-date mapping of current services across England and detailed characterisation of four different service types; a characterisation of the different patterns of frequent use of urgent and emergency care and their associated factors; detailed costs related to the pattern and use of urgent and emergency care services by frequent users; an in-depth understanding of why people use urgent and emergency services frequently; a robust evaluation of the impact of frequent user services including likely cost-effectiveness; identification of specific interventions for certain sub-groups of frequent users; understanding of the impact of COVID-19 on frequent use of urgent and emergency care networks; we will develop a framework for frequent user services which will include 'Äòideal models'Äô for each of the different service types; we will distribute the framework to all current frequent user services via a series of cascading webinars; and conduct additional multilevel knowledge transfer at a multinational, national, service and patient/ service user level. Our findings will enable urgent and emergency care networks to make informed planning choices about services for frequent users. We will provide best evidence on how interventions work (or fail) in order to guide on-the-ground delivery. By incorporating PPI in the research we will increase awareness of constructive approaches to this important issue and aim to improve care for people who use emergency care on a frequent basis.