Article Text
PDF
Abstract
Background Frequent attendances of the same users in emergency departments (ED) can intensify workload pressures and are common among children, yet little is known about the characteristics of paediatric frequent users in EDs.
Aim To describe the volume of frequent paediatric attendance in England and the demographics of frequent paediatric ED users in English hospitals.
Method We analysed the Hospital Episode Statistics dataset for April 2014–March 2017. The study included 2 308 816 children under 16 years old who attended an ED at least once. Children who attended four times or more in 2015/2016 were classified as frequent users. The preceding and subsequent years were used to capture attendances bordering with the current year. We used a mixed effects logistic regression with a random intercept to predict the odds of being a frequent user in children from different sociodemographic groups.
Results One in 11 children (9.1%) who attended an ED attended four times or more in a year. Infants had a greater likelihood of being a frequent attender (OR 3.24, 95% CI 3.19 to 3.30 vs 5 to 9 years old). Children from more deprived areas had a greater likelihood of being a frequent attender (OR 1.57, 95% CI 1.54 to 1.59 vs least deprived). Boys had a slightly greater likelihood than girls (OR 1.05, 95% CI 1.04 to 1.06). Children of Asian and mixed ethnic groups were more likely to be frequent users than those from white ethnic groups, while children from black and 'other' had a lower likelihood (OR 1.03, 95% CI 1.01 to 1.05; OR 1.04, 95% CI 1.01 to 1.06; OR 0.88, 95% CI 0.86 to 0.90; OR 0.90, 95% CI 0.87 to 0.92, respectively).
Conclusion One in 11 children was a frequent attender. Interventions for reducing paediatric frequent attendance need to target infants and families living in deprived areas.
- urgent care
Statistics from Altmetric.com
Key messages
What is already known on this subject?
Frequent paediatric attendances in emergency departments (EDs) are often avoidable and stressful for both child and family.
Serious infectious disease incidence has fallen with the success of vaccination programmes, however, amidst the ongoing global COVID-19 pandemic, pressures on hospital EDs are stretched to their limit.
The reduction of frequent attendances in EDs is a public health concern that has had its focus mainly on adult and elderly users. There is a paucity of national level data exploring frequent paediatric attendances.
What this study adds?
Using a national database (Hospital Episode Statistics) for England, we found that that 1 in 11 children (9.1%) who attended an ED attended four times or more in a year. Infants and children residing in deprived areas had a greater likelihood of frequent attendance compared with children residing in affluent areas.
Integrated, whole-system care models that support parents, particularly among infants residing in deprived areas, should be sought to provide appropriate response to users with recurring health needs.
Introduction
Increasing pressures on emergency departments (ED) present a considerable challenge worldwide, particularly during winter.1 Prior to the COVID-19 pandemic, serious infectious disease incidence has fallen with the success of vaccination programmes. However, amidst the ongoing global pandemic pressure on hospital ED are stretched to their limit. This can strain health resources and budgets and can result in poor clinical outcomes.2 Increasing demand for EDs may be driven by rising morbidity in an ageing population, poor access to primary care and increase in patient expectations. In England, in 2017/2018, there were 23.8 million ED attendances, an increase of 22% since 2008/20093; rises were higher in the under-5 (28%–30%),4 and one-third of all British children visit ED each year.5 Such increases pose immense challenges to the National Health Service (NHS) amidst significant cuts in funding,6 given that that nearly half of the health budget is spent on emergency and acute care.7
Frequent attendances (FA) of the same users play an important role in the rise of ED attendances, and the reasons for this will vary, depending, for example, on the underlying epidemiology of disease or the structure of the particular health system in a country regarding parental access to primary care. Serious infectious disease incidence has fallen with the success of vaccination programmes,8–10 but many EDs find they are faced with increased demands from those with chronic non-communicable diseases such as asthma, diabetes, epilepsy and mental health disorders.11 12 In adults, frequent ED attendances often include mental health diagnoses.13
Rising short stay admissions for both infectious illness and chronic conditions raises concern that the need for acute, first-contact care may not be met by primary care.14 15 Paediatric ED attendance rates and short stay hospital admissions have been driven by changes in out-of-hours provision by general practitioners (GPs) in England.16 Moreover, children and families will often attend services offering extended hours despite the majority being registered to a GP.17 US parents interviewed as to the reasons that they brought their children to the ED believed that paediatric ED visits could be avoided if responsive primary care alternatives were available. Many of them had originally contacted a GP and were referred to the ED, or were concerned their child was too ill to wait for an appointment.18 However, hospital-based emergency care is not designed to provide continuity of care for recurrent problems and exposes the child to hospital acquired infections.19
Several studies describe the volume and characteristics of frequent paediatric users at hospital EDs.20–22 Most studies use only either a single or several centres, which are prone to unrepresentative samples, and only a few provide statewide estimates.21 23 24 In a recent study, we found that 9.5% of all ED attenders in English hospitals attended three times or more in 1 year and accounted for 27.1% of the EDs attendances. The groups with most FAs in English EDs were infants and elderly people. Frequent ED attenders had a higher risk of hospital admission than less-frequent ED attenders.25
However, other than one single-centre study conducted by our group,26 we are unaware of any attempt to describe frequent paediatric ED attendances in England. We found that 6% of under-5 were frequent attenders, who accounted for 20% of the workload in the EDs.26 Frequent ED attenders were predominantly those who habitually return with self-limiting conditions and those with or without exacerbation of underlying long-standing illness. However, at least in adults, a Canadian study indicated low complexity is not a contributory factor in ED overcrowding, and so reducing the number of low-complexity ED patients is unlikely to lessen ED crowding.27
The UK healthcare system has different characteristics from many other countries in terms of access to emergency care and the interface between primary and emergency care, such as strong gatekeeping.28 EDs in England are hospital based and provide consultant-led 24 hours service with full resuscitation facilities for accident and emergency patients. Many major hospitals in the UK operate a triage system to EDs in the form of front-end, GP-led urgent care centres colocated next to EDs, Where only acute cases are triaged to the ED. Nurse-led minor injury units in small hospitals also provide urgent care.
Due to the lack of large-scale studies describing the volume and characteristics of frequent paediatric users at hospital EDs, and the uniqueness of the English emergency care system, we aim to describe the volume and the characteristics of frequent paediatric ED users in English hospitals.
Methods
Study design
We conducted an observational study using routine administrative data. Hospital Episode Statistics (HES) data covering all attendances at NHS hospitals in England for the period April 2014–March 2017 were used.
Setting
We included all hospital-based EDs in England that provide consultant-led 24 hours service with full resuscitation facilities for accident and emergency patients (‘Type 1’ EDs).
Subjects
We included all children (age <16) who attended at least once at one of the English EDs reporting data to HES. Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.
Data preparation
We included all unplanned attendances in major EDs in all English hospitals. Data for each of the 3 years were sorted by patient ID and date of attendances. Duplicate attendances of the same user on the same date were removed. In initial data exploration, we found that most duplicate attendances of the same patient on the same date were within less than 25 min from the first attendance on that date, which were likely to be technical duplicates rather than standalone attendances. Attendances of users with a recorded in-hospital death during each year were removed to ensure that all users have a similar follow-up period.
Outcome measure
The definition of FA is a subject of clinical and methodological discussion among researchers. Previous studies on frequent users in EDs defined a frequent user based on the number of attendances within a given time frame (normally 1 year). For example, Hunt et al 29 chose four visits or more because that group of patients accounted for 25% of ED visits.
However, counts of repeat attendance across the year provide less information about repeated attendances clustered around one episode. Counting the number of attendances within 1-year might miss ‘tails’ of attendances that occurred just before or just after the year under investigation. Definitions range from 3 to 12 attendances per year.30 Our method of calculating whether a person was an FA was different. The outcome measure was the number of attendances each user made in the financial year 2015/2016 (figure 1). The method is based on finding two index attendances for each user and counting the number of attendances that occurred within 365 days before and after each index attendance. To ensure maximum coverage of the year, we used both the first and last attendances in the FY 2015/2016 as ‘first index’ and ‘last index’. This year was chosen to ensure data is available for a preceding year (2014/2015) and succeeding year (2016/2017), to be able to count attendances occurring in the 12 months forward and backward from the index attendances. Users with four attendances or more (excluding the index attendance itself) in the preceding 365 days or the succeeding 365 days to each of the two index attendances were classified as frequent users. This threshold was based on the assumption that clinicians would not expect more than four emergency attendances in a year. We have previously reported repeat attendances in 1 year in children, finding that less than 2% of preschoolers visited more than six times.31 Similarly, we will report that the in a general population, only 1.2% attended six times or more (in press). We were concerned that, due to the Poisson-shape distribution of the repeat attendances, raising the threshold would narrow the group of FA significantly.
Calculation of frequent attendances (FA). The ‘X’s represent attendances and the ticks represent attendances that would qualify to be counted FAs. This patient had two attendances in the 12 months following the first index attendance and two attendances in the preceding 12 months to the first index attendance. They had three attendances in the 12 months following the last index attendance and three attendances in the 12 months preceding the last index attendance. They would qualify as an FA because they had three attendances in the vicinity of the last index attendance. Considering merely the first index attendance, they would not be classified as an FA. They had only two attendances proceeding to the first index attendance in the onward 12 months, and only two attendances preceding to the first index attendance in the backward 12 months.
Explanatory variables
Age of admission was categorised into four groups: under 1 year old, 1–4 years, 5–9 years and 10–15 years. Socioeconomic status was measured using the Index of Multiple Deprivation (IMD) ranging from 1 (most deprived) to 5 (least deprived). The IMD is the official measure of relative deprivation for small areas in England. It ranks every small area in England from 1 (most deprived area) to 32 844 (least deprived area). The small areas are designed to be of a similar population size with an average of 1500 residents. It is common to group the IMD into population-weighted fifths, with an equal total population in each fifth. Ethnic group was categorised into five groups: white, black, Asian, mixed and other.
Statistical methods
We used a mixed effects logistic regression model to predict the odds of being a frequent user by the explanatory variables as fixed variables. The unit of analysis was a single user. The binary outcome was the being an FA (0/1). As users are clustered into hospitals and GP practices, a nested model was considered. Clustering by GP Practice proved impractical due to a substantial number of practices with small number of users: 4625 of the 12 594 GP practices (36%) of GP practices had <100 users, and it would not make sense to group these under one virtual practice. Clustering was thus done by hospital only, by adding the hospital ID to the model as a random intercept. The regression model was run using the SAS V.9.4 GLIMMIX command. Missing values for each predictor appear in table 1. Users with missing values were excluded from the regression model.
Demographic characteristics for children under 16 years old who attended an ED at least once in 2015/2016, n=2 308 816
Findings
A total of 3 272 864 unplanned attendances of children under 16 years old occurred in major EDs in 2015/2016, made by 2 308 816 users who attended at least once in 2015/2016. The most common groups were 1–4 years old, males, users living in deprived areas and white ethnic group (table 1). A total of 210 102 (9.1%) attended four times or more and were categorised as FA.
Characteristics of frequent users
Data of 1 987 032/2 308 816 (86.1%) were analysed after excluding missing values for the explanatory variables (table 1). Infants under 1 year old had the greatest likelihood of being an FA (table 2). Boys had a slightly greater likelihood than girls. Living in a more deprived area was associated with a greater likelihood of being an FA. Asian and mixed ethnic group had greater likelihoods relative to white ethnic group, while black and ‘other’ had a lower likelihood relative to white ethnic group.
ORs and their 95% CIs for being an FA (n=1 987 032 children <16 years old)
Discussion
One in 11 children (9.1%) who attended an ED attended four times or more in a year. Infants, boys and children living in more deprived areas had greater likelihood of being an FA.
Estimates for frequent paediatric ED attendance vary widely in different countries, from 0.5%32 to 75%,21 reflecting wide definitions and variability in measurement methods as well as real differences in volumes of FAs. The predictors found in this analysis appear to be consistent with existing literature, suggesting FA among infants and the very young.21 It is unclear from the literature whether sex is a consistent characteristic of FA,20 26 and the effect in our findings was small. Similarly, there are conflicting findings about the links between paediatric FA and social disadvantage.33 34 There are also mixed results in the literature regarding the association between ethnic group and FA.20 21 26 33 These findings are strongly dependent on particular cultural and health system settings, which are different, for example, between the USA, Europe and the UK. In the UK cultural context, Asian or Asian British ethnic origin were the most prevalent non-white group among paediatric FA.26
Strengths and limitations
Our data cover a national population, and it is among the largest studies on frequent ED attendances to date. The study captures a national snapshot of frequent ED attendances in England. We implemented an algorithm to calculate the number of FAs, which is able to capture attendances in the ‘tails’ of a given year rather than a simple count of the number of attendances per year.
Whereas a national dataset such as the one we used can provide a generalisable view of ED activity nationally, a full understanding the underlying health systems and local provisions required granular and richer local data. Whereas there are some common models, such as a GP-led or nurse-led triage to ED, each hospital has a different emergency care configuration. For example, in some hospitals, there are separate pathways for adults and children, while in others there is only one pathway for all ages.
Lack of valid clinical data in the HES A&E dataset limit the ability to probe the presence of comorbidities, diagnoses and clinical reasons for attendance, and to better adjust for case mix.
Implications for research
Further examination of the underlying causes of FAs among young children could help devise interventions and innovative service configurations to provide better care. Future research is needed to better understand the impact of social determinants on FA risk such as cultural factors, education, housing and family setting, as well as geographical variations, proximity to the ED and access to primary care.35 Likewise, the outcomes of the frequent hospital attendances (in terms of admission, discharge, referral and death), as well as the length of stay at the ED, could be examined in future studies. Further investigation into the presence of comorbidities and triage diagnoses is crucial to better understand the clinical reasons for FA. It would also be of interest to examine whether there is evidence that the rate of FAs is rising, amidst evidence showing that attendances are ever increasing.4 Lastly, knowledge of the effectiveness and cost–benefit of interventions specifically aimed at reducing FAs is imperative.
Implications for practice
Specific attention to FA should be given in policy and strategy in providing adequate and efficient response to their needs, whether clinical or psychosocial. Although children with good levels of GP access are less likely to attend EDs,35 still about 12% of paediatric ED attendances are deemed inappropriate.2 This means that most paediatric ED attendances, whether incidental or frequent, are appropriate, and thus efforts should not merely focus on reducing FAs or diverting them to community primary care. Integrated, whole-system care models for infants, children and young people are showing promise in reducing avoidable ED attendances through better case management between specialists and primary care.6
In seeking to understand FAs in children the situation is complicated by reliance on the main carer’s health seeking behaviour as well as the child’s underlying medical condition. For example, parental depression is associated with increased utilisation of ED, outpatient and inpatient services by children and young people, and with increased GP consultations among adolescents.36
Currently, a very poor evidence base exists for transferable interventions to reduce ED attendances in a paediatric population, and there is inconclusive evidence to support any intervention aimed at reducing subsequent non-urgent attendances following a non-urgent attendance. The long-term impact of interventions is limited, although the effect may be maximised if delivered by primary care providers in children identified after their ED attendance.37
Conclusions
Infants and children residing in deprived areas have greater likelihood of being an FA. Interventions that support parents and contribute to reducing avoidable FAs, particularly among infants, are crucial to provide appropriate response to users with recurring health needs.
References
Footnotes
Handling editor Simon Carley
Contributors GG was involved with conception and design, conducted the data analysis and drafted the manuscript. MB was involved with acquisition of funding, conception and design, data analysis and revised various versions of the manuscript. PPA was involved with acquisition of funding conception and design, data collection and analysis, and revised various versions of the manuscript. SS was involved with acquisition of funding, conception and design, data analysis, and revised various versions of the manuscript. AM was involved with acquisition of funding, conception and design, data analysis, and revised various versions of the manuscript. AB was involved with acquisition of funding, conception and design, data collection and analysis, and revised various versions of the manuscript. All authors read and approved the final manuscript.
Funding This report is independent research supported by the National Institute for Health Research Applied Research Collaboration Northwest London. The views expressed in this publication are those of the authors and not necessarily those of the National Institute for Health Research or the Department of Health and Social Care. The Dr Foster Unit is an academic unit in the Department of Primary Care and Public Health, within the School of Public Health, Imperial College London. The unit receives research funding from the National Institute of Health Research and Dr Foster Intelligence, an independent health service research organisation (a wholly owned subsidiary of Telstra). The Dr Foster Unit at Imperial is affiliated with the National Institute of Health Research (NIHR) Imperial Patient Safety Translational Research Centre. The NIHR Imperial Patient Safety Translational Centre is a partnership between the Imperial College Healthcare NHS Trust and Imperial College London. The Department of Primary Care & Public Health at Imperial College London is grateful for support from the Imperial NIHR Biomedical Research Centre. SS is funded by the National Institute for Health Research School for Public Health Research (NIHR SPHR) and the NIHR Applied Research Collaboration (ARC). The NIHR School for Public Health Research is a partnership between the Universities of Sheffield; Bristol; Cambridge; Imperial; and University College London; The London School for Hygiene and Tropical Medicine (LSHTM); LiLaC—a collaboration between the Universities of Liverpool and Lancaster; and Fuse—The Centre for Translational Research in Public Health a collaboration between Newcastle, Durham, Northumbria, Sunderland and Teesside Universities.
Disclaimer The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Patient consent for publication Not required.
Ethics approval We have approval from the Secretary of State and the Health Research Authority under Regulation 5 of the Health Service (Control of Patient Information) Regulations 2002 to hold confidential data and analyse them for research purposes (CAG ref 15/CAG/0005). The London - South East Ethics Committee (REC ref 15/LO/0824) has approved the use of this data for quality measurement of healthcare delivery.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement No data are available. We are unable to share any patient data.