Examining patterns of urgent and emergency care service use by children and young people

Background
Demand for urgent and emergency care services such as emergency departments (EDs) is rising in the UK. Concerns have been raised regarding the consequences of increased strain on these services for specific patient groups such as children and young people. For parents and caregivers, deciding the most appropriate healthcare service for their child can be challenging, as children and young people are often extremely vulnerable with specialised healthcare needs. To assist parents and caregivers with this decision making process, the telephone triage service NHS 111 is tasked with providing callers with healthcare advice and recommending relevant healthcare services. However, there is evidence to suggest that the telephone triage service is leading to a number of avoidable ED attendances, further increasing the strain on other urgent and emergency care services. Although the use of pre-hospital services such as NHS 111 is a pivotal component of a child’s urgent and emergency care journey, another important consideration is those who frequently use urgent and emergency care services such as EDs. High frequency ED attenders often represent a small proportion of attenders but make a relatively large number of attendances. The reasons behind why these children and young people frequently attend EDs consists of confounding factors such as parental anxieties and reduced access to other healthcare services. Examining these factors in detail can highlight interventions aimed at improving the management of high frequency ED attenders within the ED, but may also improve the safety of these individuals outside the urgent and emergency care system. The overarching aim of this PhD is to analyse routinely collected patient data to provide insights into the use of urgent and emergency care services such as EDs by children and young people. This is achieved by analysing parental/caregiver, socioeconomic and medical factors associated with a child’s urgent and emergency care journey.

Overarching Aim
To use routinely collected patient data to provide insights into the use of EDs and other urgent and emergency care services by children and young people.
Primary Research Objectives
1. To understand more about the relationship between NHS 111 pathways and ED outcomes for children and young people.
2. To determine the statistical distribution underpinning the number of ED attendances children and young people make, and examine methods of assessing factors associated with high frequency ED attendance.
3. To explore factors associated with high frequency ED attendance by children and young people.
4. To measure ED attendance patterns and examine heterogeneity in reasons for attendance in children and young people.

Methods
This quantitative PhD analysed routinely collected patient data collected within the Yorkshire and Humber region of the UK, stored within the CUREd research database. Two systematic reviews were conducted to explore what had been previously published on the use of telephone triage services prior to an ED attendance in children and young people, and high frequency ED attendance by children and young people respectively. Standard approaches for systematic reviews were adopted and reported against PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines. A variety of statistical methods were implemented to address the above research objectives. These include a number of statistical models such as logistic, zero truncated poisson, quantile and mixed effects regression models, and statistical tests such as likelihood ratio and chi-squared tests. Other statistical approaches include the use of a Herfindahl index to measure the heterogeneity of reasons for ED attendances and the implementation of an expectation-maximisation algorithm to model the time taken for an individual to attend ED after contacting NHS 111.

Results
The telephone triage systematic review concluded that these services may be leading to a number of inappropriate ED attendances by children and young people (especially when the call did not involve input from a clinical expert). It also highlighted a gap in the evidence exploring urgent and emergency care pathways made by children and young people following contact with NHS 111. The high frequency ED attender systematic review found that studies often use a numeric threshold (ranging from 2-6 attendances made in a year) for defining high frequency ED attenders. This highlighted the need for a novel statistical approach to assess factors associated with high frequency ED attendance without needing to categorise high frequency ED attenders in this way. This systematic review also revealed a gap in the evidence for exploring high frequency ED attendance by children and young people over multiple years in addition to the heterogeneity of reasons for attending. The first study conducted in this PhD analysed 348,401 calls made to NHS 111 regarding children and young people, and found that 45,746 (13.3%) led to an ED attendance within 24 hours. The time taken for individuals to attend ED following contact with NHS 111 could be categorised into two relatively distinct waves, with most attending soon after the call and a proportion attending later. Of those attending an ED after being told to do so by NHS 111, over a quarter resulted in a low acuity attendance. Young age (<1) was found to be associated with low acuity ED attendances following contact with NHS 111, irrespective of the disposition provided during calls. The second study conducted in this PhD analysed the yearly ED attendances made by children and young people and provided evidence to suggest the underlying distribution follows a heavy tail distribution such as the discrete power law or lognormal. This study proposes the use of a zero truncated or one inflated zero truncated poisson model in combination with a quantile count regression model to analyse this in children and young people. As this distribution was found to be heavy-tailed in nature this provided evidence to suggest ED use by children and young people represents characteristics of a complex system. The third study conducted in this PhD found that high frequency ED attenders were more likely to be admitted to hospital after attending an ED, in comparison to occasional users. When considering factors associated with ED attendance rates in ED users, younger age groups and greater levels of deprivation were found to be associated with higher rates of attendance. This association was amplified in those attending the ED most often. Ethnicity was also found to be associated with ED attendance rates. The final study conducted during this PhD analysed ED attendances made by children and young people over a two-year period. Although a relatively small proportion of individuals were defined as high frequency ED attenders in their first observation year (9,700, 13.6%), over half of these individuals made at least one attendance in their second observation year (5,345, 55.1%). The analysis of reasons for attendance in individuals (making 7-13 ED attendances in the observation period) found those aged 8-12 were more likely to make injury related attendances and less likely to make illness related attendances. Conversely, those under the age of 1 were less likely to make injury related attendances and more likely to make illness related attendances. There was less heterogeneity in the reasons for attendance in those aged 8-12 in comparison to those under the age of 1.

Conclusion
The factors associated with the use of urgent and emergency care services by children and young people represent a complex interplay of medical, socioeconomic and parental/caregiver decision making considerations. This PhD has demonstrated how large quantities of routine patient data can be used to provide insights into the use of urgent and emergency care services by children and young people. However, when using routine patient data it can be challenging to distinguish between high frequency ED users making clinically appropriate ED attendances for reasons such as chronic conditions, and those making less clinically appropriate ED attendances due to factors such as parental anxiety and reduced access to primary care services. Although targeted interventions should be implemented to assist parents and caregivers when making health-related decisions for their child, as the urgent and emergency care system comprises a number of interacting components, these interventions must also consider the wider healthcare system and beyond.