• emergency department

In their EMJ manuscript, Burton and colleagues apply clever mathematics to the challenging issue of ED frequent attenders.¹ We must disclose here that we are not experts in the sort of analysis used and we suspect that applies to many readers. As a result, it may be tempting to give this only cursory attention and to move on to more familiar methodologies. We would urge a second and deeper look.

Emergency medicine is moving into an era with increasing access to very large datasets of linked data that are relevant to our practice. Analyses such as this are likely to become more commonplace. Importantly, they provide insights that may not be possible with more traditional research approaches and additionally, the ability to evaluate, at scale, the real-world impact of interventions.

As the authors’ note, ED frequent attendance appears to occur everywhere. It incorporates a complex mix of human beings who have a variety of problems which may be societal in addition to medical. The authors hypothesise that the pattern of frequent attendance within a regional urgent and emergency care system would show features of a complex system and explore this using power law analysis.

We may be familiar with a simple power law effect from school mathematics where, when we doubled the length of the side of a square, its area quadrupled whatever the length was in the first place. We may also experience a related effect when gentle braking in a car well ahead of us in a traffic stream results in our stopping and waiting in a traffic jam (sometimes called an accordion or concertina effect).

The study by Burton and colleagues demonstrates a power law relationship between the number of times an individual attends an ED and the proportion of all ED attenders who attend at least that number of times in a year. This power law behaviour emerges from the interaction of many parts—this emergent behaviour, rather than preplanned, is typically used to define a complex system. Economies, ecosystems and the climate are all examples of complex systems in which the overall properties of the system arise from the multifaceted interaction of the components. So, rates of frequent attendance in the ED are entwined with (and significantly impacted by) large-scale drivers within the complexity of the health system.

The over 70 age group did not appear to fit a power law. This is shown by a non-linear (convex) distribution appearance (figure 1d in the Burton et al manuscript). While the authors suggest that this identifies that frailty-related frequent attendance may be better understood at the level of the individual, there is a presumption here that the 70 plus group are necessarily presenting with frailty-related issues. As the study did not record presenting complaints, (and given the variation present with these), one should not assume that the age bracket necessarily implies that the problems are either comorbidity, long-term condition or frailty associated.

Crucially, interventions aimed specifically at individuals are unlikely to change the overall frequent attendance rates (though individuals may be impacted), because they are only one component of the system. It is important to note that the study is not saying—‘don’t bother, you can’t do anything about the frequent attenders’, rather, it is helping us see that while we may affect the individual, we will only make meaningful volume change through changes in the system itself. If the first braking vehicle causing a concertina traffic jam is red and one bans all red vehicles, then this is unlikely to prevent future similar standstills. Additionally, we are unlikely to be able to predict or effect meaningful change on the rates of frequent attenders by tweaking a single lever. This has also given us a tool by which to measure the effect of system-level changes—changes that may occur over long periods of time or which are forced on us through external circumstances (like a pandemic).

Finally, we believe that there is a valuable human angle here. At the end of a weekend of late shifts, it can sometimes be hard to feel positively about someone’s umpteenth visit to the ED that month. However, we would not wish to be ‘wearing the shoes’ of a frequent attender and this study will remind us that they are ‘caught up’ in a complex system effect.