The study by Naouri et al in this issue of BMJ Quality and Safety describes an ambitious, 24-hour cross-sectional physician survey and chart review of all the emergency departments (ED) in France to characterise the ‘inappropriateness’ of ED visits.1 The determination of inappropriateness for any given visit was based on (A) physician opinion, (B) physician determination of ambulatory care sensitivity, or (C) resource utilisation. Based on these measures, the authors concluded that between 13% and 27% of ED visits were inappropriate. Further, patients with supplemental public insurance (a proxy for the socioeconomic disadvantaged in France) were 15%–33% more likely to use the ED inappropriately.

Naouri’s study is part of a growing body of literature that characterises ED use as inappropriate, avoidable, ambulatory care sensitive or preventable.2 3 While there is precedent and potential merit in classifying healthcare services based on their value,4 this trend raises some concerns—as the authors rightly conclude—when describing the use of EDs by disadvantaged populations.

Determining the patient’s ED visit as inappropriate, without consideration of broader contextual factors, is an example of attribution bias: the tendency for people to overemphasise individual and personality-based explanations for behaviours while underemphasising situational explanations.5 We may blame patients for visiting the ED inappropriately, when in reality, healthcare systems are often designed to funnel patients towards the ED. A historical perspective may be useful in justifying this assertion.

For generations, there has been a tremendous imbalance in the revenue generated by hospital versus outpatient services. On an aggregate level, hospital stays in the USA generated US$27 billion more than outpatient visits in 2017.6 Similar disparities in revenue generation between inpatient and outpatient care exists across the globe irrespective of a country’s healthcare financing structure.7

This lopsided revenue has incentivised healthcare systems to funnel patients into hospitals instead of preventing illness in community or practice-based settings. Primary care practices and community-based services such as home care have fewer resources to manage the growing complexity of ageing, chronically ill and comorbid populations.8 High-quality primary care is even more elusive for patients with socioeconomic risk.9 Naouri’s study supports these conclusions: patients visiting EDs often did so because they could not find an available general practitioner or were referred to the ED by their general practitioner.

By contrast, EDs—gateways to the revenue-generating hospital—are convenient one-stop shops replete with resources. So, it should not surprise us when patients—especially those with social risk—visit the ED. This observation is a reflection of the system that we have built and a side effect is ED overcrowding.

So, where do we go from here? If we approached the problem of excessive hospital use solely as the consequence of patient choice, we might continue our current tiring paths. We could impose cost sharing on patients for ‘inappropriate’ ED visits. These strategies have reduced ED use10 but may result in unintended health and financial harms to patients by creating financial barriers to care, particularly among the highest risk individuals. Alternatively, we might continue designing gatekeeper strategies such as nurse access lines employed by insurance companies. These approaches raise the spectre for discrimination and inequity, as minorities and disadvantaged patients commonly use the ED and are the most likely to be diverted. Indeed, this concern is significant enough that in the USA, patients have long been protected under the US Emergency Medical Treatment and Active Labor Act, preventing EDs from turning patients away for any reason.

We recommend three key considerations for healthcare provider organisations as they try to address overcrowding in the ED.

First, hospitals and administrators should recognise the potential for implicit or explicit bias when determining the inappropriateness of ED visits based on narrow clinician perspectives. If we instead assume that patients are usually the best judge of when and where to seek care, we might reinterpret the following finding from Naouri’s study: 25% of the time, ED physicians deemed patient visits inappropriate. While Naouri et al concluded that inappropriate ED visits were higher among socially disadvantaged patients, socially disadvantaged patients might have been victims of clinicians’ biased judgements.

Second, a more user-centric approach would develop alternative paths of care that mirror the convenience and timeliness of EDs. Urgent care centres, retail health clinics and telemedicine have become exponentially popular in the USA because of consumer demands. To compete, traditional healthcare provider organisations have enhanced access to outpatient practices with extended after-hours care. However, with some exceptions, the expanded time availabilities have not been matched by expanded clinical infrastructure (ie, on-site laboratory testing and imaging). So, patients often sense that even when they get timely appointments, practices are ill equipped to help them.8 Ultimately, efforts to grow the clinical capacity of practices may have constraints. Provider shortages are a global challenge and working weekends and late evenings are unlikely to mitigate the growing wave of burnout and depression among providers. And efforts to outfit practices with more clinical infrastructure may reach a point of diminishing financial returns. Centralising these resources in the ED or hospitals may in the end be the most efficient approach.

Third, if our goal is truly to reduce ED use, then it makes sense to invest in interventions that address the underlying socioeconomic issues—hunger, joblessness, trauma—that often drive acute illness and ED utilisation in the first place. There is a growing literature that suggests that health system-based socioeconomic interventions11 12—including support from community health workers13–15—can decrease hospital use while improving health and quality.

The problem of ED utilisation in France and the USA is probably less the result of inappropriate patient decision-making, and more the result of healthcare’s inability to adapt to the needs and preferences of patients. In the USA, changes may be on the horizon. Since the passage of the 2010 Affordable Care Act, payment has shifted towards value-based payments for healthcare services, including capitated or bundled payment structures. Value flips the incentives that exist in a fee-for-service environment: instead of doing more at greater cost, hospitals make money by doing less in higher cost settings. And so, with a changing incentive structure and studies like the one by Naouri et al, healthcare provider organisation should shift towards identifying better delivery approaches that meet patient needs in more cost-effective settings rather than blaming patients for accessing the care they need.