Where are the patients? The factors affecting the use of emergency care during Covid-19

The Covid-19 pandemic has rapidly changed the narrative about many issues in health and social care over the past three months. One of the most striking changes has been, in the midst of a pandemic, the substantial and sustained drop in the number of attendances at emergency departments (EDs) across the UK.

Visits fell by 29% in March and 57% in April (compared with the same time in 2019) – the latter representing an absolute reduction of 21,000 patient attendances per major ED. This is a striking contrast to the annual average increases of 2-5% per year experienced over the past two decades.

Data from the Office of National Statistics also reported a sharp spike in the number of weekly deaths – with only 44% being directly attributable to Covid-19. This has raised concerns that patients with health conditions, such as myocardial infarction, stroke and new cancer diagnoses, have not been seeking the care they need for fear of catching the virus in hospital.

Some predicted that the number of patients needing emergency care could increase over the course of the pandemic, as patients with chronic disease decompensate and people with previously undiagnosed conditions present with major complications in the absence of usual health care. This was coupled with a warning that the system could be severely stretched by pent-up demand for care after lockdown ends.

In this piece, based on the extant literature and my own clinical experience of working in a London teaching hospital over the past three months, I review a range of factors that might be impacting on the use of emergency health care services during the pandemic. They include changes in how the system is structured, the impact of disasters on people’s health, and how individuals might be modifying aspects of their own behaviour during the crisis.

They will suggest, in my view, that the drop in ED attendances is likely to be real, do not necessarily represent pent-up or deferred demand, and that aspects of this may be sustainable in the longer term.

Who has stayed away and why

Analysis of the changes in ED presentations has found that most of the drop was in lower acuity groups and that the rate of presentation of seriously ill patients also fell, albeit to a lesser extent. No in-depth analysis has yet been conducted to explore which groups of patients have been implicated in these changes, but here I make a few suggestions.

Presentations due to trauma, drugs and alcohol

A striking feature of the pandemic internationally has been the fall in presentations relating to trauma, accidents, and drugs and alcohol.

Although national data for England for trauma service activity has not yet been released, local reports have found a marked reduction in the number of surgeries performed for emergency trauma.

This is not surprising given the decline in the more common causes of trauma – road traffic accidents (which usually account for around 30% of trauma calls), high falls (approximately 15%) and assault (around 8%). Since lockdown, traffic levels have fallen to levels not seen since 1955, construction sites (usually associated with high falls) have closed and violent crime has dropped sharply.

While the drop in trauma could have been predicted by the nature of the lockdown, some had suggested that presentations due to drugs and alcohol would rise, with people turning to substance abuse in response to the stressors of the pandemic.

In England, there has been a police crackdown on drug dealing in the quest to keep the streets safe, making illegal substances more difficult to obtain. And while retail sales of alcohol appear to be up, long queues and restrictions on the amount that any one person can buy could be helping to limit heavy binge drinking at home.

That trauma, drug and alcohol presentations peak on a Friday and Saturday night point to their underlying social nature, and so it follows that lockdown should result in a fall in these types of presentations.

Patients with chronic disease

Colleagues conducting virtual outpatient specialty clinics have reported that one positive result of their patients being concerned about coming to hospital has been much higher levels of engagement with all aspects of self-care, particularly medication compliance and exercise.

Many have also reported patients making minor adjustments to their own medications themselves or with the support of their GP, rather than waiting until the next routine appointment. Given the rates of chronic illness, even marginal improvements in medication compliance and engagement with self-care are likely to impact on ED presentations.

The numbers of patients with intermittent conditions – such as back pain, migraine and cyclical vomiting – who often attend the ED for symptomatic relief have also fallen over the past three months. Some report better use of medication and increasing their use of proven alternative therapies, such as physiotherapy, cognitive behavioural therapy and exercise.

The homeless

Although the homeless make up a small percentage of the overall population, they are often high volume consumers of emergency care. Homeless people account for 1-10% of ED attendances, with homeless or marginal sheltering being substantially more common in highly urban areas.

With high rates of drug and alcohol use, exposure to violence, mental health problems and medical illnesses such as tuberculosis and HIV, the homeless also tend to visit EDs more frequently, with around 30% re-attending within 30 days. Some studies have estimated that such frequent attendees account for only 8% of the patients attending EDs, but make up around 30% of the visits.

Unexpectedly, the pandemic has resulted in a unique opportunity to address the problems of the homeless, with initiatives to shelter and support rough sleepers in urban areas across the UK. There are reports of unusually high levels of engagement with this difficult-to-reach group, with many having their health and social needs consistently addressed for the first time. The hope is that many will not return to the streets and will come out of the crisis in substantially better health.

‘Social’ admissions

EDs have been seen as the default ‘safe’ place for situations where social care has collapsed and people are no longer able to be supported in their own homes. Although these patients make a small number of ED attendances on a daily basis, they also account for a surprisingly large number of bed days.

In a manner similar to the support given to the homeless, finding ways to safely manage these patients outside the hospital environment has become an even higher priority for social workers and councils.

People with mental health problems

Significant concerns have been raised about the impact of the pandemic on mental health, particularly the risk of suicide and attempted suicide, both during the lockdown and the years that follow.

It is important to note, however, that studies of suicides and suicidal ideation after natural disasters have found mixed results. A review by The American Emergency Mental Health and Traumatic Stress Service reported that suicide rates sometimes dropped, held steady or decreased after natural disasters, while suicidal ideation was found more frequently to decrease.

Similarly, the overall impact on mental health from disasters is unclear. Some studies have found increases in the overall rates of psychiatric disorders and post-traumatic stress disorder, but this seems to be limited to specific subgroups rather than being widespread. Overall, the literature suggests that the impact on any individual is the result of a complex interaction between the type of disaster, its length and severity, and the circumstances of individuals.

After the 2011 earthquake in Christchurch in New Zealand, for example, many local residents reported substantial improvements in their overall wellbeing in the years that followed.

Many hospitals over the past three months have made provisions to create separate areas to manage an expected increase in the number of patients with mental health problems outside of the main ED. In some places, these areas were subsequently repurposed when that spike did not materialise. This may change should lockdown continue and/or as economic circumstances change.

System-level factors

GP triage can be effective

One of the most marked changes to the UK health care system has been the mass conversion to telephone/video triage of patients prior to accessing both primary and secondary care.

The impact of this has been debated, with some arguing that remote consultation is insufficiently sensitive to pick up certain conditions – especially those where clinical examination may be critical to diagnosis and where there’s evidence that telephone consultations may increase work rather than reduce it. This follows the failure of a number of initiatives such as NHS 111, co-location of GPs with EDs and publicity campaigns to stem the tide of rising ED attendances.

It is worth remembering that GP triage prior to all but the most urgent hospital attendances was once the norm. Although it has been argued that only 15% of ED presentations are for GP-type conditions, in my view this switch to initial telephone triage for all general practice and Covid-related conditions is likely to reduce ED presentations even further, with little impact on clinical care.

Similar concerns were expressed in the Netherlands in the early 2010s, when a system was introduced where patients requiring emergency care were required to either see a GP or contact an emergency hotline before presenting to an ED. Although only 60% of patients use the system appropriately, ED attendances have remained relatively low without an accompanying increase in mortality for emergency conditions. Denmark also operates a system where all contact with the ED requires pre-authorisation by a GP, telephone advice line or ambulance service.

We still have a functioning health care system

The ‘third’ wave of decompensated chronic illness has been well documented across a range of disasters, including earthquakes, hurricanes and tsunamis (first wave being injuries caused by the disaster and the second outbreaks of infectious diseases).

Detailed analysis, however, has shown that the severity of the subsequent ‘waves’ depends on the nexus of the type of disaster, the extent to which individuals and their housing and livelihoods are affected, and the resilience of local health care systems.

In the case of Covid-19, the UK health care system remains intact, albeit reconfigured. GPs, pharmacies and emergency departments have remained open, even if they’ve been accessed in different ways. Ambulance services have also substantially increased the number of patients seen and treated on scene, and physician response services, where emergency doctors visit patients’ homes, have been expanded. That is not to say there has not been a significant impact on some patients waiting for elective surgery, or whose care for their long-term conditions has been disrupted.

Other international comparisons may be instructive here. Although the Christchurch earthquake was severe and damage was sustained by the local hospital, much of the local health care system remained functional, with GPs responding quickly and expanding capacity.

A detailed study of myocardial infarction over the following five years found that, far from the predicted outbreak of heart disease occurring, there was no change in the rate of cardiovascular disease outcomes in the least affected areas. There were 66 excess admissions from those living in the worst affected areas in the first year after the earthquake, with none in the four years after that.

Less iatrogenic harm

There are very major concerns about patients not receiving potentially life-saving interventions for a range of conditions during the pandemic.

However, reductions in the work of a hospital have paradoxically been found to be associated with falls in mortality. Retrospective review of doctors’ strikes held between 1976 and 2003 consistently found a reduction in mortality, even with strikes lasting up to 17 weeks. Subsequent studies have found similar results.

How might this observation be linked to ED presentations? Firstly, the practice of modern medicine is inherently risky, which will contribute directly to ED presentations after the index hospital stay. The re-attendance rate for patients undergoing surgery is, on average, 11%, rising to 15% for high-risk vascular procedures.

Readmissions due to adverse drug reactions to new medications that started in hospital occur in over 20% of medical patients. Chemotherapy also causes harm and can increase, rather than reduce, short-term morbidity and mortality.

So while reducing the number of patients receiving interventions may cause longer-term harms in some patients, it will substantially impact on short-term outcomes and ED presentations.

It is striking that Australia, which has suffered few deaths from Covid-19, may also be experiencing a significant drop in mortality associated with the cessation of routine medical work.

Conclusion

It is undoubted that patients have suffered and died, both directly and indirectly, from Covid-19. Most at risk are those whose urgent treatments have been delayed because of the cessation of routine work, and those who avoid seeking care because of the fear of contracting the virus in hospital.

However, in my view the drop in ED attendances is likely to reflect a combination of real decreases in the prevalence of certain types of presentations and exacerbations of chronic disease, as well as pointing to more appropriate routes of accessing care.

As lockdown restrictions continue to be lifted, we will see whether numbers will rebound or whether some of the reduction in attendances can be sustained. That would allow the system some much-needed relief and the space to clear the backlog of diagnostics and start up suspended services, such as elective surgery and cancer therapies, thereby ensuring that any damage to the nation’s health is minimised.