Impact of COVID-19 on ED attendance in the initial and later phase of the pandemic
The COVID-19 pandemic has posed a dramatic, unexpected, and unprecedented burden on many healthcare systems worldwide. Universally decreased numbers of ED visits by age, disease categories, and the characteristics and case severity of patients have been revealed by anecdotal reports [18,19,20,21,22,23,24,25,26,27,28,29, 31,32,33,34,35,36]. Most studies originated from severe pandemic areas that analyzed the change in the early stage (4–8 weeks) of the COVID-19 outbreak and reported a significant decrease, ranging from 31.0 to 63.5%, in ED patient flow compared to a historical control period [18, 19, 21, 23, 27, 31, 34]. During the initial phase, we found some similarities regarding a reduction in in ED visits, especially in pediatric and ambulatory patients, among the compared countries where the pandemic was serious [20, 21, 26, 27, 31]. In February 2020, most parameters used to measure the input of patient flow in our ED dropped sharply and reached their lowest point in April. Other tertiary hospitals and the largest healthcare organization in northern Taiwan found that ED visits decreased, similarly, from February to April 2020 [29, 33, 37].
In the later phase till the end of 2020, compared with the same period during the previous year, patient visits gradually recovered, but not completely. Throughout 2020, ambulatory visits, lower-urgency acuity (level III) cases, and the pediatric division still saw the greatest impact. The impact significantly disproportionate to age would correspond to the effect of these combinations. Compared with adults (Fig. 4a) and pediatrics (Fig. 4c), the trauma (Fig. 4b) division had the least reduction in the number of patients and recovered sooner, probably because there were few activity restrictions in Taiwan , and traumatic events usually have little to do with the spread of diseases.
Possible reasons for the decline in ED attendance
Despite a low prevalence of COVID-19 in Taiwan, the visits to the ED declined significantly in all divisions to various degrees. This phenomenon could not be explained simply by lockdown and sports restrictions, which clearly led to the reduction of traumatic cases [22, 38], since restrictions were not strictly executed in Taiwan . The reason why these patients did not visit the ED remains unclear; however, fear of contracting the virus, according to many observations worldwide, may be the cause [38,39,40,41], and this was also emphasized during the SARS epidemic [2,3,4,5, 42]. The public perception of fear was also demonstrated by the unprecedented increase in tweets expressing fear sentiments related to COVID-19 in January 2020 . In Taiwan, the fear of COVID-19 may not be as high as that of SARS because the death rate and the death toll from COVID-19 infection in Taiwan were much lower; nonetheless, the impact on both ED censuses was similar. Therefore, the observation implied other reasons for this decline. During the COVID-19 pandemic, the emergence of telemedicine [44, 45] may have kept pediatric and lower-urgency patients away from attending the ED. However, this twenty-first century advance is unlikely to affect the decline of in our ED census, since telemedicine was not yet widely applicable in Taiwan . The impact of influenza, SARS, and COVID-19 pandemics on ED visits is most pronounced in the pediatric and lower-urgency categories [2, 26, 47]. In contrast to influenza epidemics , the volume of the ED census during the SARS and COVID-19 epidemic decreased [2,3,4,5, 18,19,20,21,22, 26]. There is a high incidence of visits for upper respiratory diseases and infections, trauma, and the youngest age groups (< 5 years) in typical pediatric EDs [48, 49]. Respiratory tract infections account for the most common causes of pediatric treat-and-release ED visits . Therefore, public health strategies that help to control the spread of flu-like illnesses should reduce ED visits. For example, lockdowns could lead to a reduction in the number of pediatric patients presenting an airborne infectious disease . Coincidental with the COVID-19 pandemic, the case numbers of 14 airborne/droplet-transmitted infectious diseases decreased between January–October 2019 and January–October 2020 in Taiwan, with a reduction of 28.2% . This might have contributed to a decrease in ED attendance, and even more so in the pediatric division. In addition to fear, a significant decrease in commonly transmitted respiratory diseases achieved by improving personal hygiene and protection may play an essential role in decreasing future ED attendance by those with relatively low acuity as well as pediatric patients.
The reduction in ambulatory visits and patients with lower acuity can be reasonably attributed to public fear of contracting the virus and a diminished transmission of respiratory diseases, but the numbers of higher acuity visits (levels I and II) also decreased to a similar degree. The unchanged ratio of patients within the top three acuity levels shows that the scale of our ED shrank in general compared to the pre-pandemic period. The ED divisions that often had patients with hospitalization requirements such as those with higher acuity, the elderly, and traumatic patients, also dramatically declined at first. Most of these patients likely arrive via non-ambulatory means, such as transport by EMS or transfer-in via ambulance. Since there was no decrease in EMS demand, we assume that, taking the decreased numbers of transfer-in patients into account, primary and secondary medical facilities also faced a reduction in the number of patients, which allowed them higher capacity to treat severe patients. We also assumed that among chronically critically ill patients, the choice of hospice at home or local nursing facilities increased. Where patients with higher acuity issues went warrants more comprehensive investigation. Unlike the decrease in the number of higher acuity visits, the number of patients with OHCA did not significantly change, which is different from areas devastated by the outbreak, where an increase in the percentage of EMS-attended deaths and OHCA was observed [52, 53]. It is possible that the number of impending death cases related to severe COVID-19 infections in our community were minimal and not unusual after the pandemic.
Effects of COVID-19 on ED crowding in the initial and later phase of the pandemic
In the initial phase, despite a significant decline in total ED attendance except for the highest urgency (level I), the LOS of both discharged and hospitalized patients did not decrease. The D2d of acuity level II visits also did not shorten. These observations may imply that the novel disease, lacking information regarding its manifestations and transmission, may complicate the process of decision making and disposition made by physicians. Furthermore, the numbers of boarded beds in the hospital, usually an important factor affecting throughput and output, did not change significantly throughout the year, indicating that there was no pressure of loading reduction or even shut down in our hospital due to the impact of the pandemic. Finally, the percentage of adverse outcomes of our ED patients increased. The response capability of our medical team and administration seemed to be compromised early in the pandemic, but it could have been worse if COVID-19 cases surged simultaneously. A misjudgment could result in the implementation of cost-ineffective strategies that lead to a waste of human resources such as understaffing in screening and caring for patients with suspected COVID-19 infection with the pediatric department being overstaffed because of a drop in pediatric patient visits. A performance review of throughput and output at the ED will be essential, especially early in the pandemic, even in the low prevalence areas, to prevent the collapse of emergency medical care.
Intriguingly, a reverse change was noted in two throughput factors, D2d and LOS, in the later phase of the pandemic. Although it was short already, the D2d at our ED still demonstrated a significant decrease, making it an objective indicator of the performance and efficacy of ED staff. In addition, the LOS of hospitalized patients shortened significantly in all divisions, with a greater degree in the adult division, whereas the LOS of discharged patients shortened significantly only in the adult division. Although the reduction rate of discharge was higher than that of hospitalization, the reduction of LOS was significant among hospitalized patients rather than discharged patients. The divisions with longer LOS caused by overcrowding and more requirements for hospitalization seemed to benefit more from the overall reduction in patient numbers. Interestingly, the pediatric volume of patients decreased by nearly half, improving the LOS of hospitalized children rather than that of discharged children. This suggests that for divisions without overcrowding and prolonged LOS, a simple reduction in attendance, of which most cases are ambulatory and lower urgency, will not necessarily improve the LOS of discharged patients. The LOS affected less significantly was usually for categories approximately or less than 4 h. This outcome implies that patients who stayed longer hours at the ED contributed to crowding, which was previously mentioned . ED physicians could pay more attention to patients who needed more time for work-up when unnecessary visits decreased. Except for the universal decrease in the influx of patients, the decrease in hospitalized patients from the ED contributed to relieving the pressure of admitted output. In the United States, despite a decrease in total encounters in many EDs, the LOS of patients was longer, thus attributing to surges in COVID-19 cases . The overall improvement of crowding at our ED in a low-prevalence scenario also helped to reduce the incidence of URVs rather than IHCA and DAMA. Not only did physicians have extra time to evaluate medical problems, but the reluctance of patients to seek medical care could be another reason. The unchanged ratio of high urgency acuity patients may account for the stable incidence of IHCA and DAMA.
Experience with COVID-19 in a low prevalence region
Our ED did not experience a surge in COVID-19 cases that would contribute to a longer LOS , there should theoretically be no notable difference when compared to the pre-pandemic period. However, there was a prolonged reduction in ED visits that could be attributed to an improved medical seeking behavior from an awareness of the importance of personal protection against COVID-19 rather than an alleviated fear of the pandemic because of the apparent drop in the number of cases with confirmed infection. Hence, we were able to improve the long-term crowding problems in non-trauma adults. Nevertheless, taking into account the gradual increase in ED attendance (Fig. 2), maintenance of a sustainable behavior in seeking emergency healthcare remains vital at the later stage of the pandemic to prevent a prolonged LOS similar to that before the pandemic (Fig. 5). An analysis of the change in characteristics of patient flow in an emergency setting throughout the pandemic in the current study may shed light on the role of reducing unnecessary visits in the alleviation of ED overcrowding. In this way, the implications of our study in a low COVID-19 prevalent area like Taiwan might provide a useful reference for health administrators and hospitals to improve the condition of ED overcrowding in the post-COVID-19 era.
Comparison with SARS experience
Tracing the history of coronavirus pandemics, literature on the impact of the SARS epidemic in 2003 on the EDs at three tertiary hospitals from northern, central, and southern Taiwan were also reviewed [2,3,4,5]. The hospital from the south was our hospital (KSVGH). Compared with the pre-epidemic period, a month after the highest outbreak time in April 2003, the largest average declines in monthly visits were 51.6, 33.4, and 40%, respectively. The changes were also more significant in lower acuity and pediatric patients than in higher acuity and trauma patients. During the observation in 2003, one of these hospitals found that the total ED patient volume took a long time to return to the previous level . Although the degrees of devastation caused by the SARS and COVID-19 epidemics are different, the immediate and long-term impact on visit numbers and characteristics of patients in the ED seem to be quite similar. Therefore, this phenomenon may repeat at the next novel coronavirus pandemic. The results of an analysis from a tertiary hospital in northern Taiwan could also support this hypothesis . The historical experience from SARS to COVID-19 suggested that the unprecedented global coronavirus pandemic could be a rare but essential factor associated with the decision of non-urgent ED use that had not been previously mentioned in a systemic literature review .
Our study has several limitations. First, it is a single-center study that may not be generalizable or applicable to other emergency medicine institutions, since the hospital’s administrative policies could greatly affect the results, such as the LOS of patients. Nonetheless, the universal decrease in the need for medical care without increasing the crude death rate officially released by Taiwan’s government  was consistent with some of our findings. Second, since the cut-off point was set on January 31 instead of January 21, the date of the first reported case, comparing the changes in the initial epidemic phase between January and February 2020 was underestimated. Third, we did not adjust for multiple potentially confounding factors related to decreases in the ED flow other than the COVID-19 pandemic. Further analysis of different diagnostic categories may identify additional details about confounding factors. These factors may be influenced by decreased social distance and improved hand hygiene, resulting in a decrease in visits due to causes such as flu or gastroenteritis.