Emergency Department as the Entry Point to Inpatient Care: A Nationwide, Population-Based Study in South Korea, 2016–2018

(1) Background: The emergency department provides lifesaving treatment and has become an entry point to hospital admission. The purpose of our study was to describe the characteristics and outcomes of patients who were admitted through the emergency department to the intensive care unit or general ward. (2) Methods: We performed a retrospective, cross-sectional, descriptive analysis using the National Emergency Department Information System, analyzing patient data including disease category, diagnosis, and mortality from 1 January 2016, to 31 December 2018. (3) Results: During the study period, about 13.6% were admitted through the emergency department. Of these, the overall in-hospital mortality was 4.6%. The frequent disease class for the intensive care unit admissions was the cardiovascular system, and the classes for the general ward admissions were as follows: injury and toxicology, digestive system, and respiratory system. Cardiovascular system-related emergencies were the predominant cause of death among patients admitted to the intensive care unit; however, oncologic complications were the leading cause of death in the general ward. (4) Conclusions: Emergency departments are incrementally utilized as the entry point for hospital admission. Health care providers need to understand emergency department admission epidemiology and prepare for managing patients with certain common diagnoses.


Introduction
Emergency departments (EDs) have a crucial role in the health care system, acting as hospital admission gatekeepers and a place for the provision of lifesaving treatment. Recent studies have proven that immediate interventions and admission through an ED for various diseases, such as ST-segment elevation myocardial infarction, acute ischemic stroke, and geriatric trauma, could improve outcomes [1][2][3]. The utility of EDs has been increasing and has outpaced the growth of the general population. This has been caused primarily by an increase in the proportion of aging patients with greater rates of chronic diseases [4,5]. Admissions through an ED have also increased simultaneously [6]. Between 2003 and2009 in the United States of America (USA), hospital admissions originating in the ED increased by 17%, whereas admissions from physicians' offices and clinics decreased by 10% [7]. Furthermore, another study announced that 70% of hospital admissions were processed through the ED, and there was a linear association between age and the ED admission in the USA [8]. The characteristics of admissions through the ED greatly differ from those of regularly scheduled admissions. Patients requiring emergency admission were typically older, had more comorbidities, and had severe acute illnesses [9]. Moreover, there were potential differences in emergency admitted patients between the general ward (GW) and intensive care unit (ICU) [10,11]. Assessing the epidemiology of patients admitted through an ED is essential to prepare and manage hospital bed resource; however, data on disease classification, specific diagnosis, and mortality in patients admitted from the ED to the hospital are limited. Kwak et al. conducted a retrospective study and described the nation-wide ED utilization pattern by children in Korea [12]. Chen et al. compared the epidemiological characteristics and disease spectrum of patients of two EDs in China [13]. However, these studies did not provide specific diagnoses, classifications, and mortality for both the ICU and GW [12,13].
The primary purpose of this study was to describe common disease classifications and specific diagnoses in patients admitted to the hospital from the ED between 2016 and 2018. We also examined the frequent specific diagnoses in non-survival patients to provide insights for preparing future medical resources.

Study Design
This retrospective, cross-sectional, descriptive, nationwide study was conducted using the National Emergency Department Information System (NEDIS) between 1 January 2016, and 31 December 2018. NEDIS was started in 2003 by the Ministry of Health and Welfare and is managed by the National Emergency Medical Center of Korea. NEDIS covers all clinical and administrative data on patients who visit EDs throughout the country to maintain ED quality and enhance the emergency medical service system [14]. It is regarded as a reliable source of ED data due to the fact that >98% of EDs participate in the system and annual governmental review of the data. This study was approved by the Institutional Review Board of the study facility, and informed consent was waived because of the anonymous nature of the data.

Study Setting and Population
According to the Korean Emergency Medical Care Act, there are three categories of EDs, ranging from large to small regarding the size and role of EDs, i.e., regional emergency centers, local emergency centers, and local emergency institutes. The NEDIS has a variable to differentiate the hospitals according to the levels. We included all admitted patients through EDs of both regional emergency centers and local emergency centers. Patients were excluded if they had visited local emergency institutes because the characteristics of patient population and disease spectrum were totally different from others. Most local emergency institutes in Korea have no ICUs and find it hard to manage patients with critical illnesses. In addition, we excluded patients who were discharged, died, transferred to other hospitals in the ED, or had missing outcome data, such as lost to follow-up. Available variables were extracted, such as demographics, insurance status, route of arrival (direct from the scene or home, transfer from other outpatient clinic or hospital), causes of ED visits (medical or trauma), acuity at the initial triage, and location of admission. According to the Korean Triage and Acuity Scale (KTAS), acuity at the initial triage was rated between 1 and 5, with level 1 indicating the most severely ill patients [15,16]. Length of stay, in line with the previous study, was defined as the interval (in minutes) between patient arrival at the ED and subsequent departure [17].

Measures
We compared the number and rate of admissions by location (i.e., ICU and GW) and collected clinical outcomes on the admitted patients, including discharge, death, and transfer. Others (0.7%) included all unclassified cases discharged without adequate recommendations and patients who ran away. We used the final diagnosis on admission according to the Korean Standard Classification of Disease and Cause of Death (KCD) codes [18]. KCD codes are managed by Statistics Korea, a central organization for statistics under the Ministry of Strategy and Finance, and identical to the International Classification of Disease 10. Furthermore, the diseases and causes of death were classified based on Statistics Korea.

Data Analysis
All statistical analyses were performed with R statistics software, version 3.5.0 (R Foundation for Statistical Computing, Vienna, Austria). Continuous parameters were presented as means and standard deviations (S) and compared using the Student's t-test or the Mann-Whitney U test. Categorical parameters were expressed as a number and percentage and compared with chi-square tests. p values of < 0.05 were considered to be statistically significant.

Measures
We compared the number and rate of admissions by location (i.e., ICU and GW) and collected clinical outcomes on the admitted patients, including discharge, death, and transfer. Others (0.7%) included all unclassified cases discharged without adequate recommendations and patients who ran away. We used the final diagnosis on admission according to the Korean Standard Classification of Disease and Cause of Death (KCD) codes [18]. KCD codes are managed by Statistics Korea, a central organization for statistics under the Ministry of Strategy and Finance, and identical to the International Classification of Disease 10. Furthermore, the diseases and causes of death were classified based on Statistics Korea.

Data Analysis
All statistical analyses were performed with R statistics software, version 3.5.0 (R Foundation for Statistical Computing, Vienna, Austria). Continuous parameters were presented as means and standard deviations (S) and compared using the Student's t-test or the Mann-Whitney U test. Categorical parameters were expressed as a number and percentage and compared with chi-square tests. P values of < 0.05 were considered to be statistically significant.    Table 1 shows the baseline characteristics of patients admitted through the ED according to the admission location. Male sex was predominant with a mean age of 53.4 years old. Direct visits (61.1 for ICU vs. 68.9% for GW) and transfers from outpatient departments (2.5 vs. 5.2%) were frequent in the GW group, and transfers from other hospitals (36.4 vs. 25.9%) were common in the ICU group. Medical problems were prevalent in both groups (80.6 vs. 82.7%), and mean ED length of stay was significantly shorter in the ICU group than in the GW group (353.2 vs. 455.0 min). KTAS 1 and 2, called severe illness, were more common in the ICU group (10.0 vs. 0.8% for KTAS 1, 37.5 vs. 10.5% for KTAS 2).

Baseline Characteristics of the Study Population
Med. 2021, 10, x FOR PEER REVIEW 4 of 11 Figure 2. Trends of the total admission through the emergency department during the study period. ED, emergency department; GW, general ward; ICU, intensive care unit. Table 1 shows the baseline characteristics of patients admitted through the ED according to the admission location. Male sex was predominant with a mean age of 53.4 years old. Direct visits (61.1 for ICU vs. 68.9% for GW) and transfers from outpatient departments (2.5 vs. 5.2%) were frequent in the GW group, and transfers from other hospitals (36.4 vs. 25.9%) were common in the ICU group. Medical problems were prevalent in both groups (80.6 vs. 82.7%), and mean ED length of stay was significantly shorter in the ICU group than in the GW group (353.2 vs. 455.0 min). KTAS 1 and 2, called severe illness, were more common in the ICU group (10.0 vs. 0.8% for KTAS 1, 37.5 vs. 10.5% for KTAS 2).  Figure 2. Trends of the total admission through the emergency department during the study period. ED, emergency department; GW, general ward; ICU, intensive care unit.  Table 2 shows the frequencies of disease classes according to admission location. Over half (59.5%) of all cases comprised injury (17.2%), and emergencies related to the digestive system (15.2%), circulatory system (13.4%), and respiratory system (12.7%). The circulatory system (35.3%) was the leading cause of admission in the ICU, followed by injury (16.8%), and respiratory system (10.2%). Injury (17.3%) and digestive system (16.3%) were common in the GW, and a higher proportion of patients were admitted in the GW for neoplasms (9.1%) than in the ICU (4.0%).  Table 3 shows the common specific diagnoses of admitted patients. Pneumonia (0.045%) was the most common cause in the total population, followed by acute appendicitis, cerebral infarct, and acute kidney injury. Hemorrhagic stroke (0.088%) and acute myocardial infarction (0.083%) were predominant in the ICU, and pneumonia (0.043%) and acute appendicitis (0.032%) were frequent in the GW. Figure 3 shows the most common causes of death classification. Neoplasms were the leading causes of death in the total (30%) and GW (48%) group. Meanwhile, diseases of the circulatory system had the largest proportion of deaths (32%), followed by respiratory system emergencies (19%), and injury (12%) in the ICU group. Figure 4 shows the specific KCD codes for in-hospital deaths. Pneumonia was the most frequent diagnosis in both the ICU (n = 9392, 12.3%) and GW (n = 11,009, 12.9%). Sepsis (n = 4601, 6.1%) was second-most common, followed by hemorrhagic stroke (n = 4450, 5.9%), and cardiac arrest (n = 3825, 5.0%) in the ICU admission. Meanwhile, incidence of cancers, including lung cancer (n = 7279, 8.5%) and hepatocellular carcinoma (n = 4759, 5.5%), were more common than that of sepsis (n = 2279, 2.6%) in patients admitted to the GW.   Figure 4 shows the specific KCD codes for in-hospital deaths. Pneumonia was the most frequent diagnosis in both the ICU (n = 9392, 12.3%) and GW (n = 11,009, 12.9%). Sepsis (n = 4601, 6.1%) was second-most common, followed by hemorrhagic stroke (n =

Discussion
Information on common causes of admission through the ED and the cause of death provide insight into a community's health care system. Our results clearly showed (1) 13.6% (2.3% ICU and 11.3% GW) of patients were admitted through the ED, (2) cardiovascular system-related emergencies comprised more than one-third of the total ICU admission cases, and oncologic diseases were the leading causes of GW admission, (3) around one-twentieth died during admission and pneumonia was the most common diagnosis of death for both ICU and GW, and (4) cardiovascular system-related emergencies, such as acute stroke, cardiac arrest, and myocardial infarct were the predominant cause of death in the ICU; however, oncologic complications were the leading cause of death in the GW, and digestive issues and injuries were relatively less common.
In the 2017 Emergency Department Benchmarking Alliance (EDBA) report, the performance measures survey showed that the average proportion of patients admitted to hospital from the ED was 16.9%, and their length of stay was 303 min in the USA. Our study confirmed that Korea's ED admission rate is approximately 13%, which increased from 12.5% in 2016 to 13.1% in 2018. This growth was primarily due to the rise in admissions to the GW. Previous epidemiologic studies of ED utilization on the detailed categorization of patients with entire emergency admission records are limited. Recent nationwide analyses in the USA reported emergency care-sensitive conditions and showed sepsis, chronic obstructive pulmonary disease, pneumonia, and heart failure were predominant [19]. In line with previous research, we found that cardiovascular, pulmonary, gastrointestinal, and oncologic issues were the leading causes of urgent admissions. Moreover, pneumonia, cerebral infarction, acute kidney injury, and heart failure were the common attributing diagnoses on admission. These may reflect the geriatric populations with

Discussion
Information on common causes of admission through the ED and the cause of death provide insight into a community's health care system. Our results clearly showed (1) 13.6% (2.3% ICU and 11.3% GW) of patients were admitted through the ED, (2) cardiovascular system-related emergencies comprised more than one-third of the total ICU admission cases, and oncologic diseases were the leading causes of GW admission, (3) around onetwentieth died during admission and pneumonia was the most common diagnosis of death for both ICU and GW, and (4) cardiovascular system-related emergencies, such as acute stroke, cardiac arrest, and myocardial infarct were the predominant cause of death in the ICU; however, oncologic complications were the leading cause of death in the GW, and digestive issues and injuries were relatively less common.
In the 2017 Emergency Department Benchmarking Alliance (EDBA) report, the performance measures survey showed that the average proportion of patients admitted to hospital from the ED was 16.9%, and their length of stay was 303 min in the USA. Our study confirmed that Korea's ED admission rate is approximately 13%, which increased from 12.5% in 2016 to 13.1% in 2018. This growth was primarily due to the rise in admissions to the GW. Previous epidemiologic studies of ED utilization on the detailed categorization of patients with entire emergency admission records are limited. Recent nationwide analyses in the USA reported emergency care-sensitive conditions and showed sepsis, chronic obstructive pulmonary disease, pneumonia, and heart failure were predominant [19]. In line with previous research, we found that cardiovascular, pulmonary, gastrointestinal, and oncologic issues were the leading causes of urgent admissions. Moreover, pneumonia, cerebral infarction, acute kidney injury, and heart failure were the common attributing diagnoses on admission. These may reflect the geriatric populations with multiple comorbidities as the mainstay of hospital resources. One nationwide descriptive study in Korea with pediatric visits announced that 15.5% of patients were admitted to the ED and fever was the most common symptom [12]. However, they did not provide information on the detailed diagnosis or classification for admission.
We additionally found a common mortality rate and cause of death among admitted patients through the ED. One of the benefits of the NEDIS registry is that death information is included and can be used to calculate nationwide ED mortality. Even though mortality could be impacted by varying illness severity and study population, the rate was quite similar to a previous nationwide study in England (4.8%) [20]. Another retrospective multicenter cohort study showed that the mortality of acutely admitted patients increased up to 10-fold compared with the general population [21]. Our results showed that pneumonia was the most common cause of death in both the ICU and GW. Furthermore, cardiovascular accidents were common in the ICU, and cancer-related problems were frequent reasons for mortality in the GW. For example, patients with pneumonia tend to require more medical resources than other diseases, such as high flow nasal cannula, mechanical ventilator, and negative pressure rooms [22]. Moreover, patients with pneumonia could negatively impact ED overcrowding, which worsened patients' overall outcomes [22]. Regarding acute ischemic or hemorrhagic stroke, urgent interventions, such as administration of the thrombolytic agent, or hemispheric decompression, with emergent admission through ED to ICU, are known to be the single most important treatment to improve neurologic outcome [23]. Meanwhile, Majzoub et al. reported that early palliative care consultations after ED admission were associated with lower risk for hospital death, and limited palliative care resources needed to be directed to those patients who were admitted through ED [24]. These data could give additional insights for preparation of future administration.
This study has several limitations. First, the NEDIS data were collected from large EDs, therefore, the generalization of the results to small hospitals may not be appropriate. However, the significance of the outcomes could provide detailed information on medical resources. Second, the NEDIS database does not disclose each ED's clinical data, so appropriate patient management could not be confirmed. Third, we used the first NEDIS diagnosis only for both admission and death analyses. While this approach may have deprived analysis of the complete information, we assumed the initial diagnosis would reflect the full situation.

Conclusions
In conclusion, ED admissions are increasing. Healthcare providers need to understand the epidemiology of ED admissions and prepare to manage these patients with a specific common diagnosis.