Emergency department (ED) crowding is a critical issue affecting healthcare systems in many countries. [1,2] According to the American College of Emergency Physicians [3], crowding occurs “when the identified need for emergency services exceeds available resources for patient care in the ED, hospital, or both.” The input-throughput-output conceptual model is widely used to understand and manage ED crowding. It partitions ED crowding into three components: input (any factors that contribute to the demand for ED services), throughput (factors that involve processes within the ED), and output (factors affecting discharge processes) [4]. ED length of stay (LOS) is a crucial throughput measure that indicates an ED’s efficiency and effectiveness [5,6].

Previous studies have shown that several factors influence ED LOS [5,7,8]. Patient-related factors include triage level, age, need for admission, and arrival by ambulance [9]. Time-related factors indicate that LOS increases with a higher percentage of daily admissions, higher rates of elopements, more frequent periods of ambulance diversion, and on weekdays [10]. For critically ill patients, admission in winter and during the night shift is associated with prolonged LOS [11]. Hospital-related factors, such as the numbers of staffed beds and consultations, also significantly influence ED LOS [7,12]. Furthermore, prolonged LOS is associated with increased in-ED mortality among critically ill patients and those experiencing out-of-hospital cardiac arrest [13,14].

In the extensive literature on the subject, most studies focus on single hospitals or specific regions, limiting the generalizability of the findings [7,9,10,15]. Due to the ED’s role as a safety net in the community, comprehensive studies using national data are needed to develop effective policies [16,17]. This broad approach would provide a more accurate picture of the factors that influence ED LOS and help regulators create strategies for mitigating ED crowding on a national scale.

Our aim in this study was to use data from the National Emergency Department Information System (NEDIS) database in Korea to investigate the factors that influence ED LOS. By analyzing data from this extensive nationwide database, we sought to uncover trends and variables that significantly affect the time that patients spend in EDs.

This study used logistic regression analyses to identify and analyze how patient, environmental, and organizational factors and patient outcomes influence ED LOS. Elderly and male patients, medical aid beneficiaries, individuals transferred from other hospitals or referred by outpatient clinics, and those transported to the ED by 119 ambulance were more likely than others to experience extended ED stays, along with patients who visited an ED during the spring or at night. Higher level EDs, the presence of emergency symptoms, and abnormal initial vital signs (excluding systolic blood pressure and SpO₂) also correlated with increased ED LOS. Certain diagnoses (sepsis, COVID-19, and other severe conditions) were associated with longer stays. Among outcomes in the ED, transfer to another hospital, ward admission, and in-hospital mortality were significantly linked to prolonged ED LOS.

Addressing ED crowding by reducing ED LOS is pivotal for enhancing patient care. Various countries have established distinct LOS time goals to alleviate ED crowding. For example, Canada mandates an ED LOS less than 8 hours for patients with severe conditions (defined as Canadian Triage and Acuity Scale [CTAS] level 1, 2, or 3) and less than 4 hours for less severe cases (defined as CTAS level of 4 or 5) [25]. The United Kingdom targets a maximum ED LOS of 4 hours, and Korea has set a goal of 5 to 6 hours for severe cases [24,26].

Numerous studies have identified factors influencing ED LOS. Driesen et al. [7] attributed prolonged ED LOS primarily to external organizational factors such as bed shortages, delays in radiological imaging, and the need for sequential specialist consultations. Pines et al. [27] identified "inappropriate use" as a significant contributor to ED crowding. Lee et al. [12] observed that age older than 65 years, being a medical aid beneficiary, nighttime arrivals, hospitals with more than 1,000 beds, high ED levels, and a metropolitan location were associated with ED LOS exceeding 6 hours.

In our study, elderly patients were significantly more prevalent in the LOS ≥6 hours group, indicating a strong association between age and prolonged ED LOS. This finding aligns with previous research, which suggested that older adults tend to experience prolonged ED stays due to preexisting medical conditions, higher admission needs, and more complex symptom presentations [5]. In contrast, pediatric patients typically experience shorter ED LOS, likely due to the presence of independent pediatric EDs with a more limited range of tests and generally less complex cases.

Geographically, Jeollabuk-do exhibited a longer ED LOS than the other regions, which might be attributable to its small number of level I or II EDs. Patients with medical aid also experienced prolonged ED LOS, reflecting the ED's role as a safety net for this population, who often present with more severe conditions due to lower medication adherence and participation in screening programs [28–30].

Higher level EDs, despite having more resources, are associated with prolonged ED LOS due to the complexity and severity of the cases they handle [12]. Patients referred from outpatient services or transferred from other hospitals also exhibit longer ED LOS because they often require hospital admission or extensive care. Additionally, patients arriving by 119 ambulances tend to have prolonged ED LOS, likely reflecting the severity of their conditions.

Seasonal variations were noted, with spring showing longer ED LOS, potentially due to the annual turnover of medical residents in March, which affects ED efficiency [31]. Admission time also influenced ED LOS, with nighttime admissions associated with longer stays, possibly due to the reduced staffing levels in both EDs and wards [32].

The severity of a patient's condition significantly influences their ED LOS. Higher level EDs, emergency symptoms, and abnormal initial vital signs are associated with prolonged ED LOS. An exception is an abnormal initial systolic blood pressure and SpO₂, which correlate with shorter stays. Diagnoses such as sepsis, COVID-19 infections, and severe illness extend ED LOS. During the COVID-19 pandemic, these delays were exacerbated by quarantine protocols, prolonged diagnostic testing, and boarding delays related to insufficient isolation room capacity [33]. Conditions such as myocardial infarction and cerebral infarction typically result in shorter LOS due to targeted treatment protocols, such as door-to-balloon time and door-to-needle time, and fast-track systems.

ED outcomes are also related to LOS. Patients requiring transfer to another hospital or admission to a general ward often experience longer ED stays, influenced by both throughput components such as diagnostic testing and disposition time and output factors, such as ED boarding. ED boarding significantly contributes to crowding and ambulance diversion because transferring patients requires specialized resources and delays other treatments and exacerbates overall ED LOS.

Sensitivity analyses comparing ED LOS at 12 and 18 hours indicated that the overall trends were similar to those observed at 6 hours (Supplementary Tables 3–6). The primary factor associated with prolonged ED LOS at these extended durations was the wait for transfer to other hospitals and general ward admission. To mitigate ED LOS and crowding, proactive efforts to transfer patients within 24 hours and efficient inpatient bed management are essential.

A key strength of this study is our use of national data from a single country to provide a comprehensive and representative sample of ED visits. This enhances the generalizability and relevance of our findings within Korea. Furthermore, we used a robust framework by incorporating input, throughput, and output factors to identify the determinants of prolonged ED LOS. This holistic approach enabled a thorough examination of the various elements that contribute to extended stays in the ED, facilitating a deeper understanding of the multifactorial nature of ED LOS.

This study also has several limitations. First, as a retrospective observational study, it could be subject to selection bias and missing data. Second, some variables from the NEDIS database were excluded due to incomplete information, potentially affecting the accuracy and generalizability of the results. Third, data from level III EDs were excluded, so the findings mainly apply to higher level EDs. Fourth, because the study is based on Korea's emergency medical system, differences in healthcare systems, patient demographics, and resources could limit the generalizability of our findings to other countries. Fifth, despite adjustments in a multivariate analysis, unmeasured confounding factors such as hospital practices and staffing levels might have influenced ED LOS. Sixth, the study period includes the COVID-19 pandemic, which affected ED operations. Seventh, the outlier threshold was set at 24 hours, leading to the exclusion of 2.26% of patients. In Korea, policy mandates that the proportion of patients staying in the ED for more than 24 hours be less than 5% annually. This exclusion might have introduced selection bias. Last, the optimal duration for ED LOS has not been established. Although we analyzed factors associated with a stay exceeding 6 hours, our results should not be interpreted as suggesting that ED stays must be limited to a specific timeframe.

In summary, this study has elucidated several factors associated with prolonged ED LOS. These findings underscore the multifactorial nature of ED LOS, which is influenced by patient demographics, clinical severity, and systemic healthcare variables. Targeted interventions focusing on older adults, severe illness, and operational inefficiencies, such as hospital transfers, are essential for reducing ED LOS and improving overall emergency care delivery.