The emergency department (ED) is an important public service that provides immediate access and stabilization for patients with emergency conditions [1]. The immediate access nature of the ED has led an increasing number of people to use it as a preferential route to access treatment, causing overcrowding. ED overcrowding, more correctly defined as access block, has become a worldwide phenomenon, causing serious consequences on patient outcomes, staff work and the system, with a general increase in costs [2].
The problem of crowding (or overcrowding) of the ED, as well as that of long patient wait times, occurs domestically and internationally. Several techniques have been adopted to improve the efficiency of the procedures in the ED [3, 4]. In recent years, in particular, there has been an increasing use of data analysis and artificial intelligence to enhance biomedical data and signal analysis, for example as support for diagnosis [5, 6], for the development of simulation models to support the characterization of flows [7,8,9] or for the optimization of processes through the support of appropriate performance indicators [10, 11]. There are several studies that have provided for the implementation of these techniques directly in the ED, for the study of waiting times [12], length of stay [13, 14] and drop-out rates [15, 16]. The ED, with its complex mission, namely, that of providing an adequate, timely and optimal response to patients who present themselves in an unscheduled manner and addressing clinically critical situations by implementing all the necessary life-saving practices [17], needs even more rigorous analysis and efficiency. The main problem is that there is no universal standard definition of overcrowding in emergency rooms because there is no single standard measure of hospital performance [18, 19]. One of the definitions that seems more complete is that provided by the American College of Emergency Physicians (ACEP) Crowding Resources Task Force, according to which overcrowding can be defined “a state in which the identified need for emergency services exceeds available resources in the ED. This situation occurs in hospital EDs when there are more patients than staffed ED treatment beds and when wait times exceed a reasonable period” [20].
ED overcrowding is often measured through the mean occupancy rate or by dividing the number of patients in the emergency department by the number of treatment spaces [21].
In recent years, a large number of scientific studies have addressed this problem, including contributions from different areas of research [22, 23]. In 1990, the United Kingdom became the first country to introduce a few clinical indices [24] based on the quickness in assessing the condition of the patient as a parameter [25]. However, in 2010, Jones and Schimanski [26] demonstrated that the introduction of an ED time target effectively abolished this measure [27], and the associated massive financial investment has not resulted in a consistent improvement in the United Kingdom. Therefore, new indices have been proposed [24] to indicate the quality of care through a wider range of variables that can be monitored: 1) time to initial nursing review, 2) duration of treatment, 3) number of outpatients, and 4) number of patients who leave the department without being seen. A problematic point, however, was that the theoretical indices usually did not reflect the actual conditions in which clinicians found themselves working. Zhou et al. [28] collected subjective and objective emergency department occupancy (EDO) data three times a day (1:00, 9:00, and 17:00) over a period of six months and analysed them using Bland–Altman and Kappa tests. The results showed that the NEDOCS (National Emergency Department Overcrowding Scale) index did not consistently reflect the sense of overcrowding perceived by the staff in the emergency room, calculated by the VAS (Visual Analogue Scale) method. This situation highlights the importance of identifying a complete taxonomy of the ED crowding indices present in the scientific literature.
The scientific literature identifies 16 main indices of ED crowding, including four multidimensional indices (EDWIN [Emergency Department Work Index], Hazard Stairs, READI [Real-time Emergency Analysis of Demand Indicators], and NEDOCS), five input indices (total capacity for first aid, number of patient arrivals in six hours, ambulance transport number, number of patients waiting for medical treatment, and number of patients in the waiting room), three indices of throughput (length of stay in the emergency room [ED LOS], wait time for a first appointment, and time spent in the waiting room), and two indices of output (number of patients in the emergency room and percentage of total beds occupied) [29].
Once the indices were defined, various studies were conducted to validate them. Tekwani et al. [30] conducted a survey based on an interval of eight months on a sample of patients discharged from the emergency room to quantitatively assess the effect of crowding on patient satisfaction using a variety of questions associated with a Likert scale score, a widely used scale for healthcare quality assessment studies [31, 32].
McCarthy et al. [33] compared the emergency room occupancy rate, calculated as the ratio between the total number of hospital stays and the total number of hospital beds in a given period, to measure emergency room crowding, with a validated EDWIN index. Although not extremely accurate, the latter index can be used to quantify crowding and has the advantage of being simpler and more intuitive than the other indices. Several studies were also conducted to assess the correlation between the EDWIN score and the frequency of medical errors [34] or the delayed antibiotics for sepsis [35]. In a study by Todisco [36], after the introduction of six beds into the emergency room, there was a 10.11% reduction in the NEDOCS. There are also several examples of the application of the NEDOCS index to measure ED crowding [19, 37, 38].
Several works have also compared the performance of the EDWIN and NEDOCS indices in evaluating overcrowding. Weiss et al. [39] demonstrated that both indices, and in particular the NEDOCS index, show good accuracy in measuring the overcrowding of an emergency room. Instead, Bernstein et al. [40] showed a strong correlation between the EDWIN index and the staff’s crowding assessment.
This work aims to correlate the EDWIN and NEDOCS indices to verify their validity for evaluating ED crowding at the “A. Cardarelli” Hospital of Naples. The possibility of including a run-time instrument in the information system of the ED is also considered.