Physician performance scores used to predict emergency department admission numbers and excessive admissions burden.

BACKGROUND: Overcrowding in hospitals is associated with a panoply of adverse events. Inappropriate decisions in the emergency department (ED) contribute to overcrowding. The performance of individual physicians as part of the admitting team is a critical factor in determining the overall rate of admissions. While previous attempts to model admission numbers have been based on a range of variables, none have included measures of individual staff performance. We construct reliable objective measures of staff performance and use these, among other factors, to predict the number of daily admissions. Such modelling will enable enhanced workforce planning and timely intervention to reduce inappropriate admissions and overcrowding. METHODS: A database was created of 232 245 ED attendances at Meir Medical Center in central Israel, spanning the years 2016-2021. We use several measures of physician performance together with historic caseload data and other variables to derive statistical models for the prediction of ED arrival and admission numbers. RESULTS: Our models predict arrival numbers with a mean absolute percentage error (MAPE) of 6.85%, and admission numbers with a MAPE of 10.6%, and provide a same-day alert for heavy admissions burden with 75% sensitivity for a false-positive rate of 20%. The inclusion of physician performance measures provides an essential boost to model performance. CONCLUSIONS: Arrival number and admission numbers can be predicted with sufficient fidelity to enable interventions to reduce excess admissions and smooth patient flow. Individual staff performance has a strong effect on admission rates and is a critical variable for the effective modelling of admission numbers.

Reduced Time to Admit Emergency Department Patients to Inpatient Beds Using Outflow Barrier Analysis and Process Improvement.

Objective: Because admitted emergency department (ED) patients waiting for an inpatient bed contribute to dangerous ED crowding, we conducted a patient flow investigation to discover and solve outflow delays. After solution implementation, we measured whether the time admitted ED patients waited to leave the ED was reduced. Methods: In June 2022, a team using Lean Healthcare methodologies identified flow delays and underlying barriers in a Midwest, mid-sized hospital. We calculated barriers' magnitudes of burden by the frequency of involvement in delays. During October-December 2022, solutions targeting barriers were implemented. In October 2023, we tested whether waiting time, defined as daily median time in minutes from admission disposition to departure (ADtoD), declined by conducting independent sample, single-tailed t-test comparing pre- to post-intervention time periods, January 1-September 30, 2022 (273 days) to January 1-September 30, 2023 (273 days). Additionally, we regressed ADtoD onto pre-/post period while controlling for ED volume (total daily admissions and ED daily encounters) and hospital occupancy. A run chart analysis of monthly median ADtoD assessed improvement sustainability. Results: Process mapping revealed that three departments (ED, environmental services [EVS], and transport services) co-produced the outflow of admitted ED patients wherein 18 delays were identified. The EVS-clinical care collaboration failures explained 61% (11/18) of delays. Technology contributed to 78% (14/18) of delays primarily because staff's technology did not display needed information, a condition we coined "digital blindness." Comparing pre- and post-intervention days (3,144 patients admitted pre-intervention and 3,256 patients post), the median minutes a patient waited (ADtoD) significantly decreased (96.4 to 87.1 minutes, P = 0.04), even while daily ED encounter volume significantly increased (110.7 to 117.3 encounters per day, P < 0.001). After controlling in regression for other factors associated with waiting, the intervention reduced ADtoD by 12.7 minutes per patient (standard error 5.10, P = 0.01; 95% confidence interval -22.7, -2.7). We estimate that the intervention translated to ED staff avoiding 689 hours of admitted patient boarding over nine months (ADtoD coefficient [-12.7 minutes] multiplied by post-intervention ED admissions [3,256] and divided by 60). Run chart analysis substantiated the intervention's sustainability over nine months. Conclusion: After systemwide patient flow investigation, solutions resolving digital blindness and environmental services-clinical care collaboration failures significantly reduced ED admitted patient boarding.

Larval crowding enhances dengue virus loads in Aedes aegypti, a relationship that might increase transmission in urban environments.

BACKGROUND: Climate change and urbanization will alter the global distribution of disease vectors, changing the disease burden in yet unpredictable ways. Aedes aegypti is a mosquito responsible for transmitting dengue, Zika, chikungunya, and yellow fever viruses that breeds in containers associated with urban environments. We sought to understand how ambient temperature and larval densities in the immature aquatic phases determine adult life history traits and dengue virus loads post-infection. We predicted that larval crowding and high temperatures would both lead to smaller mosquitoes that might struggle to invest in an immune response and, hence, would exhibit high viral loads. METHODS: We first examined larval densities from urban and rural areas via a meta-analysis. We then used these data to inform a laboratory-based 2x2 design examining the interacting effects of temperature (21 vs. 26°C) and density (0.2 vs. 0.4 larvae/mL) on adult life history and dengue virus loads. RESULTS: We found that urban areas had an ~8-fold increase in larval densities compared to more rural sites. In the lab, we found that crowding had more impact on mosquito traits than temperature. Crowding led to slower development, smaller mosquitoes, less survival, lower fecundity, and higher viral loads, as predicted. The higher temperature led to faster development, reduced fecundity, and lower viral loads. The virus-reducing effect of higher temperature rearing was, however, overwhelmed by the impact of larval crowding when both factors were present. CONCLUSIONS: These data reveal complex interactions between the environmental effects experienced by immature mosquitoes and adult traits. They especially highlight the importance of crowding with respect to adult viral loads. Together, these data suggest that urban environments might enhance dengue virus loads and, therefore, possibly transmission, a concerning result given the increasing rates of urbanization globally.

Hospital overcrowding and care of stroke patients: Irish national audit of stroke.

INTRODUCTION: Hospital overcrowding where patient admissions exceed capacity is associated with worse outcomes in Emergency Department. Developments in emergency stroke care have been associated with improvements in stroke outcome but are dependent on effective, organised care. We examined if overcrowding in the hospital system was associated with negative changes in stroke outcome. METHODS: Data on overcrowding were obtained from the Irish Nurses and Midwives Organisation (INMO) 'Trolley Count' database recording the number of patients cared for on trolleys/chairs in all acute hospitals each midnight. These were compared with quarterly data from the Irish National Audit of Stroke from 2013 to 2021 inclusive. Variables analysed were inpatient mortality rate, thrombolysis rate for ischaemic stroke, median door to needle time and median length of stay. RESULTS: 579449 patient episodes were recorded by Trolley Watch over the period, (Quarterly Median 16719.5, range 3389-27015). Average Quarterly Thrombolysis rate was 11.3% (sd 1.3%) Median Quarterly Inpatient Mortality rate was 11.8% (Range 8.9-14.0%). Median Quarterly Length of stay was 9 days (8-11 days). Median quarterly door to needle was 65 min (45-80 min). Q1 was typically the worst for overcrowding with on average 19777 incidences (sd 4786). This was significantly higher than for Q2 (mean 13540 (sd 4785) p = 0.005 t-test) and for Q3 (mean 14542 (sd 4753) p = 0.03). No significant correlation was found between quarterly Trolley watch episodes and inpatient mortality (r = 0.084, p = 0.63), median length of stay r=-0.15, p = 0.37) or thrombolysis rate (r = 0.089 p = 0.61). There was an unexpected significant negative correlation between trolley watch data and median door to needle time (r=-0.36, p = 0.03). CONCLUSION: Despite increasing hospital overcrowding, stroke services still managed to preserve standard of care. We could find no association between levels of overcrowding and deterioration in selected indices of patient care.

Redirection of low-acuity emergency department patients to nearby medical clinics using an electronic medical support system: effects on emergency department performance indicators.

BACKGROUND: Overcrowded emergency departments (EDs) are associated with higher morbidity and mortality and suboptimal quality-of-care. Most ED flow management strategies focus on early identification and redirection of low-acuity patients to primary care settings. To assess the impact of redirecting low-acuity ED patients to medical clinics using an electronic clinical decision support system on four ED performance indicators. METHODS: We performed a retrospective observational study in the ED of a Canadian tertiary trauma center where a redirection process for low-acuity patients was implemented. The process was based on a clinical decision support system relying on an algorithm based on chief complaint, performed by nurses at triage and not involving physician assessment. All patients visiting the ED from 2013 to 2017 were included. We compared ED performance indicators before and after implementation of the redirection process (June 2015): length-of-triage, time-to-initial-physician-assessment, length-of-stay and rate of patients leaving without being seen. We performed an interrupted time series analysis adjusted for age, gender, time of visit, triage category and overcrowding. RESULTS: Of 242,972 ED attendees over the study period, 9546 (8% of 121,116 post-intervention patients) were redirected to a nearby primary medical clinic. After the redirection process was implemented, length-of-triage increased by 1 min [1;2], time-to-initial assessment decreased by 13 min [-16;-11], length-of-stay for non-redirected patients increased by 29 min [13;44] (p < 0.001), minus 20 min [-42;1] (p = 0.066) for patients assigned to triage 5 category. The rate of patients leaving without being seen decreased by 2% [-3;-2] (p < 0.001). CONCLUSION: Implementing a redirection process for low-acuity ED patients based on a clinical support system was associated with improvements in two of four ED performance indicators.

Deadly yet Preventable? Lessons From South Korea's Halloween Crowd Crush.

Avoidable disasters are both saddening and baffling. In 2022, 159 people, mostly in their 20s, and 30s were crushed to death in Itaewon's narrow alleyway amid South Korea's first pandemic-restrictions-free Halloween celebration. What is particularly sobering about this tragedy is that although many people called police hotlines as crowds became cramped and static, their calls went unheeded for hours. Rather than order independent investigations into the catastrophe (as of January 2024), the President of South Korea at the time focused on superficial issues such as asking the public to refer to the disaster as an "accident" (which it was not, it was an avoidable disaster) and the casualties as "the dead" (who are casualties indeed, instead of victims of a preventable tragedy). In this paper, we examine how officials' complacency about public health and safety dangers, ineffective disaster prevention, and preparedness systems, as well as the government's chronic lack of prioritization of public health and safety may have contributed to the disaster. Furthermore, we discuss the importance of creating integrated public health and safety protection systems to prevent similar tragedies from happening.

Crowd Counting Using Meta-Test-Time Adaptation.

Machine learning algorithms are commonly used for quickly and efficiently counting people from a crowd. Test-time adaptation methods for crowd counting adjust model parameters and employ additional data augmentation to better adapt the model to the specific conditions encountered during testing. The majority of current studies concentrate on unsupervised domain adaptation. These approaches commonly perform hundreds of epochs of training iterations, requiring a sizable number of unannotated data of every new target domain apart from annotated data of the source domain. Unlike these methods, we propose a meta-test-time adaptive crowd counting approach called CrowdTTA, which integrates the concept of test-time adaptation into the meta-learning framework and makes it easier for the counting model to adapt to the unknown test distributions. To facilitate the reliable supervision signal at the pixel level, we introduce uncertainty by inserting the dropout layer into the counting model. The uncertainty is then used to generate valuable pseudo labels, serving as effective supervisory signals for adapting the model. In the context of meta-learning, one image can be regarded as one task for crowd counting. In each iteration, our approach is a dual-level optimization process. In the inner update, we employ a self-supervised consistency loss function to optimize the model so as to simulate the parameters update process that occurs during the test phase. In the outer update, we authentically update the parameters based on the image with ground truth, improving the model's performance and making the pseudo labels more accurate in the next iteration. At test time, the input image is used for adapting the model before testing the image. In comparison to various supervised learning and domain adaptation methods, our results via extensive experiments on diverse datasets showcase the general adaptive capability of our approach across datasets with varying crowd densities and scales.

Improving emergency department flow by introducing four interventions simultaneously. A quality improvement project.

BACKGROUND: Emergency department (ED) crowding is a widespread issue with adverse effects on patient care and outcomes. LOCAL PROBLEM: ED crowding exacerbates wait times and compromises patient care, prompting opportunities for internal process improvement. METHOD: Over one week, the ED flow project team implemented four interventions, including an additional triage station, to optimize patient flow. We compared triage times, length of stay, crowding levels, and patient experiences with two control periods. RESULTS: During peak hours, waiting times to triage decreased significantly with a median of 20 min (IQR 15-30) in the project week and 26 min (IQR 18-37) in the control weeks. Self-referrals decreased, while general practitioner referrals remained unchanged. Individual patient length of stay was unaffected, but crowding reduced notably during the project week. We found no difference in patient experiences between the periods. CONCLUSION: The interventions contributed to reduced crowding and improved patient flow. The dedication of the ED flow project team and the ED nurses was crucial to these outcomes. An additional triage station during peak hours in the ED was established as a structural change.

Identification of avoidable patients at triage in a Paediatric Emergency Department: a decision support system using predictive analytics.

BACKGROUND: Crowding has been a longstanding issue in emergency departments. To address this, a fast-track system for avoidable patients is being implemented in the Paediatric Emergency Department where our study is conducted. Our goal is to develop an optimized Decision Support System that helps in directing patients to this fast track. We evaluated various Machine Learning models, focusing on a balance between complexity, predictive performance, and interpretability. METHODS: This is a retrospective study considering all visits to a university-affiliated metropolitan hospital's PED between 2014 and 2019. Using information available at the time of triage, we trained several models to predict whether a visit is avoidable and should be directed to a fast-track area. RESULTS: A total of 507,708 visits to the PED were used in the training and testing of the models. Regarding the outcome, 41.6% of the visits were considered avoidable. Except for the classification made by triage rules, i.e. considering levels 1,2, and 3 as non-avoidable and 4 and 5 as avoidable, all models had similar results in model's evaluation metrics, e.g. Area Under the Curve ranging from 74% to 80%. CONCLUSIONS: Regarding predictive performance, the pruned decision tree had evaluation metrics results that were comparable to the other ML models. Furthermore, it offers a low complexity and easy to implement solution. When considering interpretability, a paramount requisite in healthcare since it relates to the trustworthiness and transparency of the system, the pruned decision tree excels. Overall, this paper contributes to the growing body of research on the use of machine learning in healthcare. It highlights practical benefits for patients and healthcare systems of the use ML-based DSS in emergency medicine. Moreover, the obtained results can potentially help to design patients' flow management strategies in PED settings, which has been sought as a solution for addressing the long-standing problem of overcrowding.

Beating the Clock: A Prediction Model for Timely Care in Emergency Departments.

We developed and validated a statistical prediction model using 2.5 electronic health records from 24 German emergency departments (EDs) to estimate treatment timeliness at triage. The model's moderate fit and reliance on interoperable, routine data suggest its potential for implementation in ED crowding management.

Exploring Hospital Overcrowding with an Explainable Time-to-Event Machine Learning Approach.

Emergency department (ED) overcrowding is a complex problem that is intricately linked with the operations of other hospital departments. Leveraging ED real-world production data provides a unique opportunity to comprehend this multifaceted problem holistically. This paper introduces a novel approach to analyse healthcare production data, treating the length of stay of patients, and the follow up decision regarding discharge or admission to the hospital as a time-to-event analysis problem. Our methodology employs traditional survival estimators and machine learning models, and Shapley additive explanations values to interpret the model outcomes. The most relevant features influencing length of stay were whether the patient received a scan at the ED, emergency room urgent visit, age, triage level, and the medical alarm unit category. The clinical insights derived from the explanation of the models holds promise for increase understanding of the overcrowding from the data. Our work demonstrates that a time-to-event approach to the over- crowding serves as a valuable initial to uncover crucial insights for further investigation and policy design.

Abated crowding by fast-tracking the Throughput component of the ED for patients in no need of hospitalization with competency managed personnel.

BACKGROUND: Emergency department (ED) crowding is a major patient safety concern and has a negative impact on healthcare systems and healthcare providers. We hypothesized that it would be feasible to control crowding by employing a multifaceted approach consisting of systematically fast-tracking patients who are mostly not in need of a hospital stay as assessed by an initial nurse and treated by decision competent physicians. METHODS: Data from 120,901 patients registered in a secondary care ED from the 4tth quarter of 2021 to the 1st quarter of 2024 was drawn from the electronic health record's data warehouse using the SAP Web Intelligence tool and processed in the Python programming language. Crowding was compared before and after ED transformation from a uniform department into a high flow (α) and a low flow (ß) section with patient placement in gurneys/chairs or beds, respectively. Patients putatively not in need of hospitalization were identified by nurse, placed in in the α setting and assessed and treated by decision competent physicians. Incidence of crowding, number of patients admitted per day and readmittances within 72 h following ED admission before and after changes were determined. Values are number of patients, mean ± SEM and mean differences with 95% CIs. Statistical significance was ascertained using Student's two tailed t-test for unpaired values. RESULTS: Before and after ED changes crowding of 130% amounted to 123.8 h and 19.3 h in the latter. This is a difference of -104.6 ± 23.9 h with a 95% CI of -159.9 to -49.3, Δ% -84 (p = 0.002). There was the same amount of patients / day amounting to 135.8 and 133.5 patients / day Δ% = -1.7 patients 95% CI -6.3 to 1.6 (p = 0.21). There was no change in readmittances within 72 h before and after changes amounting to 9.0% versus 9.5%, Δ% = 0.5, 95%, CI -0.007 to 1.0 (p > 0.052). CONCLUSION: It appears feasible to abate crowding with unchanged patient admission and without an increase in readmittances by fast-track assessment and treatment of patients who are not in need of hospitalization.

Daily crowding stress has limited, yet detectable effects on skin and head kidney gene expression in surgically tagged atlantic salmon (Salmo salar).

To ensure welfare-friendly and effective internal tagging, the tagging process should not cause a long-term burden on individuals given that tagged fish serve as representatives for the entire population in telemetry applications. To some extent, stress is inevitable within regular aquaculture practices, and thus, the consequences of long-term stress should be described in terms of their effects on internal tagging. In fish, stressors activate the Hypothalamus-Pituitary-Interrenal (HPI) and Brain-Sympathetic-Chromaffin Cell (BSC) axes, leading to neuroimmunoendocrine communication and paracrine interactions among stress hormones. The interrelation between wound healing and stress is complex, owing to their shared components, pathways, and energy demands. This study assessed 14 genes (mmp9, mmp13, il-2, il-4, il-8a, il-10, il-12, il-17d, il-1b, tnfa, ifng, leg-3, igm, and crh) in the skin (1.5 cm from the wound) and head kidney over eight weeks. These genes, associated with cell signaling in immunity, wound healing, and stress, have previously been identified as influenced and regulated by these processes. Half of a group of Atlantic salmon (n = 90) with surgically implanted dummy smart-tags were exposed to daily crowding stress. The goal was to investigate how this gene panel responds to a wound alone and then to the combined effects of wounding and daily crowding stress. Our observations indicate that chronic stress impacts inflammation and impedes wound healing, as seen through the expression of matrix metalloproteinases genes in the skin but not in the head kidney. This difference is likely due to the ongoing internal wound repair, in contrast to the externally healed wound incision. Cytokine expression, when significant in the skin, was mainly downregulated in both treatments compared to control values, particularly in the study's first half. Conversely, the head kidney showed initial cytokine downregulation followed by upregulation. Across all weeks observed and combining both tissues, the significantly expressed gene differences were 12 % between the Wound and Stress+ groups, 28 % between Wound and Control, and 25 % between Stress+ and Control. Despite significant fluctuations in cytokines, sustained variations across multiple weeks are only evident in a few select genes. Furthermore, Stress+ individuals demonstrated the most cytokine correlations within the head kidney, which may suggest that chronic stress affects cytokine expression. This investigation unveils that the presence of stress and prolonged activation of the HPI axis in an eight weeklong study has limited yet detectable effects on the selected gene expression within immunity, wound healing, and stress, with notable tissue-specific differences.

Prison overcrowding and ill-treatment: sentence reduction as a reparation measure. A view from Latin America and Europe.

INTRODUCTION: In recent years mass protest movements have taken to the streets in many countries across the world. Despite strong international and domestic legal protections for the right to freedom of peaceful assembly and other fundamental human rights, entire assemblies are frequently labelled violent and less lethal weapons are used to disperse them. METHODS: This article examines the weapons often used by police against public assemblies. Focusing on striking weapons (batons), chemical irri-tants, kinetic impact projectiles and stun grenades, the article uses examples from various countries to illustrate how these weapons are being used and the associated human rights and health impacts. Re-sults: Worrying trends identified include the use of dangerous or untested equipment, such as thermal foggers to deploy chemical irritants; the use of inherently abusive weapons, such as whips or sjam-boks; and the increasing use of certain types of munitions, specifically indiscriminate kinetic impact projectiles. DISCUSSION: The article seeks to support medical and legal professionals becoming more familiar with the weapons being used in the countries they practice in, the effects of those weapons, and clinical aspects in the presentation and care of those exposed.

Behavioral transition of a fish school in a crowded environment.

In open water, social fish gather to form schools, in which fish generally align with each other. In this work, we study how this social behavior evolves when perturbed by artificial obstacles. We measure the behavior of a group of zebrafish in the presence of a periodic array of pillars. When the pillar density is low, the fish regroup with a typical interdistance and a well-polarized state with parallel orientations, similarly to their behavior in open-water conditions. Above a critical density of pillars, their social interactions, which are mostly based on vision, are screened and the fish spread randomly through the aquarium, orienting themselves along the free axes of the pillar lattice. The abrupt transition from natural to artificial orientation happens when the pillar interdistance is comparable to the social distance of the fish, i.e., their most probable interdistance. We develop a stochastic model of the relative orientation between fish pairs, taking into account alignment, antialignment, and tumbling, from a distribution biased by the environment. This model provides a good description of the experimental probability distribution of the relative orientation between the fish and captures the behavioral transition. Using the model to fit the experimental data provides qualitative information on the evolution of cognitive parameters, such as the alignment or the tumbling rates, as the pillar density increases. At high pillar density, we find that the artificial environment imposes its geometrical constraints to the fish school, drastically increasing the tumbling rate.

A prospective randomised controlled trial investigating household SARS-CoV-2 transmission in a densely populated community in Cape Town, South Africa - the transmission of COVID-19 in crowded environments (TRACE) study.

BACKGROUND: South Africa's first SARS-CoV-2 case was identified 5th March 2020 and national lockdown followed March 26th. Households are an important location for secondary SARS-CoV-2 infection. Physical distancing and sanitation - infection mitigation recommended by the World Health Organization (WHO) at the time - are difficult to implement in limited-resource settings because of overcrowded living conditions. METHODS: This study (ClinicalTrials.gov NCT05119348) was conducted from August 2020 to September 2021 in two densely populated, low socioeconomic Cape Town community sub-districts. New COVID-19 index cases (ICs) identified at public clinics were randomised to an infection mitigation intervention (STOPCOV) delivered by lay community health workers (CHWs) or standard of care group. STOPCOV mitigation measures included one initial household assessment conducted by a CHW in which face masks, sanitiser, bleach and written information on managing and preventing spread were provided. This was followed by regular telephonic follow-up from CHWs. SARS-CoV-2 PCR and IgM/IgG serology was performed at baseline, weeks 1, 2, 3 and 4 of follow-up. RESULTS: The study randomised 81 ICs with 245 HHCs. At baseline, no HHCs in the control and 7 (5%) in the intervention group had prevalent SARS-CoV-2. The secondary infection rate (SIR) based on SARS-CoV-2 PCR testing was 1.9% (n = 2) in control and 2.9% (n = 4) in intervention HHCs (p = 0.598). At baseline, SARS-CoV-2 antibodies were present in 15% (16/108) of control and 38% (52/137) of intervention participants. At study end incidence was 8.3% (9/108) and 8.03% (11/137) in the intervention and control groups respectively. Antibodies were present in 23% (25/108) of control HHCs over the course of the study vs. 46% (63/137) in the intervention arm. CHWs made twelve clinic and 47 food parcel referrals for individuals in intervention households in need. DISCUSSION: Participants had significant exposure to SARS-CoV-2 infections prior to the study. In this setting, household transmission mitigation was ineffective. However, CHWs may have facilitated other important healthcare and social referrals.

Non-emergency department (ED) interventions to reduce ED utilization: a scoping review.

BACKGROUND: Emergency department (ED) crowding is a global burden. Interventions to reduce ED utilization have been widely discussed in the literature, but previous reviews have mainly focused on specific interventions or patient groups within the EDs. The purpose of this scoping review was to identify, summarize, and categorize the various types of non-ED-based interventions designed to reduce unnecessary visits to EDs. METHODS: This scoping review followed the JBI Manual for Evidence Synthesis and the PRISMA-SCR checklist. A comprehensive structured literature search was performed in the databases MEDLINE and Embase from 2008 to March 2024. The inclusion criteria covered studies reporting on interventions outside the ED that aimed to reduce ED visits. Two reviewers independently screened the records and categorized the included articles by intervention type, location, and population. RESULTS: Among the 15,324 screened records, we included 210 studies, comprising 183 intervention studies and 27 systematic reviews. In the primary studies, care coordination/case management or other care programs were the most commonly examined out of 15 different intervention categories. The majority of interventions took place in clinics or medical centers, in patients' homes, followed by hospitals and primary care settings - and targeted patients with specific medical conditions. CONCLUSION: A large number of studies have been published investigating interventions to mitigate the influx of patients to EDs. Many of these targeted patients with specific medical conditions, frequent users and high-risk patients. Further research is needed to address other high prevalent groups in the ED - including older adults and mental health patients (who are ill but may not need the ED). There is also room for further research on new interventions to reduce ED utilization in low-acuity patients and in the general patient population.

Unraveling the causes of the Seoul Halloween crowd-crush disaster.

As the world steadily recovers from the COVID-19 pandemic, managing large gatherings becomes a critical concern for ensuring crowd safety. The crowd-crush disaster in Seoul in 2022 highlights the need for effective predictive crowd management techniques. In this study, an empirical analysis of this incident is conducted using data from various sources, and model-based simulations are created to replicate hazardous crowd conditions in high-risk areas. In the empirical analysis, mobile device data indicates a significant increase in population above normal levels in the disaster area just hours before the incident occurred. In the simulations, a hydrodynamic model is employed to simulate a bidirectional collision, which quantitatively demonstrates that the average density during the crush reached 7.57 ped/m2 (with a maximum of (9.95)ped/m2). Additionally, the average crowd pressure peaked at 1,063 N/m (with a maximum of 1,961 N/m), and the maximum velocity entropy was 10.99. Based on these findings, it can be concluded that the primary causes of the disaster were the substantial population, bidirectional collision, and escalating panic. The results of controlled simulations under various management strategies are then presented. By implementing effective crowd management techniques, crowd safety can be enhanced through quantitative comparisons of these key indicators.