Association between Initial Fibrinogen Levels and the Need for Massive Transfusion in Emergency Department Patients with Primary Postpartum Hemorrhage: A Retrospective Study from a Single Center in Korea.

BACKGROUND: This study aimed to evaluate the association between initial fibrinogen levels and massive transfusion (MT) in emergency department (ED) patients with primary postpartum hemorrhage (PPH). METHODS: This retrospective study was conducted in the ED of a university-affiliated, tertiary referral center from January 2004 to August 2023. Patients were divided into two groups: the MT group, which included those who received a transfusion of 10 or more units of packed red blood cells within the first 24 h, and the Non-MT group. RESULTS: Out of the 364 patients included in the study, 97 (26.6%) required MT. Fibrinogen, shock index, and lactate were independently associated with MT (odds ratio [OR] 0.987; 95% confidence interval [CI] 0.983-0.991; p < 0.001, OR 7.277; 95% CI 1.856-28.535; p = 0.004, and OR 1.261; 95% CI 1.021-1.557; p = 0.031, respectively). The area under the receiver operating characteristic curve for fibrinogen, shock index, and lactate in predicting MT was 0.871 (95% CI 0.832-0.904; p < 0.001), 0.821 (95% CI 0.778-0.859; p < 0.001), and 0.784 (95% CI 0.738-0.825; p < 0.001), respectively. When the cutoff value of fibrinogen was 400 mg/dL, both the sensitivity and negative predictive values for predicting MT were 100.0%. When the cutoff value of fibrinogen was 100 mg/dL, the specificity and positive predictive values were 91.8% and 70.7%, respectively. CONCLUSION: The initial fibrinogen levels were independently associated with the need for MT in ED patients with primary PPH.

Comparison of Modified Shock Index and Shock Index for Predicting Massive Transfusion in Women with Primary Postpartum Hemorrhage: A Retrospective Study.

BACKGROUND The modified shock index (MSI) is calculated as the ratio of heart rate (HR) to mean arterial pressure (MAP) and has been used to predict the need for massive transfusion (MT) in trauma patients. This retrospective study from a single center aimed to compare the MSI with the traditional shock index (SI) to predict the need for MT in 612 women diagnosed with primary postpartum hemorrhage (PPH) at the Emergency Department (ED) between January 2004 and August 2023. MATERIAL AND METHODS The patients were divided into the MT group and the non-MT group. The predictive power of MSI and SI was compared using the areas under the receiver operating characteristic curve (AUC). The sensitivity, specificity, positive predictive value (PPV), and negative predictive value were calculated. RESULTS Out of 612 patients, 105 (17.2%) required MT. The MT group had higher median values than the non-MT group for MSI (1.58 vs 1.07, P<0.001) and SI (1.22 vs 0.80, P<0.001). The AUC for MSI, with a value of 0.811 (95% confidence interval [CI], 0.778-0.841), did not demonstrate a significant difference compared to the AUC for SI, which was 0.829 (95% CI, 0.797-0.858) (P=0.066). The optimal cutoff values for MSI and SI were 1.34 and 1.07, respectively. The specificity and PPV for MT were 77.1% and 40.2% for MSI, and 83.2% and 45.9% for SI. CONCLUSIONS Both MSI and SI were effective in predicting MT in patients with primary PPH. However, MSI did not demonstrate superior performance to SI.

Prediction of hospitalization and waiting time within 24 hours of emergency department patients with unstructured text data.

Overcrowding of emergency departments is a global concern, leading to numerous negative consequences. This study aimed to develop a useful and inexpensive tool derived from electronic medical records that supports clinical decision-making and can be easily utilized by emergency department physicians. We presented machine learning models that predicted the likelihood of hospitalizations within 24 hours and estimated waiting times. Moreover, we revealed the enhanced performance of these machine learning models compared to existing models by incorporating unstructured text data. Among several evaluated models, the extreme gradient boosting model that incorporated text data yielded the best performance. This model achieved an area under the receiver operating characteristic curve score of 0.922 and an area under the precision-recall curve score of 0.687. The mean absolute error revealed a difference of approximately 3 hours. Using this model, we classified the probability of patients not being admitted within 24 hours as Low, Medium, or High and identified important variables influencing this classification through explainable artificial intelligence. The model results are readily displayed on an electronic dashboard to support the decision-making of emergency department physicians and alleviate overcrowding, thereby resulting in socioeconomic benefits for medical facilities.

Modification of HEART Pathway for Patients With Chest Pain: A Korean Perspective.

BACKGROUND AND OBJECTIVES: The History, Electrocardiography, Age, Risk factors, and Troponin (HEART) pathway was developed to identify patients at low risk of a major adverse cardiac event (MACE) among patients presenting with chest pain to the emergency department. METHODS: We modified the HEART pathway by replacing the Korean cut-off of 25 kg/m² with the conventional threshold of 30 kg/m² in the definition of obesity among risk factors. The primary outcome was a MACE within 30 days, which included acute myocardial infarction, primary coronary intervention, coronary artery bypass grafting, and all-cause death. RESULTS: Of the 1,304 patients prospectively enrolled, MACE occurred in 320 (24.5%). The modified HEART pathway identified 37.3% of patients as low-risk compared with 38.3% using the HEART pathway. Of the 500 patients classified as low-risk with HEART pathway, 8 (1.6%) experienced MACE, and of the 486 low-risk patients with modified HEART pathway, 4 (0.8%) experienced MACE. The modified HEART pathway had a sensitivity of 98.8%, a negative predictive value (NPV) of 99.2%, a specificity of 49.0%, and a positive predictive value (PPV) of 38.6%, compared with the original HEART pathway, with a sensitivity of 97.5%, a NPV of 98.4%, a specificity of 50.0%, and a PPV of 38.8%. CONCLUSIONS: When applied to Korean population, modified HEART pathway could identify patients safe for early discharge more accurately by using body mass index cut-off levels suggested for Koreans.

An Experimental Study on Defect Detection of Anchor Bolts Using Non-Destructive Testing Techniques.

Anchor bolts are often used for fixing information boards, supports, and soundproof walls in various facilities. Corrosion of anchor bolts and fatigue cracks occur frequently due to the various external environments, and visual inspection and hammering inspection are mainly used. In visual inspection, it is difficult to confirm corrosion or fatigue cracks of anchor bolts in the area where foundations, nuts, and base plates are installed. Additionally, the hammering inspection is easily affected by the surrounding environment and the subjective reaction of the tester. Therefore, it is necessary to develop a method that can easily and accurately detect defects such as cracks and corrosion occurring in anchor bolts installed in road facilities using non-destructive testing techniques. In this paper, the possibility and reliability of anchor bolt defects such as corrosion and cracks were experimentally verified by applying ultrasonic inspection among non-destructive inspection techniques for anchor bolt maintenance.

Incidence and short-term outcomes of central line-related bloodstream infection in patients admitted to the emergency department: a single-center retrospective study.

Central line-related bloodstream infection (CRBSI) is a common complication during hospital admissions; however, there is insufficient data regarding CRBSI in the emergency department. Therefore, we evaluated the incidence and clinical impact of CRBSI using a single-center retrospective study to analyze medical data of 2189 adult patients (median age: 65 years, 58.8% males) who underwent central line insertion in ED from 2013 to 2015. CRBSI was defined if the same pathogens were identified at peripheral and catheter tips or the differential time to positivity was > 2 h. CRBSI-related in-hospital mortality and risk factors were evaluated. CRBSI occurred in 80 patients (3.7%), of which 51 survived and 29 died; those with CRBSI had higher incidence of subclavian vein insertion and retry rates. Staphylococcus epidermidis was the most common pathogen, followed by Staphylococcus aureus, Enterococcus faecium, and Escherichia coli. Using multivariate analysis, we found that CRBSI development was an independent risk factor for in-hospital mortality (adjusted odds ratio: 1.93, 95% confidence intervals: 1.19-3.14, p < 0.01). Our findings suggest that CRBSI after central line insertion in the emergency department is common and associated with poor outcomes. Infection prevention and management measures to reduce CRBSI incidence are essential to improve clinical outcomes.

An interactive and realistic phantom for cricothyroidotomy simulation of a patient with obesity through a reusable design using 3D-printing and Arduino.

BACKGROUND AND OBJECTIVES: Proper airway management during emergencies can prevent serious complications. However, cricothyroidotomy is challenging in patients with obesity. Since this technique is not performed frequently but at a critical time, the opportunity for trainees is rare. Simulators for these procedures are also lacking. Therefore, we proposed a realistic and interactive cricothyroidotomy simulator. METHODS: All anatomical structures were modeled based on computed tomography images of a patient with obesity. To mimic the feeling of incision during cricothyroidotomy, the incision site was modeled to distinguish between the skin and fat. To reinforce the educational purpose, capacitive touch sensors were attached to the artery, vein, and thyroid to generate audio feedback. The tensile strength of the silicone-cast skin was measured to verify the similarity of the mechanical properties between humans and our model. The fabrication and assembly accuracies of the phantom between the Standard Tessellation Language and the fabricated model were evaluated. Audio feedback through sensing the anatomy parts and utilization was evaluated. RESULTS: The body, skull, clavicle, artery, vein, and thyroid were fabricated using fused deposition modeling (FDM) with polylactic acid. A skin mold was fabricated using FDM with thermoplastic polyurethane. A fat mold was fabricated using stereolithography apparatus (SLA) with a clear resin. The airway and tongue were fabricated using SLA with an elastic resin. The tensile strength of the skin using silicone with and without polyester mesh was 2.63 ± 0.68 and 2.46 ± 0.21 MPa. The measurement errors for fabricating and assembling parts of the phantom between the STL and the fabricated models were -0.08 ± 0.19 mm and 0.13 ± 0.64 mm. The measurement errors internal anatomy embodied surfaces in fat part were 0.41 ± 0.89 mm. Audio feedback was generated 100% in all the areas tested. The realism, understanding of clinical skills, and intention to retrain were 7.1, 8.8, and 8.3 average points. CONCLUSIONS: Our simulator can provide a realistic simulation experience for trainees through a realistic feeling of incision and audio feedback, which can be used for actual clinical education.

Performance analysis considering endpoints for three accelerated diagnostic protocols for chest pain.

INTRODUCTION: The modified accelerated diagnostic protocol (ADP) to assess patients with chest pain symptoms using troponin as the only biomarker (mADAPT), the History, ECG, Age, Risk factors, and Troponin (HEART) pathway, and the Emergency Department Assessment of Chest Pain Rule (EDACS)-ADP, are the three most well-known ADPs for patients with chest pain. These ADPs define major adverse cardiac event (MACE) as components of acute myocardial infarction, revascularization, and death; unstable angina is not included as an endpoint. METHODS: We performed a single-center prospective observational study comparing the performance of these 3 ADPs for patients with 30-day MACE with and without unstable angina. We hypothesized that these ADPs will have high sensitivities for MACE without unstable angina, a definition used for score derivation studies. However, when unstable angina is included in the MACE, their performances would be lower than the acceptable rate of >99% sensitivity. RESULTS: A total of 1,214 patients were included in the analysis. When unstable angina was not included in the endpoint, sensitivities for MACE were 99.1% (95% confidence interval [CI]: 96.7-99.9%), 99.5% (95% CI: 97.4-100%), and 100% (95% CI: 98.3-100%) for mADAPT, EDACS-ADP, and HEART pathway, respectively. The HEART pathway had the highest proportion of patients classified as low risk (39.2%, 95% CI: 35.8-42.9%), followed by EDACS-ADP (31.3%, 95% CI: 28.2-34.6%) and mADAPT (29.3%, 95% CI: 26.4-32.5%). However, when unstable angina was included in the MACE, sensitivities were 96.6% (95% CI: 94.4-98.1%) for mADAPT, 97.3% (95% CI: 95.3-98.6%) for EDACS-ADP, and 97.3% (95% CI: 95.3-98.6%) for the HEART pathway, respectively. There were 15 false-negative cases with mADAPT, and 12 false-negative cases each for EDACS-ADP and HEART pathway. CONCLUSION: All three ADPs-mADAPT, EDACS-ADP, and HEART pathway-were similarly accurate in their discriminatory performance for the risk stratification of ED patients presenting with possible ACS when unstable angina was not included in the endpoint. The HEART pathway showed the best combination of sensitivity and proportion of patients that can be classified as safe for early discharge. However, when unstable angina was added to the endpoint, all three ADPs did not show appropriate safety levels and their performances were lower than the acceptable risk of MACE.

Analyses of Pore-Size-Dependent Ionic Transport in Nanopores in the Presence of Concentration and Temperature Gradients.

Mass transport through nanopores occurs in various natural systems, including the human body. For example, ion transport across nerve cell membranes plays a significant role in neural signal transmission, which can be significantly affected by the electrolyte and temperature conditions. To better understand and control the underlying nanoscopic transport, it is necessary to develop multiphysical transport models as well as validate them using enhanced experimental methods for facile nanopore fabrication and precise nanoscale transport characterization. Here, we report a nanopore-integrated microfluidic platform to characterize ion transport in the presence of electrolyte and temperature gradients; we employ our previous self-assembled particle membrane (SAPM)-integrated microfluidic platform to produce various nanopores with different pore sizes. Subsequently, we quantify pore-size-dependent ionic transport by measuring the short circuit current (SCC) and open circuit voltage (OCV) across various nanopores by manipulating the electrolyte and temperature gradients. We establish three simple theoretical models that heavily depend on pore size, electrolyte concentration, and temperature and subsequently validate them with the experimental results. Finally, we anticipate that the results of this study would help clarify ion transport phenomena at low-temperature conditions, not only providing a fundamental understanding but also enabling practical applications of cryo-anesthesia in the near future.

Characteristics of Patients Who Visited Emergency Department: A Nationwide Population-Based Study in South Korea (2016-2018).

The utilization of the emergency department (ED) has been continuously increasing and has become a burden for ED resources. The aim of this study was to describe the characteristics, outcomes, common diagnoses, and disease classifications of patients who were referred to the ED. This nationwide epidemiologic study examined the data from adult patients (>18 years) who visited EDs from 1 January 2016 to 31 December 2018. Most EDs in Korea provide data from ED patients to the National Emergency Medical Center (NEMC). The disposition of ED patients was classified as discharge, admission, death, and re-transfer. From 2016 to 2018, the proportion of referred patients out of the total ED visits increased from 7.3% to 7.8%. The referred patients were older (61.1 vs. 50.5 years), had worse vital signs, longer ED lengths of stay (409.1 vs. 153.3 min), and higher admission (62.3 vs. 16.9%) and re-transfer rates (4.4 vs. 1.9%) than the direct-visit patients. Among the referred patients in the 3 years, 62.3% were hospitalized, and the most common disease classification was "disease of the digestive system" (19.8%). The most common diagnosis was pneumonia (6.0%), followed by urinary tract infection, gastrointestinal bleeding, and hepatobiliary infection. The number of patients referred to EDs is increasing, and more than 60% of referred patients are hospitalized. Detailed characteristics of these patients will be helpful for improving ED management and the distribution of medical resources.

Gate-controlled gas sensor utilizing 1D-2D hybrid nanowires network.

Novel gas sensors that work at room temperature are attracting attention due to their low energy consumption and stability in the presence of toxic gases. However, the development of sensing characteristics at room temperature is still a primary challenge. Diverse reaction pathways and low adsorption energy for gas molecules are required to fabricate a gas sensor that works at room temperature with high sensitivity, selectivity, and efficiency. Therefore, we enhanced the gas sensing performance at room temperature by constructing hybridized nanostructure of 1D-2D hybrid of SnSe2 layers and SnO2 nanowire networks and by controlling the back-gate bias (Vg = 1.5 V). The response time was dramatically reduced by lowering the energy barrier for the adsorption on the reactive sites, which are controlled by the back gate. Consequently, we believe that this research could contribute to improving the performance of gas sensors that work at room temperature.

Late Awakening Is Common in Settings Without Withdrawal of Life-Sustaining Therapy in Out-of-Hospital Cardiac Arrest Survivors Who Undergo Targeted Temperature Management.

OBJECTIVES: We investigated awakening time and characteristics of awakening compared nonawakening and factors contributing to poor neurologic outcomes in out-of-hospital cardiac arrest survivors in no withdrawal of life-sustaining therapy settings. DESIGN: Retrospective analysis of the Korean Hypothermia Network Pro registry. SETTING: Multicenter ICU. PATIENTS: Adult (≥ 18 yr) comatose out-of-hospital cardiac arrest survivors who underwent targeted temperature management at 33-36°C between October 2015 and December 2018. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We measured the time from the end of rewarming to awakening, defined as a total Glasgow Coma Scale score greater than or equal to 9 or Glasgow Coma Scale motor score equals to 6. The primary outcome was awakening time. The secondary outcome was 6-month neurologic outcomes (poor outcome: Cerebral Performance Category 3-5). Among 1,145 out-of-hospital cardiac arrest survivors, 477 patients (41.7%) regained consciousness 30 hours (6-71 hr) later, and 116 patients (24.3%) awakened late (72 hr after the end of rewarming). Young age, witnessed arrest, shockable rhythm, cardiac etiology, shorter time to return of spontaneous circulation, lower serum lactate level, absence of seizures, and multisedative requirement were associated with awakening. Of the 477 who woke up, 74 (15.5%) had poor neurologic outcomes. Older age, liver cirrhosis, nonshockable rhythm, noncardiac etiology, a higher Sequential Organ Failure Assessment score, and higher serum lactate levels were associated with poor neurologic outcomes. Late awakeners were more common in the poor than in the good neurologic outcome group (38/74 [51.4%] vs 78/403 [19.4%]; p < 0.001). The awakening time (odds ratio, 1.005; 95% CIs, 1.003-1.008) and late awakening (odds ratio, 3.194; 95% CIs, 1.776-5.746) were independently associated with poor neurologic outcomes. CONCLUSIONS: Late awakening after out-of-hospital cardiac arrest was common in no withdrawal of life-sustaining therapy settings and the probability of awakening decreased over time.

Subcutaneous fat area at the upper thigh level is a useful prognostic marker in the elderly with femur fracture.

BACKGROUND: The aim of this study was to evaluate prognostic value of body tissue composition at the upper thigh level for 1 year mortality in elderly patients with proximal femur fracture. METHODS: This retrospective cohort study included consecutive elderly (aged ≥65) patients diagnosed with proximal femur fracture based on the findings of pelvic bone computed tomography (CT) performed at the emergency department of a tertiary care hospital and treated with surgery between 2010 and 2017. The cross-sectional area of subcutaneous fat and skeletal muscle at the upper thigh level was measured using CT. Adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for 1 year mortality were estimated using a Cox proportional hazards model. Survival based on the SFA quartiles was assessed using nonparametric Kaplan-Meier survival analysis and compared used log-rank tests. RESULTS: Among 876 elderly patients included in this study, the median age was 79.0 years, and 646 (73.7%) patients were female. A total of 93 (10.6%) died within 1 year after admission to the emergency department. Survivors had a significantly higher median subcutaneous fat area (SFA) than non-survivors (170.2 vs. 133.0 cm2 , P < 0.001), but no significant difference was observed between the skeletal muscle area (median, 156.7 vs. 160.3 cm2 , P = 0.504) and muscle density (median, 19.0 vs. 19.1 HU, P = 0.861) of both groups. After adjustment of other clinical characteristics and body compositions, the multivariate Cox proportional hazard analysis showed that SFA (adjusted HR, 0.987; 95% CI, 0.982-0.992; P < 0.001) was independently associated with 1 year mortality. With 384 deaths during 51 322 person-months of follow-up, the median estimated survival duration of all the patients was 92.8 months (95% CI, 80.8-104.7 months). The patients with SFA in the third (165.6-195.0 cm2 ) and fourth (>195.0 cm2 ) quartiles showed significantly longer survival duration than those with SFA in the first (<131.4 cm2 ; median survival time, 51.3 months) and second (131.4-165.5 cm2 ; median survival time, 88.7 months) quartiles (P < 0.001 by log-rank test). CONCLUSIONS: The SFAs measured at the upper thigh level and 1 year mortality are positively associated in elderly patients with proximal femur fracture. SFA may be an independent prognostic biomarker for 1 year mortality of femur fracture.

Prediction of Neurologically Intact Survival in Cardiac Arrest Patients without Pre-Hospital Return of Spontaneous Circulation: Machine Learning Approach.

Current multimodal approaches for the prognostication of out-of-hospital cardiac arrest (OHCA) are based mainly on the prediction of poor neurological outcomes; however, it is challenging to identify patients expected to have a favorable outcome, especially before the return of spontaneous circulation (ROSC). We developed and validated a machine learning-based system to predict good outcome in OHCA patients before ROSC. This prospective, multicenter, registry-based study analyzed non-traumatic OHCA data collected between October 2015 and June 2017. We used information available before ROSC as predictor variables, and the primary outcome was neurologically intact survival at discharge, defined as cerebral performance category 1 or 2. The developed models' robustness were evaluated and compared with various score metrics to confirm their performance. The model using a voting classifier had the best performance in predicting good neurological outcome (area under the curve = 0.926). We confirmed that the six top-weighted variables predicting neurological outcomes, such as several duration variables after the instant of OHCA and several electrocardiogram variables in the voting classifier model, showed significant differences between the two neurological outcome groups. These findings demonstrate the potential utility of a machine learning model to predict good neurological outcome of OHCA patients before ROSC.

Types of cancer and outcomes in patients with cancer requiring admission from the emergency department: A nationwide, population-based study, 2016-2017.

BACKGROUND: Emergency department (ED) utilization and emergency admissions by patients with cancer have increased. The authors aimed to evaluate the characteristics of patients with cancer admitted through the ED and determine whether cancer types are related to in-hospital mortality. METHODS: The National Emergency Department Information System database of patients visiting EDs in South Korea between 2016 and 2017 was analyzed. Among 6,179,088 adult patients who presented to an ED with nontraumatic medical illness, patients with cancer were identified. The primary outcome was in-hospital mortality. RESULTS: Patients with cancer accounted for 6.8% of ED visits, and 239,630 patients (57.0%) were admitted to the hospital (intensive care unit [ICU], 9.5%; others, 90.5%). The prevalent cancers requiring hospitalization were lung cancer (15.7%), liver cancer (14.2%), and colon cancer (11.6%). The commonest reasons for admission other than cancer-related medical problems (41.4%) were pneumonia (4.8%) and hepatobiliary infection (2.8%). Overall in-hospital mortality was 16.1% (ICU, 28.3%; general wards, 14.8%); lung cancer (22.9%), liver cancer (19.7%), and leukemia/multiple myeloma (17.8%) showed the highest mortality rates. The highest odds for mortality were for lung cancer (adjusted odds ratio [OR], 2.227; 95% confidence interval [CI], 2.124-2.335; P < .001) and liver cancer (adjusted OR, 1.839; 95% CI, 1.751-1.930; P < .001), which were referenced to genitourinary cancer by multivariable logistic regression analysis. CONCLUSIONS: More than half of the patients with cancer visiting EDs were admitted to the hospital with a mortality rate of 16.1%. Physicians treating patients with cancer and policymakers and planners designing health systems should understand the different prevalences and outcomes of oncological emergencies by cancer type to improve patient care.

Effects of Patient-Generated Health Data: Comparison of Two Versions of Long-Term Mobile Personal Health Record Usage Logs.

Patient-generated health data (PGHD) can be managed easily by a mobile personal health record (mPHR) and can increase patient engagement. This study investigated the effect of PGHD functions on mPHR usage. We collected usage log data from an mPHR app, My Chart in My Hand (MCMH), for seven years. We analyzed the number of accesses and trends for each menu by age and sex according to the version-up. Generalized estimating equation (GEE) analysis was used to determine the likelihood of continuous app usage according to the menus and version-up. The total number of users of each version were 15,357 and 51,553, respectively. Adult females under 50 years were the most prevalent user group (30.0%). The "My Chart" menu was the most accessed menu, and the total access count increased by ~10 times after the version-up. The "Health Management" menu designed for PGHD showed the largest degree of increase in its likelihood of continuous usage after the version-up (1.245; p < 0.0001) across menus (range: 0.925-1.050). Notably, improvement of PGHD management in adult females over 50 years is needed.

Modification of the HEART pathway by adding coronary computed tomography angiography for patients suspected of acute coronary syndrome in the emergency department.

The HEART (history, ECG, age, risk factors, troponin) pathway (HP) was developed for identifying low-risk patients for early discharge among patients presenting with chest pain to the emergency department (ED). We investigated whether adding coronary computed tomography angiography (CCTA) results to selected patients could improve the diagnostic accuracy of the HP. Patients suspected of acute coronary syndrome who had undergone CCTA were included. The HP was modified by adding CCTA results of stenosis of any major epicardial coronary arteries to patients either with 0-3 points and a positive troponin test or with 4-6 points. All patients were reclassified into low and increased risk groups. We then compared the accuracy of the modified HP, the HP, and the HEART score. The primary outcome was the 30-day major adverse cardiac events (MACE). Of the total 1239 patients included, MACE occurred in 206 (16.6%) patients. Adding the CCTA results increased the proportion of patients with low risk (68.7%) compared with the HP (40.0%) and the HEART score (47.4%). Using the modified HP, 50.4% of patients with intermediate-risk by the HEART score could be discharged from the ED and had no MACE. Incorporation of CCTA results improved the accuracy rate for the prediction of MACE compared with the HP and the HEART score (net reclassification improvements were 34.5 and 39.6%, respectively). Using the CCTA after the HP in selected patients could be a better strategy to discharge more patients early and safely.

Prediction of Adverse Events in Stable Non-Variceal Gastrointestinal Bleeding Using Machine Learning.

Clinical risk-scoring systems are important for identifying patients with upper gastrointestinal bleeding (UGIB) who are at a high risk of hemodynamic instability. We developed an algorithm that predicts adverse events in patients with initially stable non-variceal UGIB using machine learning (ML). Using prospective observational registry, 1439 out of 3363 consecutive patients were enrolled. Primary outcomes included adverse events such as mortality, hypotension, and rebleeding within 7 days. Four machine learning algorithms, namely, logistic regression with regularization (LR), random forest classifier (RF), gradient boosting classifier (GB), and voting classifier (VC), were compared with the Glasgow-Blatchford score (GBS) and Rockall scores. The RF model showed the highest accuracies and significant improvement over conventional methods for predicting mortality (area under the curve: RF 0.917 vs. GBS 0.710), but the performance of the VC model was best in hypotension (VC 0.757 vs. GBS 0.668) and rebleeding within 7 days (VC 0.733 vs. GBS 0.694). Clinically significant variables including blood urea nitrogen, albumin, hemoglobin, platelet, prothrombin time, age, and lactate were identified by the global feature importance analysis. These results suggest that ML models will be useful early predictive tools for identifying high-risk patients with initially stable non-variceal UGIB admitted at an emergency department.

Correction to: Maximum emergency department overcrowding is correlated with occurrence of unexpected cardiac arrest.

An amendment to this paper has been published and can be accessed via the original article.