Development and External Validation of Clinical Features-based Machine Learning Models for Predicting COVID-19 in the Emergency Department.

Introduction: Timely diagnosis of patients affected by an emerging infectious disease plays a crucial role in treating patients and avoiding disease spread. In prior research, we developed an approach by using machine learning (ML) algorithms to predict serious acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection based on clinical features of patients visiting an emergency department (ED) during the early coronavirus 2019 (COVID-19) pandemic. In this study, we aimed to externally validate this approach within a distinct ED population. Methods: To create our training/validation cohort (model development) we collected data retrospectively from suspected COVID-19 patients at a US ED from February 23-May 12, 2020. Another dataset was collected as an external validation (testing) cohort from an ED in another country from May 12-June 15, 2021. Clinical features including patient demographics and triage information were used to train and test the models. The primary outcome was the confirmed diagnosis of COVID-19, defined as a positive reverse transcription polymerase chain reaction test result for SARS-CoV-2. We employed three different ML algorithms, including gradient boosting, random forest, and extra trees classifiers, to construct the predictive model. The predictive performances were evaluated with the area under the receiver operating characteristic curve (AUC) in the testing cohort. Results: In total, 580 and 946 ED patients were included in the training and testing cohorts, respectively. Of them, 98 (16.9%) and 180 (19.0%) were diagnosed with COVID-19. All the constructed ML models showed acceptable discrimination, as indicated by the AUC. Among them, random forest (0.785, 95% confidence interval [CI] 0.747-0.822) performed better than gradient boosting (0.774, 95% CI 0.739-0.811) and extra trees classifier (0.72, 95% CI 0.677-0.762). There was no significant difference between the constructed models. Conclusion: Our study validates the use of ML for predicting COVID-19 in the ED and demonstrates its potential for predicting emerging infectious diseases based on models built by clinical features with temporal and spatial heterogeneity. This approach holds promise for scenarios where effective diagnostic tools for an emerging infectious disease may be lacking in the future.

Development of a prediction model for emergency medical service witnessed traumatic out-of-hospital cardiac arrest: A multicenter cohort study.

BACKGROUND/PURPOSE: To develop a prediction model for emergency medical technicians (EMTs) to identify trauma patients at high risk of deterioration to emergency medical service (EMS)-witnessed traumatic cardiac arrest (TCA) on the scene or en route. METHODS: We developed a prediction model using the classical cross-validation method from the Pan-Asia Trauma Outcomes Study (PATOS) database from 1 January 2015 to 31 December 2020. Eligible patients aged ≥18 years were transported to the hospital by the EMS. The primary outcome (EMS-witnessed TCA) was defined based on changes in vital signs measured on the scene or en route. We included variables that were immediately measurable as potential predictors when EMTs arrived. An integer point value system was built using multivariable logistic regression. The area under the receiver operating characteristic (AUROC) curve and Hosmer-Lemeshow (HL) test were used to examine discrimination and calibration in the derivation and validation cohorts. RESULTS: In total, 74,844 patients were eligible for database review. The model comprised five prehospital predictors: age <40 years, systolic blood pressure <100 mmHg, respiration rate >20/minute, pulse oximetry <94%, and levels of consciousness to pain or unresponsiveness. The AUROC in the derivation and validation cohorts was 0.767 and 0.782, respectively. The HL test revealed good calibration of the model (p = 0.906). CONCLUSION: We established a prediction model using variables from the PATOS database and measured them immediately after EMS personnel arrived to predict EMS-witnessed TCA. The model allows prehospital medical personnel to focus on high-risk patients and promptly administer optimal treatment.

Cardiac screening in school children: Combining auscultation and electrocardiography with a crowdsourcing model.

BACKGROUND/PURPOSE: School-based cardiac screening is useful for identifying children and adolescents with a high risk of sudden cardiac death. However, because of challenges associated with cost, distance, and human resources, cardiac screening is not widely implemented, especially in rural areas with limited medical resources. This study aims to establish a cloud-based system suitable for mass cardiac screening of schoolchildren in rural areas with limited medical resources. METHODS: Students from three schools were included. They or their guardians completed a simple questionnaire, administered in paper or electronic form. Heart sounds were recorded using an electronic stethoscope. Twelve-lead electrocardiograms (ECGs) were recorded and digitalized. The signals were transmitted through Bluetooth to a tablet computer and then uploaded to a cloud server over Wi-Fi. Crowdsourced pediatric cardiologists reviewed those data from a web-based platform and provided remote consultation. In cases in which abnormal heart sounds or ECGs were noted, the students were referred to the hospital for further evaluation. RESULTS: A total of 1004 students were enrolled in this study. Of the 138 students referred, 62 were diagnosed as having an abnormal heart condition and most had previously been undiagnosed. The interrater agreeability was high. CONCLUSION: An innovative strategy combining a cloud-based cardiac screening system with remote consultation by crowdsourced experts was established. This system allows pediatric cardiologists to provide consultation and make reliable diagnoses. Combined with crowdsourcing, the system constitutes a viable approach for mass cardiac screening in children and adolescents living in rural areas with insufficient medical resources.

Effect of using G-FAST to recognize emergent large vessel occlusion: A city-wide community experience.

BACKGROUND/PURPOSE: A prehospital bypass strategy was suggested for large vessel occlusion. This study aimed to evaluate the effect of a bypass strategy using the gaze-face-arm-speech-time test (G-FAST) implemented in a metropolitan community. METHODS: Pre-notified patients with positive Cincinnati Prehospital Stroke Scale and symptom onset <3 h from July 2016 to December 2017 (pre-intervention period) and those with positive G-FAST and symptom onset <6 h from July 2019 to December 2020 (intervention period) were included. Patients aged <20 years and those with missing in-hospital data were excluded. The primary outcomes were the rates of receiving endovascular thrombectomy (EVT) and intravenous thrombolysis (IVT). The secondary outcomes were total prehospital time, door-to-computed tomography (CT) time, door-to-needle (DTN) time, and door-to-puncture (DTP) time. RESULTS: We included 802 and 695 pre-notified patients from the pre-intervention and intervention periods, respectively. The characteristics of the patients in the two periods were similar. In the primary outcomes, pre-notified patients during the intervention period showed higher rates of receiving EVT (4.49% vs. 15.25%, p < 0.001) and IVT (15.34% vs. 21.58%, p = 0.002). In the secondary outcomes, pre-notified patients during intervention period had longer total prehospital time (mean 23.38 vs 25.23 min, p < 0.001), longer door-to-CT time (median 10 vs 11 min, p < 0.001), longer DTN time (median 53 vs 54.5 min, p < 0.001) but shorter DTP time (median 141 vs 139.5 min, p < 0.001). CONCLUSION: The prehospital bypass strategy with G-FAST showed benefits for stroke patients.

The predictive performance of current termination-of-resuscitation rules in patients following out-of-hospital cardiac arrest in Asian countries: A cross-sectional multicentre study.

BACKGROUND: Termination-of-resuscitation rules (TORRs) in out-of-hospital cardiac arrest (OHCA) patients have been applied in western countries; in Asia, two TORRs were developed and have not been externally validated widely. We aimed to externally validate the TORRs using the registry of Pan-Asian Resuscitation Outcomes Study (PAROS). METHODS: PAROS enrolled 66,780 OHCA patients in seven Asian countries from 1 January 2009 to 31 December 2012. The American Heart Association-Basic Life Support and AHA-ALS (AHA-BLS), AHA-Advanced Life Support (AHA-ALS), Goto, and Shibahashi TORRs were selected. The diagnostic test characteristics and area under the receiver operating characteristic curve (AUC) were calculated. We further determined the most suitable TORR in Asia and analysed the variable differences between subgroups. RESULTS: We included 55,064 patients in the final analysis. The sensitivity, specificity, negative predictive value, positive predictive value, and AUC, respectively, for AHA-BLS, AHA-ALS, Goto, Shibashi TORRs were 79.0%, 80.0%, 19.6%, 98.5%, and 0.80; 48.6%, 88.3%, 9.8%, 98.5%, and 0.60; 53.8%, 91.4%, 11.2%, 99.0%, and 0.73; and 35.0%, 94.2%, 8.4%, 99.0%, and 0.65. In countries using the Goto TORR with PPV<99%, OHCA patients were younger, had more males, a higher rate of shockable rhythm, witnessed collapse, pre-hospital defibrillation, and survival to discharge, compared with countries using the Goto TORR with PPV ≥99%. CONCLUSIONS: There was no single TORR fit for all Asian countries. The Goto TORR can be considered the most suitable; however, a high predictive performance with PPV ≥99% was not achieved in three countries using it (Korea, Malaysia, and Taiwan).

Effect of Field Triage Training on Emergency Medical Technicians in Taipei City.

Injury is a leading cause of death among young adults. An accurately implemented fi led triage scheme (FTS) by emergency medical technicians (EMTs) is the first step for delivering right patients to the right hospital. However, the training effect of FTS on EMTs with different levels and backgrounds has scarcely been reported. We evaluated training effects of FTS among EMTs in Taipei. Standard FTS contains physiologic status, anatomical sites of injury, and mechanism of injury criteria. The intervention was a 30-minute lecture and pre-and-post tests, each containing five questions about trauma severity judgment (i.e., mechanism of injury [2 questions], anatomic sites of injury [2 questions], and physiological status [1 question]). The change in EMT accuracy was measured before and after training. Subgroup analyses were performed across EMTs with different levels and seniorities. From September 1, 2015 to March 31, 2016, 821 EMTs were enrolled, including 740 EMT-intermediates and 81 paramedics. Overall, EMT accuracy improved after the intervention in the intermediate (73.2% vs. 85.5%, p < 0.05) and paramedic (76.0% vs. 85.7%, p <0.01) groups. All trainees showed improvements in physiology and mechanism criteria, but paramedics showed decreased accuracy in anatomic criteria. The subgroup analysis showed that accuracy positively associated with prehospital care experience for major trauma cases 1 year before the training course, and the anatomical criterion accuracy was adversely associated with paramedic seniority. Field triage training can improve EMT accuracy for FTS. The anatomical aspect is more diffi cult to improve and should be emphasized in FTS training courses.

Barriers to bystanders defibrillation: A national survey on public awareness and willingness of bystanders defibrillation☆.

BACKGROUND: After years of setting up public automated external defibrillators (AEDs), the rate of bystander AED use remains low all over the world. This study aimed to assess the public awareness and willingness of bystanders to use AEDs and to investigate the awareness on the Good Samaritan Law (GSL) and the factors associated with the low rate of bystander AED use. METHODS: Using stratified random sampling, national telephone interviews were conducted using an author-designed structured questionnaire. The results were weighted to match the census data in Taiwan. The factors associated with public awareness and willingness of bystanders to use AEDs were analysed by logistic regression. RESULTS: Of the 1073 respondents, only 15.2% had the confidence to recognise public AEDs, and 5.3% of them had the confidence to use the AED. Concerns on immature technique and legal issues remain the most common barriers to AED use by bystanders. Moreover, only 30.8% thought that the public should use AEDs at the scene. Few respondents (9.6%) ever heard of the GSL in Taiwan, and less than 3% understood the meaning of GSL. Positive awareness on AEDs was associated with high willingness of bystanders to use AEDs. Respondents who were less likely to use AEDs as bystanders were healthcare personnel and women. CONCLUSION: The importance of active awareness and the barriers to the use of AEDs among bystanders seemed to have been underestimated in the past years. The relatively low willingness to use AEDs among bystander healthcare providers and women needs further investigation.

A prediction model for patients with emergency medical service witnessed out-of-hospital cardiac arrest.

BACKGROUND/PURPOSE: The study aim was to develop a model for predicting patients with emergency medical service (EMS) witnessed out-of-hospital cardiac arrest (OHCA). METHODS: We used fire-based EMS data from Taipei city to develop the prediction model. Patients included in this study were those who were initially alive, non-traumatic, and age ≧20 years. Data were extracted from electronic records of ambulance run sheets and an Utstein-style OHCA registry. The primary outcome (EMS-witnessed OHCA) was defined as cardiac arrest occurring during the service of emergency medical technicians before arrival at a receiving hospital. Area under the receiver operating characteristic curve (AUROC) and the Hosmer-Lemeshow (HL) test were used to examine discrimination and calibration. The point value system with Youden's J Index was used to find the optimal cut-off value. RESULTS: From 2011 to 2015, a total of 252,771 patients were included. Of them, 660 (0.26%) were EMS-witnessed OHCA. The model, including the predictors of male gender, respiratory rate≦10 cycles/min, heart rate <60 or ≧120 beats/min, systolic blood pressure <100 mmHg, level of consciousness, and oxygen saturation <94%, reached excellent discrimination with an AUROC of 0.94 [95% confidence interval (CI), 0.93-0.95] and excellent calibration (p = 0.42 for HL test) in a randomly selected derivation cohort. The results were comparable to those found in a validation cohort. The optimal cut-off value (≧13) of the tool demonstrated high sensitivity (87.84%) and specificity (86.20%). CONCLUSION: This newly developed prediction model will help identify high-risk patients with EMS-witnessed OHCA.

Correction: Optimal paramedic numbers in resuscitation of patients with out-of-hospital cardiac arrest: A randomized controlled study in a simulation setting.

[This corrects the article DOI: 10.1371/journal.pone.0235315.].

Public knowledge, attitudes and willingness regarding bystander cardiopulmonary resuscitation: A nationwide survey in Taiwan.

BACKGROUND: A low bystander cardiopulmonary resuscitation (CPR) rate is one of the factors associated with low cardiac arrest survival. This study aimed to assess knowledge, attitudes, and willingness towards performing CPR and the barriers for implementation of bystander-initiated CPR. METHODS: Telephone interviews were conducted using an author-designed and validated structured questionnaire in Taiwan. After obtaining a stratified random sample from the census, the results were weighted to match population data. The factors affecting bystander-initiated CPR were analysed using logistic regression. RESULTS: Of the 1073 respondents, half of them stated that they knew how to perform CPR correctly, although 86.7% indicated a willingness to perform CPR on strangers. The barriers to CPR performance reported by the respondents included fear of legal consequences (44%) and concern about harming patients (36.5%). Most participants expressed a willingness to attend only an hour-long CPR course. Respondents who were less likely to indicate a willingness to perform CPR were female, healthcare providers, those who had no cohabiting family members older than 65 years, those who had a history of a stroke, and those who expressed a negative attitude toward CPR. CONCLUSION: The expressed willingness to perform bystander CPR was high if the respondents possessed the required skills. Attempts should be made to recruit potential bystanders for CPR courses or education, targeting those respondent subgroups less likely to express willingness to perform CPR. The reason for lower bystander CPR willingness among healthcare providers deserves further investigation.

The effect and associated factors of dispatcher recognition of stroke: A retrospective observational study.

BACKGROUND/PURPOSE: Details of the communication between the caller and dispatcher have not been reported previously in Taiwan. This study aimed to: (1) understand the details of the communication between the caller and dispatcher among the calls for stroke patients, (2) identify factors associated with stroke recognition by dispatchers, and (3) evaluate the association between stroke recognition by dispatchers and stroke management. METHODS: We conducted a retrospective observational study involving patients with stroke or transient ischemic stroke transported by the emergency medical service, and arriving at 9 hospitals in Taipei within 3 h of symptom onset from January 1, 2013 to February 28, 2014. Patients were excluded if tape-recording data or prehospital information were not available. Data of the enrolled patients were reviewed. We used stroke dispatch determination as the surrogate for stroke recognition by dispatchers. Multivariable logistic regression was used to identify the factors associated with stroke dispatch determination. RESULTS: A total of 507 patients were included. In approximately 50% of cases, callers were close family members. Ninety-one patients (17.9%) had stroke dispatch determination. After adjustment, stroke reported spontaneously, any symptom included in the Cincinnati Prehospital Stroke Scale reported spontaneously, and dispatcher adherence to the protocol, were associated with stroke dispatch determination significantly. Stroke dispatch determination was associated with receiving pre-arrival notification, shorter door-to-computed tomography time, and thrombolytic therapy. CONCLUSION: The dispatchers should spend more time identifying stroke patients by following the dispatch protocol. Recognition of stroke by dispatchers was associated with improved stroke care.

Cardiac arrest and cardiopulmonary resuscitation outcome reports: update of the Utstein Resuscitation Registry Templates for Out-of-Hospital Cardiac Arrest: a statement for healthcare professionals from a task force of the International Liaison Committee on Resuscitation (American Heart Association, European Resuscitation Council, Australian and New Zealand Council on Resuscitation, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Council of Southern Africa, Resuscitation Council of Asia); and the American Heart Association Emergency Cardiovascular Care Committee and the Council on Cardiopulmonary, Critical Care, Perioperative and Resuscitation.

Utstein-style guidelines contribute to improved public health internationally by providing a structured framework with which to compare emergency medical services systems. Advances in resuscitation science, new insights into important predictors of outcome from out-of-hospital cardiac arrest, and lessons learned from methodological research prompted this review and update of the 2004 Utstein guidelines. Representatives of the International Liaison Committee on Resuscitation developed an updated Utstein reporting framework iteratively by meeting face to face, by teleconference, and by Web survey during 2012 through 2014. Herein are recommendations for reporting out-of-hospital cardiac arrest. Data elements were grouped by system factors, dispatch/recognition, patient variables, resuscitation/postresuscitation processes, and outcomes. Elements were classified as core or supplemental using a modified Delphi process primarily based on respondents' assessment of the evidence-based importance of capturing those elements, tempered by the challenges to collect them. New or modified elements reflected consensus on the need to account for emergency medical services system factors, increasing availability of automated external defibrillators, data collection processes, epidemiology trends, increasing use of dispatcher-assisted cardiopulmonary resuscitation, emerging field treatments, postresuscitation care, prognostication tools, and trends in organ recovery. A standard reporting template is recommended to promote standardized reporting. This template facilitates reporting of the bystander-witnessed, shockable rhythm as a measure of emergency medical services system efficacy and all emergency medical services system-treated arrests as a measure of system effectiveness. Several additional important subgroups are identified that enable an estimate of the specific contribution of rhythm and bystander actions that are key determinants of outcome.

Global health and emergency care: a resuscitation research agenda--part 1.

At the 2013 Academic Emergency Medicine global health consensus conference, a breakout session on a resuscitation research agenda was held. Two articles focusing on cardiac arrest and trauma resuscitation are the result of that discussion. This article describes the burden of disease and outcomes, issues in resuscitation research, and global trends in resuscitation research funding priorities. Globally, cardiovascular disease and trauma cause a high burden of disease that receives a disproportionately smaller research investment. International resuscitation research faces unique ethical challenges. It needs reliable baseline statistics regarding quality of care and outcomes; data linkages between providers; reliable and comparable national databases; and an effective, efficient, and sustainable resuscitation research infrastructure to advance the field. Research in resuscitation in low- and middle-income countries is needed to understand the epidemiology, infrastructure and systems context, level of training needed, and potential for cost-effective care to improve outcomes. Research is needed on low-cost models of population-based research, ways to disseminate information to the developing world, and finding the most cost-effective strategies to improve outcomes.

Direct endotracheal intubation using a novel detachable optic probe (Sunscope) by emergency medical technicians with various training backgrounds.

OBJECTIVE: Effective airway management requires both proper technique and the appropriate devices. With the widespread implementation of advanced life-support techniques in emergency medical services (EMS), orotracheal intubation is now performed not only by professional practitioners but, in many occasions, nonprofessionals. With extensively diversified skill equipped, we tested whether the Sunscope, a patented tracheal intubation device with a digital display, is able to facilitate tracheal intubation by naïve EMS personnel with various training backgrounds. METHODS: We conducted a study to determine rate of success and time required to insert an orotracheal tube into a mannequin using Sunscope. The participants were placed into the professional group (i.e., anesthesiologists and emergency medical paramedics; EMT-P) or the nonprofessional group (i.e., emergency medical technicians; EMT). Intubation required three steps: equipment preparation, vocal cord exposure, and tube insertion. The time required for each step was recorded by a senior staff member, and the data were analyzed by nonparametric statistics. RESULTS: Each consecutive step in the operating procedure was significantly shorter for the professional group in comparison with the nonprofessional group during the first trial: equipment preparation, 10.5 ± 2.1 vs. 11.9 ± 4.1 seconds; vocal cord exposure, 7.4 ± 7.7 vs. 12.2 ± 7.7 seconds; tube insertion, 8.8 ± 4.8 vs. 17.6 ± 9.4 seconds; and total time required for intubation, 26.7 ± 8.8 vs. 35.8 ± 19.6 seconds. The professional practitioners showed no significant improvement, in terms of time reduction, on the following three trials. On the other hand, the nonprofessional practitioners showed no significant differences, in terms of time required to expose the vocal cords and total operation time, following the third trial in comparison with the professional practitioners. CONCLUSION: Our research demonstrates that professional practitioners are able to use the Sunscope on their first attempt. Despite a lack of training in conventional endotracheal intubation, emergency medical technicians (EMT-I and -II) were able to complete intubation on their first attempt; a significant reduction in the time required to intubate was noted after repeated practice. All levels of naïve EMTs were able to readily visualize the vocal cords through Sunscope and, thereby, reliably insert the endotracheal tube in less than 1 minute, regardless of their skills before testing.

Ultrasound may decrease the emergency surgery rate of incarcerated inguinal hernia.

OBJECTIVE: Manual reduction is the standard procedure for incarcerated inguinal hernia reduction. The role of ultrasound in incarcerated inguinal hernia reduction has not been defined. The aim of this study was to determine whether ultrasound can improve the ability to reduce incarcerated inguinal hernia safely when manual reduction fails and thereby decrease the emergency surgery rate. MATERIAL AND METHODS: Between January 1994 and December 2003, 112 adult patients with incarcerated inguinal hernias were admitted to a university medical center and classified into two groups. Group I consisted of 61 patients who received emergency surgical reduction after the failure of two attempts at manual reduction. In Group II, comprising 51 patients, ultrasound-guided reduction was performed when two attempts at manual reduction failed. Emergency surgical reduction was scheduled if both ultrasound-guided reduction and two attempts at manual reduction failed. The difference in emergency surgery rates between groups I and II was compared. RESULTS: In group I, manual reduction was successful in 55 cases (90.2%) and 6 patients underwent emergency surgery. In group II, manual reduction was successful in 45 cases, ultrasound-guided reduction in 4 cases, 1 reduction was avoided after ultrasonic examination, and 1 patient underwent emergency surgery. The emergency surgery rates in groups I and II were 9.8% and 2.0%, respectively. CONCLUSIONS: Ultrasound can improve the ability to reduce incarcerated inguinal hernia safely when manual reduction fails and may decrease the rate of emergency surgery.