Abstract 10992: Machine Learning to Predict In-Hospital Cardiac Arrest in Patients Admitted from the Emergency Department with COVID-19 and Suspected Pneumonia

Circulation ◽  
2021 ◽  
Vol 144 (Suppl_2) ◽  
Author(s):  
Tsung-Chien Lu ◽  
Eric H Chou ◽  
CHIH-HUNG WANG ◽  
Amir Mostafavi ◽  
Mario Tovar ◽  
...  
Keyword(s):  
Machine Learning ◽  
Cardiac Arrest ◽  
Characteristic Curve ◽  
Vital Signs ◽  
Early Warning Score ◽  
Gradient Boosting ◽  
Patients At Risk ◽  
End Of Life Decision ◽  
Hospital Cardiac Arrest ◽  
Suspected Pneumonia

Introduction: There are only scarce models developed for stratifying the risk of cardiac arrest from COVID-19 patients presenting to the ED with suspected pneumonia. By using the machine learning (ML) approach, we aimed to develop and validate the ML models to predict in-hospital cardiac arrest (IHCA) in patients admitted from the ED. Hypothesis: We hypothesized that ML approach can serve as a valuable tool in identifying patients at risk of IHCA in a timely fashion. Methods: We included the COVID-19 patients admitted from the EDs of five hospitals in Texas between March and November 2020. All adult (≥ 18 years) patients were included if they had positive RT-PCR for SARS-CoV-2 and also received CXR examination for suspected pneumonia. Patients’ demographic, past medical history, vital signs at ED triage, CXR findings, and laboratory results were retrieved from the EMR system. The primary outcome (IHCA) was identified via a resuscitation code. Patients presented as OHCA or without any blood testing were excluded. Nonrandom splitting strategy based on different location was used to divide the dataset into the training (one urban and two suburban hospitals) and testing cohort (one urban and one suburban hospital) at around 2-to-1 ratio. Three supervised ML models were trained and performances were evaluated and compared with the National Early Warning Score (NEWS) by the area under the receiver operating characteristic curve (AUC). Results: We included 1,485 records for analysis. Of them, 190 (12.8%) developed IHCA. Of the constructed ML models, Random Forest outperformed the others with the best AUC result (0.930, 95% CI: 0.896-0.958), followed by Gradient Boosting (0.929, 95% CI: 0.891-0.959) and Extra Trees classifier (0.909, 95% CI: 0.875-0.943). All constructed ML models performed significantly better than by using the NEWS scoring system (AUC: 0.787, 95% CI: 0.725-0.840). The top six important features selected were age, oxygen saturation at triage, and lab data of APTT, lactic acid, and LDH. Conclusions: The ML approach showed excellent discriminatory performance to identify IHCA for patients with COVID-19 and suspected pneumonia. It has the potential to save more life or provide end-of-life decision making if successfully implemented in the EMR system.

scholarly journals A Multicenter Validation Study of the Deep Learning-based Early Warning Score for Predicting in-hospital Cardiac Arrest in Patients Admitted to General Wards

2020 ◽  
Author(s):  
Yeon Joo Lee ◽  
Kyung-Jae Cho ◽  
Oyeon Kwon ◽  
Hyunho Park ◽  
Yeha Lee ◽  
...  
Keyword(s):  
Cardiac Arrest ◽  
Deep Learning ◽  
Early Warning ◽  
Characteristic Curve ◽  
Predictive Performance ◽  
Early Warning Score ◽  
Patients At Risk ◽  
Key Aspects ◽  
General Wards ◽  
Hospital Cardiac Arrest

Abstract Background: The recently developed deep learning (DL)-based early warning score (DEWS) has shown a potential in predicting deteriorating patients. We aimed to validate DEWS in multiple centers and compare the prediction, alarming and timeliness performance with those of the modified early warning score (MEWS) to identify patients at risk for in-hospital cardiac arrest (IHCA).Methods: This retrospective cohort study included adult patients admitted to the general wards of five hospitals during a 12-month period. We validated DEWS internally at two hospitals and externally at the other three hospitals. The occurrence of IHCA within 24 hours of vital sign observation was the outcome of interest. We used the area under the receiver operating characteristic curve (AUROC) as the main performance metric.Results: The study population consisted of 173,368 patients (224 IHCAs). The predictive performance of DEWS was superior to that of MEWS in both the internal (AUROC: 0.860 vs. 0.754, respectively) and external (AUROC: 0.905 vs. 0.785, respectively) validation cohorts. At the same specificity, DEWS had a higher sensitivity than MEWS, and at the same sensitivity, DEWS had a lower mean alarm count than MEWS, with nearly half of the alarm rate in MEWS. Additionally, DEWS was able to predict more IHCA patients in the 24 to 0.5 hours before the outcome.Conclusion: Our study showed that DEWS was superior to MEWS in the three key aspects (IHCA predictive, alarming, and timeliness performance). This study demonstrates the potential of DEWS as an effective, efficient screening tool in rapid response systems (RRSs) to identify high-risk patients.

Abstract 332: Differences in Vital Signs Between Elderly and Non-Elderly Patients Prior to Ward Cardiac Arrest

Circulation ◽  
2014 ◽  
Vol 130 (suppl_2) ◽  
Author(s):  
Matthew M Churpek ◽  
Trevor C Yuen ◽  
Christopher Winslow ◽  
Jesse B Hall ◽  
Dana P Edelson
Keyword(s):  
Cardiac Arrest ◽  
Elderly Patients ◽  
Characteristic Curve ◽  
Vital Signs ◽  
Patient Characteristics ◽  
Early Warning Score ◽  
Risk Scores ◽  
Study Patient ◽  
Elderly Age ◽  
Age Related

BACKGROUND: Vital signs and composite risk scores, such as the Modified Early Warning Score (MEWS), are used to identify high-risk patients, trigger rapid response teams, and assist with triage decisions. Although age-related vital sign changes are known to occur, little is known about the differences in vital signs between elderly and non-elderly patients prior to ward cardiac arrest (CA). We compared the accuracy of vital signs and the MEWS for detecting CA between elderly and non-elderly patients. METHODS: All patients hospitalized on the wards from five hospitals were included in the study. Patient characteristics and vital signs prior to CA were compared between elderly (age ≥65 years) and non-elderly (age <65 years) patients. The area under the receiver operating characteristic curve (AUC) was also calculated for vital signs and the MEWS for elderly and non-elderly ward patients for CA. RESULTS: A total of 269,956 admissions with documented age occurred during the study period, which included 422 index ward CAs. Within four hours of CA, elderly patients had significantly lower mean heart rate (88 vs. 99 beats per minute; P<0.001), diastolic blood pressure (60 vs. 66 mm Hg; P=0.007), and shock index (0.82 vs. 0.93; P<0.001), and higher pulse pressure index (0.45 vs. 0.41; P<0.001) and temperature (97.6 vs. 97.3 °F; P=0.047). The AUCs for all vital signs and the MEWS were higher for non-elderly patients than elderly patients (MEWS AUC 0.85 (95% CI 0.82-0.88) vs. 0.71 (95% CI 0.68-0.75); P<0.001). While the incidence of CA increased with age, accuracy of the MEWS decreased (Figure). CONCLUSIONS: Vital signs are much more accurate for detecting CA on the wards in non-elderly patients compared to elderly patients, which has important implications for how they are used for identifying critically ill patients. Further investigation into improving the accuracy of risk stratification for elderly patients is necessary in order to decrease their risk for this devastating event.

scholarly journals Implementasi Early Warning Score pada Kejadian Henti Jantung di Ruang Perawatan Rumah Sakit Dr. Hasan Sadikin Bandung yang Ditangani Tim Code Blue Selama Tahun 2017

2019 ◽  
Vol 7 (1) ◽  
pp. 33-41
Author(s):  
Nurul Subhan ◽  
Gezy Weita Giwangkencana ◽  
M. Andy Prihartono ◽  
Doddy Tavianto
Keyword(s):  
Cardiac Arrest ◽  
General Hospital ◽  
Early Warning ◽  
Medical Records ◽  
Vital Signs ◽  
Early Warning Score ◽  
Ward Staff ◽  
Cross Sectional ◽  
Code Blue ◽  
Hospital Cardiac Arrest

Angka kejadian henti jantung di rumah sakit sangat bervariasi. Sebagian besar kasus henti jantung didahului oleh penurunan kondisi pasien yang digambarkan dengan gangguan parameter tanda vital. Keberhasilan Early warning score (EWS) dalam menurunkan angka kejadian henti jantung dipengaruhi oleh implementasi yang baik dari instrumen EWS sesuai dengan pedoman yang ditetapkan. Penelitian ini bertujuan melihat implementasi EWS di RSUP Dr. Hasan Sadikin Bandung. Penelitian bersifat deskriptif dengan desain potong lintang menggunakan data rekam medis pasien henti jantung di ruang perawatan yang ditangani oleh tim Code Blue selama tahun 2017, dan dilakukan pada bulan November 2018. Data EWS 6 jam sebelum dan saat henti jantung, serta tindak lanjut yang dilakukan setelah penilaian EWS dicatat. Didapatkan 87 data rekam medis henti jantung yang memenuhi kriteria inklusi dan tidak termasuk eksklusi. Di antaranya, 72% memiliki catatan EWS lengkap, 9% memiliki catatan EWS tidak lengkap, dan 18% tidak memiliki data EWS. Dari 63 data rekam medis yang memiliki data EWS lengkap hanya 21% yang mendapat tindak lanjut yang sesuai dengan standar prosedur operasional EWS. Simpulan penelitian ini adalah implementasi EWS di ruang rawat inap RSUP Dr. Hasan Sadikin belum cukup memuaskan. Tindak lanjut yang dilakukan setelah penilaian EWS belum sesuai dengan standar prosedur operasional EWS yang berlaku.Implementation of Early Warning Score to Patients with In-Hospital Cardiac Arrest in Dr. Hasan Sadikin General Hospital Managed  by Code Blue Team Incidence of in-hospital cardiac arrest varies greatly around the world. Most in-hospital cardiac arrests are preceded with physiological deteriorations that manifest as alterations in vital signs. The success of early warning score (EWS) in reducing the incidence of cardiac arrest is influenced by the good implementation of EWS instruments by ward staff in accordance with the guidelines The aim of this study was to assess to what degree EWS was implemented at Dr. Hasan Sadikin General Hospital Bandung. This was a cross sectional descriptive study on patients with in-hospital cardiac arrest managed by the code blue team during 2017 that was conducted in November 2018. EWS 6 hour prior to cardiac arrest event, EWS at the event, and action taken upon finding an abnormal value were obtained from medical records.  Eighty seven medical records were included. Of these, 72% medical records had complete EWS data, 9 medical records had incomplete EWS data, and 18% medical records had no EWS recorded. From those 63 medical records with complete EWS recorded, only 21% had been managed correctly according to the EWS guideline. This study concludes that the implementation of EWS in the wards of Dr. Hasan Sadikin General Hospital Bandung has not been completely satisfactorily. Actions taken after EWS assessment are still not accordance with the EWS guideline.

scholarly journals Development and Validation of an Interpretable 3-day Intensive Care Unit Readmission Prediction Model Using Explainable Boosting Machines

2021 ◽  
Author(s):  
Stefan Hegselmann ◽  
Christian Ertmer ◽  
Thomas Volkert ◽  
Antje Gottschalk ◽  
Martin Dugas ◽  
...  
Keyword(s):  
Machine Learning ◽  
Intensive Care Unit ◽  
Intensive Care ◽  
Medical Information ◽  
Characteristic Curve ◽  
External Validation ◽  
Simplified Acute Physiology Score ◽  
Gradient Boosting ◽  
Patients At Risk ◽  
High Level

Intensive care unit readmissions are associated with mortality and bad outcomes. Machine learning could help to identify patients at risk to improve discharge decisions. However, many models are black boxes, so that dangerous properties might remain unnoticed. In this study, an inherently interpretable model for 3-day ICU readmission prediction was developed. We used a retrospective cohort of 15,589 ICU stays and 169 variables collected between 2006 and 2019. A team of doctors inspected the model, checked the plausibility of each component, and removed problematic parts. Qualitative feedback revealed several challenges for interpretable machine learning in healthcare. The resulting model used 67 features and showed an area under the precision-recall curve of 0.119+/-0.020 and an area under the receiver operating characteristic curve of 0.680+/-0.025. This is on par with state-of-the-art gradient boosting machines and outperforms the Simplified Acute Physiology Score II. External validation with the Medical Information Mart for Intensive Care database version IV confirmed our findings. Hence, a machine learning model for readmission prediction with a high level of human control is feasible without sacrificing performance.

Development of A Pharmaceutical Early Warning Scoring Model for Cardiac Arrest Using Deep Learning: Retrospective Cohort Study (Preprint)

2020 ◽  
Author(s):  
Hsiao-Ko Chang ◽  
Hui-Chih Wang ◽  
Chih-Fen Huang ◽  
Feipei Lai
Keyword(s):  
Time Series ◽  
Cardiac Arrest ◽  
Early Warning ◽  
Time Series Data ◽  
Predictive Accuracy ◽  
Vital Signs ◽  
Sliding Window ◽  
Early Warning Score ◽  
Series Data ◽  
Dynamic Time

BACKGROUND In most of Taiwan’s medical institutions, congestion is a serious problem for emergency departments. Due to a lack of beds, patients spend more time in emergency retention zones, which make it difficult to detect cardiac arrest (CA). OBJECTIVE We seek to develop a pharmaceutical early warning model to predict cardiac arrest in emergency departments via drug classification and medical expert suggestion. METHODS We propose a new early warning score model for detecting cardiac arrest via pharmaceutical classification and by using a sliding window; we apply learning-based algorithms to time-series data for a Pharmaceutical Early Warning Scoring Model (PEWSM). By treating pharmaceutical features as a dynamic time-series factor for cardiopulmonary resuscitation (CPR) patients, we increase sensitivity, reduce false alarm rates and mortality, and increase the model’s accuracy. To evaluate the proposed model we use the area under the receiver operating characteristic curve (AUROC). RESULTS Four important findings are as follows: (1) We identify the most important drug predictors: bits, and replenishers and regulators of water and electrolytes. The best AUROC of bits is 85%; that of replenishers and regulators of water and electrolytes is 86%. These two features are the most influential of the drug features in the task. (2) We verify feature selection, in which accounting for drugs improve the accuracy: In Task 1, the best AUROC of vital signs is 77%, and that of all features is 86%. In Task 2, the best AUROC of all features is 85%, which demonstrates that thus accounting for the drugs significantly affects prediction. (3) We use a better model: For traditional machine learning, this study adds a new AI technology: the long short-term memory (LSTM) model with the best time-series accuracy, comparable to the traditional random forest (RF) model; the two AUROC measures are 85%. (4) We determine whether the event can be predicted beforehand: The best classifier is still an RF model, in which the observational starting time is 4 hours before the CPR event. Although the accuracy is impaired, the predictive accuracy still reaches 70%. Therefore, we believe that CPR events can be predicted four hours before the event. CONCLUSIONS This paper uses a sliding window to account for dynamic time-series data consisting of the patient’s vital signs and drug injections. In a comparison with NEWS, we improve predictive accuracy via feature selection, which includes drugs as features. In addition, LSTM yields better performance with time-series data. The proposed PEWSM, which offers 4-hour predictions, is better than the National Early Warning Score (NEWS) in the literature. This also confirms that the doctor’s heuristic rules are consistent with the results found by machine learning algorithms.

Development of A Drug Early Warning System Model for Cardiac Arrest Using Deep Learning: Retrospective Cohort Study (Preprint)

2020 ◽  
Author(s):  
Hsiao-Ko Chang ◽  
Hui-Chih Wang ◽  
Chih-Fen Huang ◽  
Feipei Lai
Keyword(s):  
Machine Learning ◽  
Time Series ◽  
Cardiac Arrest ◽  
Early Warning ◽  
Time Series Data ◽  
Predictive Accuracy ◽  
Vital Signs ◽  
Warning System ◽  
Series Data ◽  
Dynamic Time

BACKGROUND In most of Taiwan’s medical institutions, congestion is a serious problem for emergency departments. Due to a lack of beds, patients spend more time in emergency retention zones, which make it difficult to detect cardiac arrest (CA). OBJECTIVE We seek to develop a Drug Early Warning System Model (DEWSM), it included drug injections and vital signs as this research important features. We use it to predict cardiac arrest in emergency departments via drug classification and medical expert suggestion. METHODS We propose this new model for detecting cardiac arrest via drug classification and by using a sliding window; we apply learning-based algorithms to time-series data for a DEWSM. By treating drug features as a dynamic time-series factor for cardiopulmonary resuscitation (CPR) patients, we increase sensitivity, reduce false alarm rates and mortality, and increase the model’s accuracy. To evaluate the proposed model, we use the area under the receiver operating characteristic curve (AUROC). RESULTS Four important findings are as follows: (1) We identify the most important drug predictors: bits (intravenous therapy), and replenishers and regulators of water and electrolytes (fluid and electrolyte supplement). The best AUROC of bits is 85%, it means the medical expert suggest the drug features: bits, it will affect the vital signs, and then the evaluate this model correctly classified patients with CPR reach 85%; that of replenishers and regulators of water and electrolytes is 86%. These two features are the most influential of the drug features in the task. (2) We verify feature selection, in which accounting for drugs improve the accuracy: In Task 1, the best AUROC of vital signs is 77%, and that of all features is 86%. In Task 2, the best AUROC of all features is 85%, which demonstrates that thus accounting for the drugs significantly affects prediction. (3) We use a better model: For traditional machine learning, this study adds a new AI technology: the long short-term memory (LSTM) model with the best time-series accuracy, comparable to the traditional random forest (RF) model; the two AUROC measures are 85%. It can be seen that the use of new AI technology will achieve better results, currently comparable to the accuracy of traditional common RF, and the LSTM model can be adjusted in the future to obtain better results. (4) We determine whether the event can be predicted beforehand: The best classifier is still an RF model, in which the observational starting time is 4 hours before the CPR event. Although the accuracy is impaired, the predictive accuracy still reaches 70%. Therefore, we believe that CPR events can be predicted four hours before the event. CONCLUSIONS This paper uses a sliding window to account for dynamic time-series data consisting of the patient’s vital signs and drug injections. The National Early Warning Score (NEWS) only focuses on the score of vital signs, and does not include factors related to drug injections. In this study, the experimental results of adding the drug injections are better than only vital signs. In a comparison with NEWS, we improve predictive accuracy via feature selection, which includes drugs as features. In addition, we use traditional machine learning methods and deep learning (using LSTM method as the main processing time series data) as the basis for comparison of this research. The proposed DEWSM, which offers 4-hour predictions, is better than the NEWS in the literature. This also confirms that the doctor’s heuristic rules are consistent with the results found by machine learning algorithms.

scholarly journals An Early Predictive Scoring Model for In-Hospital Cardiac Arrest of Emergent Hemodialysis Patients

2021 ◽  
Vol 10 (15) ◽  
pp. 3241
Author(s):  
Shih-Hao Chen ◽  
Ya-Yun Cheng ◽  
Chih-Hao Lin
Keyword(s):  
Cardiac Arrest ◽  
Risk Group ◽  
Vital Signs ◽  
Area Under The Curve ◽  
Hemodialysis Patients ◽  
High Risk Group ◽  
Predictive Score ◽  
Primary Group ◽  
Sum Scores ◽  
Hospital Cardiac Arrest

Background: Patients undergoing hemodialysis are prone to cardiac arrests. Methods: This study aimed to develop a risk score to predict in-hospital cardiac arrest (IHCA) in emergency department (ED) patients undergoing emergency hemodialysis. Patients were included if they received urgent hemodialysis within 24 h after ED arrival. The primary outcome was IHCA within three days. Predictors included three domains: comorbidity, triage information (vital signs), and initial biochemical results. The final model was generated from data collected between 2015 and 2018 and validated using data from 2019. Results: A total of 257 patients, including 52 with IHCA, were analyzed. Statistical analysis selected significant variables with higher sensitivity cutoff, and scores were assigned based on relative beta coefficient ratio: K > 5.5 mmol/L (score 1), pH < 7.35 (score 1), oxygen saturation < 85% (score 1), and mean arterial pressure < 80 mmHg (score 2). The final scoring system had an area under the curve of 0.78 (p < 0.001) in the primary group and 0.75 (p = 0.023) in the validation group. The high-risk group (defined as sum scores ≥ 3) had an IHCA risk of 47.2% and 41.7%, while the low-risk group (sum scores < 3) had 18.3% and 7%, in the primary and validation databases, respectively. Conclusions: This predictive score model for IHCA in emergent hemodialysis patients could help healthcare providers to take necessary precautions and allocate resources.

scholarly journals Development and validation of an early warning tool for sepsis and decompensation in children during emergency department triage

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Louis Ehwerhemuepha ◽  
Theodore Heyming ◽  
Rachel Marano ◽  
Mary Jane Piroutek ◽  
Antonio C. Arrieta ◽  
...  
Keyword(s):  
Emergency Department ◽  
Severe Sepsis ◽  
Early Warning ◽  
Characteristic Curve ◽  
Vital Signs ◽  
Warning System ◽  
Receiver Operator Characteristic Curve ◽  
Warning Signs ◽  
Gradient Boosting ◽  
Stochastic Gradient Boosting

AbstractThis study was designed to develop and validate an early warning system for sepsis based on a predictive model of critical decompensation. Data from the electronic medical records for 537,837 visits to a pediatric Emergency Department (ED) from March 2013 to December 2019 were collected. A multiclass stochastic gradient boosting model was built to identify early warning signs associated with death, severe sepsis, non-severe sepsis, and bacteremia. Model features included triage vital signs, previous diagnoses, medications, and healthcare utilizations within 6 months of the index ED visit. There were 483 patients who had severe sepsis and/or died, 1102 had non-severe sepsis, 1103 had positive bacteremia tests, and the remaining had none of the events. The most important predictors were age, heart rate, length of stay of previous hospitalizations, temperature, systolic blood pressure, and prior sepsis. The one-versus-all area under the receiver operator characteristic curve (AUROC) were 0.979 (0.967, 0.991), 0.990 (0.985, 0.995), 0.976 (0.972, 0.981), and 0.968 (0.962, 0.974) for death, severe sepsis, non-severe sepsis, and bacteremia without sepsis respectively. The multi-class macro average AUROC and area under the precision recall curve were 0.977 and 0.316 respectively. The study findings were used to develop an automated early warning decision tool for sepsis. Implementation of this model in pediatric EDs will allow sepsis-related critical decompensation to be predicted accurately after a few seconds of triage.

scholarly journals Predictive Value of Estimated Lean Body Mass for Neurological Outcomes after Out-of-Hospital Cardiac Arrest

2020 ◽  
Vol 10 (1) ◽  
pp. 71
Author(s):  
Sung Eun Lee ◽  
Hyuk Hoon Kim ◽  
Minjung Kathy Chae ◽  
Eun Jung Park ◽  
Sangchun Choi
Keyword(s):  
Cardiac Arrest ◽  
Critically Ill ◽  
Lean Body Mass ◽  
Body Mass ◽  
Multivariate Regression ◽  
Characteristic Curve ◽  
Cerebral Performance Category ◽  
High Body Mass Index ◽  
Neurological Outcomes ◽  
Hospital Cardiac Arrest

Background: Postcardiac arrest patients with a return of spontaneous circulation (ROSC) are critically ill, and high body mass index (BMI) is ascertained to be associated with good prognosis in patients with a critically ill condition. However, the exact mechanism has been unknown. To assess the effectiveness of skeletal muscles in reducing neuronal injury after the initial damage owing to cardiac arrest, we investigated the relationship between estimated lean body mass (LBM) and the prognosis of postcardiac arrest patients. Methods: This retrospective cohort study included adult patients with ROSC after out-of-hospital cardiac arrest from January 2015 to March 2020. The enrolled patients were allocated into good- and poor-outcome groups (cerebral performance category (CPC) scores 1–2 and 3–5, respectively). Estimated LBM was categorized into quartiles. Multivariate regression models were used to evaluate the association between LBM and a good CPC score. The area under the receiver operating characteristic curve (AUROC) was assessed. Results: In total, 155 patients were analyzed (CPC score 1–2 vs. 3–5, n = 70 vs. n = 85). Patients’ age, first monitored rhythm, no-flow time, presumed cause of arrest, BMI, and LBM were different (p < 0.05). Fourth-quartile LBM (≥48.98 kg) was associated with good neurological outcome of postcardiac arrest patients (odds ratio = 4.81, 95% confidence interval (CI), 1.10–25.55, p = 0.04). Initial high LBM was also a predictor of good neurological outcomes (AUROC of multivariate regression model including LBM: 0.918). Conclusions: Initial LBM above 48.98kg is a feasible prognostic factor for good neurological outcomes in postcardiac arrest patients.

Sign in / Sign up

Export Citation Format

Share Document