scholarly journals Clinical Phenotyping of Out-of-Hospital Cardiac Arrest Patients With Shockable Rhythm　― Machine Learning-Based Unsupervised Cluster Analysis ―

2021 ◽  
Author(s):  
Yohei Okada ◽  
Sho Komukai ◽  
Tetsuhisa Kitamura ◽  
Takeyuki Kiguchi ◽  
Taro Irisawa ◽  
...  
Keyword(s):  
Machine Learning ◽  
Cluster Analysis ◽  
Cardiac Arrest ◽  
Unsupervised Cluster ◽  
Shockable Rhythm ◽  
Clinical Phenotyping ◽  
Hospital Cardiac Arrest ◽  
Unsupervised Cluster Analysis

Abstract 408: Predictors of Survival to Hospital Discharge, a Machine Learning Modelfor Out-Of-Hospital Cardiac Arrest in the Young

Circulation ◽  
2019 ◽  
Vol 140 (Suppl_2) ◽  
Author(s):  
Alyssa Vermeulen ◽  
Hoang Nguyen ◽  
Hai Nguyen ◽  
Teri Campbell ◽  
Marina Del Rios
Keyword(s):  
Machine Learning ◽  
Cardiac Arrest ◽  
Hospital Discharge ◽  
Sampling Technique ◽  
Healthcare Facility ◽  
Machine Learning Techniques ◽  
Positive Trend ◽  
Shockable Rhythm ◽  
Advanced Airway ◽  
Hospital Cardiac Arrest

Background: Many efforts in the area of Out of Hospital Cardiac Arrest (OHCA) have been made to enhance diagnosis, interventions, and increase survival; there is little data on OHCA in the young (1-35 years old). Machine learning (ML) enhances medical diagnosis and decision making with recent models allowing clinicians better control and interpretation of their features. Objective: To develop a machine learning model to predict a young patient with OHCA survival to hospital discharge. The ML model will be used to identify important factors contributing to this predictive model. Methods: Utilizing the CARES database in Chicago, IL, from 2013 to 2017, all OHCA in ages 1-35 years were analyzed; the primary outcome of interest is survival to hospital discharge. Eight machine learning techniques were applied to classify survival to hospital discharge. XGBoost was used with decision trees. Synthetic Minority Over-sampling Technique was used to over-sample the under-represented population. All statistics were performed using Python. Results: 744 events were analyzed from 2013 to 2017. Median age was 24.6 years, of these 19.2% were 18 years and younger. The sample was 46% black, 31% caucasian and 19% Hispanic and 68% were male. Presumed cardiac etiology was identified in 61%; 13% had a shockable rhythm; 59% of events were unwitnessed. The model was able to classify survival and death with an accuracy of 90% and AUC of 0.98. Strongest positive correlation of survival to discharge was seen with sustained ROSC, first rhythm type being a shockable rhythm, and use of hypothermia. There was correlation with the year arrest occurred, with positive trend of survival over the study period time. There was a positive trend toward survival with advanced airway and epinephrine use. There was a negative trend toward survival with unwitnessed arrest and location outside of home or healthcare facility. Conclusions: The XGBoost model showed excellent classification of those who survive to discharge and die with OHCA in those 1-35 years old. Most rhythms in this population were not shockable rhythms despite over half of arrests being presumed cardiac in nature. Standard machine learning models can help identify and determine relevant interventions to improve OHCA in the young.

scholarly journals Outcomes of out-of-hospital cardiac arrest in Beijing: a 5-year cross-sectional study

BMJ Open ◽  
2021 ◽  
Vol 11 (4) ◽  
pp. e041917
Author(s):  
Fei Shao ◽  
Haibin Li ◽  
Shengkui Ma ◽  
Dou Li ◽  
Chunsheng Li
Keyword(s):  
Cardiac Arrest ◽  
Urban Areas ◽  
Patient Characteristics ◽  
Cross Sectional Study ◽  
Sectional Study ◽  
Cross Sectional ◽  
Term Survival ◽  
Shockable Rhythm ◽  
Bystander Cpr ◽  
Hospital Cardiac Arrest

ObjectiveThe purpose of this study was to assess the trends in outcomes of out-of-hospital cardiac arrest (OHCA) in Beijing over 5 years.DesignCross-sectional study.MethodsAdult patients with OHCA of all aetiologies who were treated by the Beijing emergency medical service (EMS) between January 2013 and December 2017 were analysed. Data were collected using the Utstein Style. Cases were followed up for 1 year. Descriptive statistics were used to characterise the sample and logistic regression was performed.ResultsOverall, 5016 patients with OHCA underwent attempted resuscitation by the EMS in urban areas of Beijing during the study period. Survival to hospital discharge was 1.2% in 2013 and 1.6% in 2017 (adjusted rate ratio=1.0, p for trend=0.60). Survival to admission and neurological outcome at discharge did not significantly improve from 2013 to 2017. Patient characteristics and the aetiology and location of cardiac arrest were consistent, but there was a decrease in the initial shockable rhythm (from 6.5% to 5.6%) over the 5 years. The rate of bystander cardiopulmonary resuscitation (CPR) increased steadily over the years (from 10.4% to 19.4%).ConclusionSurvival after OHCA in urban areas of Beijing did not improve significantly over 5 years, with long-term survival being unchanged, although the rate of bystander CPR increased steadily, which enhanced the outcomes of patients who underwent bystander CPR.

scholarly journals Multiple cluster analysis for the identification of high-risk census tracts for out-of-hospital cardiac arrest (OHCA) in Denver, Colorado

Resuscitation ◽  
2014 ◽  
Vol 85 (12) ◽  
pp. 1667-1673 ◽  
Author(s):  
Ariann F. Nassel ◽  
Elisabeth D. Root ◽  
Jason S. Haukoos ◽  
Kevin McVaney ◽  
Christopher Colwell ◽  
...  
Keyword(s):  
Cluster Analysis ◽  
Cardiac Arrest ◽  
High Risk ◽  
Census Tracts ◽  
Hospital Cardiac Arrest ◽  
Multiple Cluster

Abstract 170: Challenges in Implementing an ECPR Program for Out of Hospital Cardiac Arrest

Circulation ◽  
2018 ◽  
Vol 138 (Suppl_2) ◽  
Author(s):  
Meshe Chonde ◽  
Jeremiah Escajeda ◽  
Jonathan Elmer ◽  
Frank X Guyette ◽  
Arthur Boujoukos ◽  
...  
Keyword(s):  
Cardiac Arrest ◽  
Inclusion Criteria ◽  
Extracorporeal Cardiopulmonary Resuscitation ◽  
Implementation Challenges ◽  
Shockable Rhythm ◽  
Bystander Cpr ◽  
Stay And Play ◽  
Load And Go ◽  
Mechanical Cpr ◽  
Hospital Cardiac Arrest

Introduction: Extracorporeal cardiopulmonary resuscitation (ECPR) can treat cardiac arrest refractory to conventional therapy. Many institutions are interested in developing their own ECPR program. However, there are challenges in logistics and implementation. Hypothesis: Development of an ECPR team and identification of UPMC Presbyterian as a receiving center will increase recognition of potential ECPR candidates. Methods: We developed an infrastructure of Emergency Medical Services (EMS), Medic Command, and an in-hospital ECPR team. We identified inclusion criteria for patients with an out of hospital cardiac arrest (OHCA) likely to have a reversible arrest etiology and developed them into a simple checklist. These criteria were: witnessed arrest with bystander CPR, shockable rhythm, and ages 18 to 60. We trained local EMS crews to screen patients and review the checklist with a Command Physician prior to transport to our hospital. Results: From October 2015 to March 31 st 2018, there were 1165 dispatches for OHCA, of which 664 (57%) were treated and transported to the hospital and 120 to our institution. Of these, five patients underwent ECPR. Of the remaining cases, 64 (53%) had nonshockable rhythms, 48 (40%) were unwitnessed arrests, 50 (42%) were over age 60 and the remaining 20 (17%) had no documented reasons for exclusion. Prehospital CPR duration was 26 [IQR 25-40] min. Four patients (80%) underwent mechanical CPR with LUCAS device. Time from arrest to arrive on scene was 5 [IQR 4-6] min and time call MD command was 13 [IQR 7-21] min. Time to transport was 20 [IQR 19-21] min. Time from arrest to initiation of ECMO was 63 [IQR 59-69] min. Conclusions: ECPR is a relatively infrequent occurrence. Implementation challenges include prompt identification of patients with reversible OHCA causes, preferential transport to an ECPR capable facility and changing the focus of EMS in these select patients from a “stay and play” to a “load and go” mentality.

scholarly journals Emergency dispatch process and patient outcome in bystander-witnessed out-of-hospital cardiac arrest with a shockable rhythm

2015 ◽  
Vol 22 (4) ◽  
pp. 266-272 ◽  
Author(s):  
Pamela V.C. Hiltunen ◽  
Tom O. Silfvast ◽  
T. Helena Jäntti ◽  
Markku J. Kuisma ◽  
Jouni O. Kurola
Keyword(s):  
Patient Outcome ◽  
Cardiac Arrest ◽  
Shockable Rhythm ◽  
Hospital Cardiac Arrest

scholarly journals Identification of Tumor Microenvironment-Related Alternative Splicing Events to Predict the Prognosis of Endometrial Cancer

2021 ◽  
Vol 11 ◽  
Author(s):  
Xuan Liu ◽  
Chuan Liu ◽  
Jie Liu ◽  
Ying Song ◽  
Shanshan Wang ◽  
...  
Keyword(s):  
Cluster Analysis ◽  
Endometrial Cancer ◽  
Alternative Splicing ◽  
Tumor Microenvironment ◽  
Risk Score ◽  
Risk Group ◽  
Gene Signature ◽  
Unsupervised Cluster ◽  
Score Model ◽  
Unsupervised Cluster Analysis

BackgroundEndometrial cancer (EC) is one of the most common female malignant tumors. The immunity is believed to be associated with EC patients’ survival, and growing studies have shown that aberrant alternative splicing (AS) might contribute to the progression of cancers.MethodsWe downloaded the clinical information and mRNA expression profiles of 542 tumor tissues and 23 normal tissues from The Cancer Genome Atlas (TCGA) database. ESTIMATE algorithm was carried out on each EC sample, and the OS-related different expressed AS (DEAS) events were identified by comparing the high and low stromal/immune scores groups. Next, we constructed a risk score model to predict the prognosis of EC patients. Finally, we used unsupervised cluster analysis to compare the relationship between prognosis and tumor immune microenvironment.ResultsThe prognostic risk score model was constructed based on 16 OS-related DEAS events finally identified, and then we found that compared with high-risk group the OS in the low-risk group was notably better. Furthermore, according to the results of unsupervised cluster analysis, we found that the better the prognosis, the higher the patient’s ESTIMATE score and the higher the infiltration of immune cells.ConclusionsWe used bioinformatics to construct a gene signature to predict the prognosis of patients with EC. The gene signature was combined with tumor microenvironment (TME) and AS events, which allowed a deeper understanding of the immune status of EC patients, and also provided new insights for clinical patients with EC.

Sign in / Sign up

Export Citation Format

Share Document