Prediction of Good Neurological Recovery after Out-of-hospital Cardiac Arrest: A Machine Learning Analysis.

BACKGROUND This study aimed to train, validate and compare predictive models that use machine learning analysis for good neurological recovery in OHCA patients. METHODS Adult OHCA patients who had a presumed cardiac etiology and a sustained return of spontaneous circulation between 2013 and 2016 were analyzed; 80% of the individuals were analyzed for training and 20% were analyzed for validation. We developed using six machine learning algorithms: logistic regression (LR), extreme gradient boosting (XGB), support vector machine, random forest, elastic net (EN), and neural network. Variables that could be obtained within 24hours of the emergency department visit were used. The area under the receiver operation curve (AUROC) was calculated to assess the discrimination. Calibration was assessed by the Hosmer-Lemeshow test. Reclassification was assessed by using the continuous net reclassification index (NRI). RESULTS A total of 19,860 OHCA patients were included in the analysis. Of the 15,888 patients in the training group, 2,228 (14.0%) had a good neurological recovery; of the 3,972 patients in the validation group, 577 (14.5%) had a good neurological recovery. The LR, XGB, and EN models showed the highest discrimination powers (AUROC (95% CI)) of 0.949 (0.941∼0.957) for all), and all three models were well calibrated (Hosmer-Lemeshow test: p >0.05). The XGB model reclassified patients according to their true risk better than the LR model (NRI: 0.110), but the EN model reclassified patients worse than the LR model (NRI: -1.239). CONCLUSION The best performing machine learning algorithm was the XGB and LR algorithm.

[1]  R. Berg,et al.  Global incidences of out-of-hospital cardiac arrest and survival rates: Systematic review of 67 prospective studies. , 2010, Resuscitation.

[2]  Gavin D Perkins,et al.  EuReCa ONE-27 Nations, ONE Europe, ONE Registry: A prospective one month analysis of out-of-hospital cardiac arrest outcomes in 27 countries in Europe. , 2016, Resuscitation.

[3]  K. Sugiyama,et al.  Neurological outcomes and duration from cardiac arrest to the initiation of extracorporeal membrane oxygenation in patients with out-of-hospital cardiac arrest: a retrospective study , 2017, Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine.

[4]  H. Zou,et al.  Regularization and variable selection via the elastic net , 2005 .

[5]  Han Nee Gan,et al.  Modifiable Factors Associated With Survival After Out‐of‐Hospital Cardiac Arrest in the Pan‐Asian Resuscitation Outcomes Study , 2017, Annals of emergency medicine.

[6]  Chang Bae Park,et al.  A comparison of outcomes of out-of-hospital cardiac arrest with non-cardiac etiology between emergency departments with low- and high-resuscitation case volume. , 2012, Resuscitation.

[7]  Yoshikazu Goto,et al.  Decision-tree model for predicting outcomes after out-of-hospital cardiac arrest in the emergency department , 2013, Critical Care.

[8]  Alain Cariou,et al.  The CAHP (Cardiac Arrest Hospital Prognosis) score: a tool for risk stratification after out-of-hospital cardiac arrest. , 2016, European heart journal.

[9]  Mohammad Mehdi Sepehri,et al.  Toward analyzing and synthesizing previous research in early prediction of cardiac arrest using machine learning based on a multi-layered integrative framework , 2018, J. Biomed. Informatics.

[10]  Kei-Hoi Cheung,et al.  Predicting urinary tract infections in the emergency department with machine learning , 2018, PloS one.

[11]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[12]  Tianqi Chen,et al.  Higgs Boson Discovery with Boosted Trees , 2014, HEPML@NIPS.

[13]  K. Song,et al.  Scene time interval and good neurological recovery in out‐of‐hospital cardiac arrest , 2017, The American journal of emergency medicine.

[14]  Henry J. Lowe,et al.  Discretization of continuous features in clinical datasets , 2013, J. Am. Medical Informatics Assoc..

[15]  Nan Liu,et al.  Prediction of adverse cardiac events in emergency department patients with chest pain using machine learning for variable selection , 2014, BMC Medical Informatics and Decision Making.

[16]  K. Song,et al.  Effect of national implementation of utstein recommendation from the global resuscitation alliance on ten steps to improve outcomes from Out-of-Hospital cardiac arrest: a ten-year observational study in Korea , 2017, BMJ Open.

[17]  K. Enfield,et al.  C‐GRApH: A Validated Scoring System for Early Stratification of Neurologic Outcome After Out‐of‐Hospital Cardiac Arrest Treated With Targeted Temperature Management , 2017, Journal of the American Heart Association.

[18]  Marti A. Hearst.  Trends & Controversies: Support Vector Machines , 1998, IEEE Intell. Syst..

[19]  B. McNally,et al.  Increasing hospital volume is not associated with improved survival in out of hospital cardiac arrest of cardiac etiology. , 2012, Resuscitation.

[20]  Caroline Fenlon,et al.  A discussion of calibration techniques for evaluating binary and categorical predictive models. , 2018, Preventive veterinary medicine.

[21]  Sang Do Shin,et al.  A trend in epidemiology and outcomes of out-of-hospital cardiac arrest by urbanization level: a nationwide observational study from 2006 to 2010 in South Korea. , 2013, Resuscitation.

[22]  E. Vicaut,et al.  Cardiopulmonary resuscitation duration and survival in out-of-hospital cardiac arrest patients. , 2017, Resuscitation.

[23]  P. Steen,et al.  Predicting survival with good neurologic recovery at hospital admission after successful resuscitation of out-of-hospital cardiac arrest: the OHCA score. , 2007, European heart journal.

[24]  H. Moon,et al.  Prehospital advanced cardiac life support by EMT with a smartphone‐based direct medical control for nursing home cardiac arrest , 2018, The American journal of emergency medicine.

[25]  J. Kai,et al.  Can machine-learning improve cardiovascular risk prediction using routine clinical data? , 2017, PloS one.

[26]  Peter Kokol,et al.  Intelligent analysis in predicting outcome of out-of-hospital cardiac arrest , 2009, Comput. Methods Programs Biomed..

[27]  L. Morrison,et al.  A comparison of the universal TOR Guideline to the absence of prehospital ROSC and duration of resuscitation in predicting futility from out-of-hospital cardiac arrest. , 2017, Resuscitation.

[28]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[29]  Aung Myat,et al.  Out-of-hospital cardiac arrest: current concepts , 2018, The Lancet.

[30]  B. Mourvillier,et al.  Predicting survival with good neurological recovery at hospital admission after successful resuscitation of out-of-hospital cardiac arrest: the OHCA score. , 2006, European heart journal.

[31]  G. Dorffner,et al.  A prediction tool for initial out-of-hospital cardiac arrest survivors. , 2014, Resuscitation.

[32]  C. Callaway,et al.  Predicting survival after out-of-hospital cardiac arrest: role of the Utstein data elements. , 2010, Annals of emergency medicine.

[33]  C. Callaway,et al.  Early coronary angiography and percutaneous coronary intervention are associated with improved outcomes after out of hospital cardiac arrest. , 2018, Resuscitation.

[34]  J. Kwon,et al.  An Algorithm Based on Deep Learning for Predicting In‐Hospital Cardiac Arrest , 2018, Journal of the American Heart Association.