Machine Learning Models for Analysis of Vital Signs Dynamics: A Case for Sepsis Onset Prediction

Objective Achieving accurate prediction of sepsis detection moment based on bedside monitor data in the intensive care unit (ICU). A good clinical outcome is more probable when onset is suspected and treated on time, thus early insight of sepsis onset may save lives and reduce costs. Methodology We present a novel approach for feature extraction, which focuses on the hypothesis that unstable patients are more prone to develop sepsis during ICU stay. These features are used in machine learning algorithms to provide a prediction of a patient's likelihood to develop sepsis during ICU stay, hours before it is diagnosed. Results Five machine learning algorithms were implemented using R software packages. The algorithms were trained and tested with a set of 4 features which represent the variability in vital signs. These algorithms aimed to calculate a patient's probability to become septic within the next 4 hours, based on recordings from the last 8 hours. The best area under the curve (AUC) was achieved with Support Vector Machine (SVM) with radial basis function, which was 88.38%. Conclusions The high level of predictive accuracy along with the simplicity and availability of input variables present great potential if applied in ICUs. Variability of a patient's vital signs proves to be a good indicator of one's chance to become septic during ICU stay.

[1]  Max Kuhn,et al.  Building Predictive Models in R Using the caret Package , 2008 .

[2]  The Role of Approximate Entropy in Predicting Ventricular Defibrillation Threshold , 2002, Journal of cardiovascular pharmacology and therapeutics.

[3]  Manabu Kano,et al.  Epileptic Seizure Prediction Based on Multivariate Statistical Process Control of Heart Rate Variability Features , 2016, IEEE Transactions on Biomedical Engineering.

[4]  P. Melillo,et al.  Automatic Prediction of Cardiovascular and Cerebrovascular Events Using Heart Rate Variability Analysis , 2015, PloS one.

[5]  M. Kubát.  An Introduction to Machine Learning , 2017, Springer International Publishing.

[6]  M. Engoren,et al.  Approximate entropy of respiratory rate and tidal volume during weaning from mechanical ventilation. , 1998, Critical care medicine.

[7]  A. Auerbach,et al.  Diagnostic accuracy and effectiveness of automated electronic sepsis alert systems: A systematic review. , 2015, Journal of hospital medicine.

[8]  R. Bellomo,et al.  The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). , 2016, JAMA.

[9]  C. Moore,et al.  Predictive models for severe sepsis in adult ICU patients , 2015, 2015 Systems and Information Engineering Design Symposium.

[10]  H. Renz,et al.  Inflammatory markers in SIRS, sepsis and septic shock. , 2008, Current medicinal chemistry.

[11]  J. E. Carceller.  American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis , 1992, Critical care medicine.

[12]  Cheng-Deng Kuo,et al.  Heart rate variability measures as predictors of in-hospital mortality in ED patients with sepsis. , 2008, The American journal of emergency medicine.

[13]  Uli K. Chettipally,et al.  Multicentre validation of a sepsis prediction algorithm using only vital sign data in the emergency department, general ward and ICU , 2018, BMJ Open.

[14]  J. Vincent,et al.  The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure , 1996, Intensive Care Medicine.

[15]  Christopher W. Barton,et al.  A computational approach to early sepsis detection , 2016, Comput. Biol. Medicine.

[16]  S. Calvano,et al.  Physiologic Variability at the Verge of Systemic Inflammation: Multiscale Entropy of Heart Rate Variability Is Affected by Very Low Doses of Endotoxin , 2015, Shock.

[17]  K. Giuliano,et al.  Impact of protocol watch on compliance with the surviving sepsis campaign. , 2011, American journal of critical care : an official publication, American Association of Critical-Care Nurses.

[18]  C. Sprung,et al.  Surviving Sepsis Campaign: International Guidelines for Management of Severe Sepsis and Septic Shock 2012 , 2013, Critical care medicine.

[19]  Shamim Nemati,et al.  An Interpretable Machine Learning Model for Accurate Prediction of Sepsis in the ICU , 2017, Critical care medicine.

[20]  U. Rajendra Acharya,et al.  Sudden cardiac death (SCD) prediction based on nonlinear heart rate variability features and SCD index , 2016, Appl. Soft Comput..

[21]  E. Santiago,et al.  Physiologically-based, predictive analytics using the heart-rate-to-Systolic-Ratio significantly improves the timeliness and accuracy of sepsis prediction compared to SIRS. , 2017, American journal of surgery.

[22]  Steven M. Pincus,et al.  Approximate Entropy of Heart Rate as a Correlate of Postoperative Ventricular Dysfunction , 1993, Anesthesiology.

[23]  G. Decavalas,et al.  Severe Sepsis and Septic Shock , 2018 .

[24]  Alice L Givan,et al.  Decreased physiologic variability as a generalized response to human endotoxemia* , 2005, Critical care medicine.

[25]  Nan Liu,et al.  Heart rate variability based machine learning models for risk prediction of suspected sepsis patients in the emergency department , 2019, Medicine.

[26]  T G Buchman,et al.  Experimental human endotoxemia increases cardiac regularity: results from a prospective, randomized, crossover trial. , 1996, Critical care medicine.