2016/10/27 by Ahmed M. Alaa, Alaa, Ahmed M., Jinsung Yoon +5 · 1 citation
Computer Science · Medicine · #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Gaussian Processes and Bayesian Inference #Healthcare Technology and Patient Monitoring #Machine Learning in Healthcare #Sepsis Diagnosis and Treatment
paper · pdf · doi:10.48550/arxiv.1610.08853
openalex publication_date 2016/10/27 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Objective: In this paper, we develop a personalized real-time risk scoring\nalgorithm that provides timely and granular assessments for the clinical acuity\nof ward patients based on their (temporal) lab tests and vital signs; the\nproposed risk scoring system ensures timely intensive care unit (ICU)\nadmissions for clinically deteriorating patients. Methods: The risk scoring\nsystem learns a set of latent patient subtypes from the offline electronic\nhealth record data, and trains a mixture of Gaussian Process (GP) experts,\nwhere each expert models the physiological data streams associated with a\nspecific patient subtype. Transfer learning techniques are used to learn the\nrelationship between a patient's latent subtype and her static admission\ninformation (e.g. age, gender, transfer status, ICD-9 codes, etc). Results:\nExperiments conducted on data from a heterogeneous cohort of 6,321 patients\nadmitted to Ronald Reagan UCLA medical center show that our risk score\nsignificantly and consistently outperforms the currently deployed risk scores,\nsuch as the Rothman index, MEWS, APACHE and SOFA scores, in terms of\ntimeliness, true positive rate (TPR), and positive predictive value (PPV).\nConclusion: Our results reflect the importance of adopting the concepts of\npersonalized medicine in critical care settings; significant accuracy and\ntimeliness gains can be achieved by accounting for the patients' heterogeneity.\nSignificance: The proposed risk scoring methodology can confer huge clinical\nand social benefits on more than 200,000 critically ill inpatient who exhibit\ncardiac arrests in the US every year.\n