Predicting the deterioration of COVID-19 patients in the emergency department

• Aakash Kaku  |  Ben Zhang  |  Carlos Fernandez-Granda  |  David Kudlowitz  |  Duo Wang  |  Farah E. Shamout  |  Jan Witowski  |  Jungkyu Park  |  Krzysztof J. Geras  |  Lea Azour  |  Meng Cao  |  Nan Wu  |  Narges Razavian  |  Siddhant Dogra  |  Stanislaw Jastrzebski  |  Taro Makino  |  William Moore  |  Yindalon Aphinyanaphongs  |  Yiqiu Shen  |  Yvonne W. Lui

• Center for Advanced Imaging Innovation and Research NYU Langone Health  |  Center for Data Science New York University  |  Department of Mathematics Courant Institute New York University  |  Department of Medicine NYU Langone Health  |  Department of Population Health NYU Langone Health  |  Department of Radiology NYU Langone Health  |  Engineering Division NYU Abu Dhabi  |  Vilcek Institute of Graduate Biomedical Sciences NYU Grossman School of Medicin

During the coronavirus disease 2019 (COVID-19) pandemic, rapid and accurate triage of patients at the emergency department is critical to inform decision-making. This Model proposes a data-driven approach for automatic prediction of deterioration risk using a deep neural network that learns from chest X-ray images and a gradient boosting model that learns from routine clinical variables. Model's AI prognosis system, trained using data from 3,661 patients, achieves an area under the receiver operating characteristic curve (AUC) of 0.786 (95% CI: 0.745-0.830) when predicting deterioration within 96 hours. The deep neural network extracts informative areas of chest X-ray images to assist clinicians in interpreting the predictions and performs comparably to two radiologists in a reader study. In order to verify performance in a real clinical setting, model owners silently deployed a preliminary version of the deep neural network at New York University Langone Health during the first wave of the pandemic, which produced accurate predictions in real-time. In summary, model findings demonstrate the potential of the proposed system for assisting front-line physicians in the triage of COVID-19 patients.

Input variables : EHR Records, Chest X-Ray, RT- PCR Report
Output Variables : 24 hrs, 48 hrs, 72 hrs, 96 hrs findings

Visit Model : github.com