Prediction of survival for severe Covid-19 patients with three clinical features: development of a machine learning-based prognostic model with clinical data in Wuhan

Abstract

The swift spread of COVID-19 epidemic has attracted worldwide attentions since Dec., 2019. Till date, 77,041 confirmed Chinese cases have been reported by National Health Commission of P.R. China with 9,126 critical cases whose survival rate is quite low. Meanwhile, COVID-19 epidemic emergence within the other countries (e.g., Korea, Italy, Japan and Iran) is also remarkable with the increasing spread speed. It plays a more and more important role to efficiently and precisely predict the survival rate for critically ill Covid-19 patients as more fatal cases can be targeted interfered in advanced. However, the survival rates of all the present critically ill COVID -19 patients are estimated manually from over 300 laboratory and clinical features, which inevitably leads to high misdiagnose and missed-diagnose rate due to lack of experience and priori knowledge. As a remedy, we have developed a machine learning-based prognostic model that precisely predicts the survival for individual severe patients with more than 90% accuracy with clinical data collected from Tongji hospital, Wuhan. Significantly, such model only requires three key clinical features, i.e., lactic dehydrogenase (LDH), lymphocyte and High-sensitivity C-reactive protein (hsCRP) out of all 300+ features. The rationality of such mere three features may lie in that they are the representatives of tissue injury-, immunity- and inflammation-typed indices, respectively. From the COVID-19 patient diagnosis aspect, the work actualizes low-cost and prompt criticality classification and survival prediction before targeted intervening and diagnosis, especially for the triage of the large-scale explosive epidemic COVID -19 cases.