A model supported in the clinical covid patient record using DL for pandemic preparedness

This paper presents a model based on a Deep Learning approach to aid in improving our assessment of the risk of death in COVID-19 patients only based on their clinical record when admitted. It is aimed to provide an alternative fast tool for doctors and researchers to focus on a rapid selection of patients with a high likelihood of death, which is a critical aspect having in mind the growing infection dynamics of the current and perhaps future pandemics. The dataset used in this research is open-access, available for algorithm benchmarking, and represents the knowledge of the cases examined in Mexico before the immunization campaigns. The massive amount of information used to feed the algorithm provides robustness and aids in detecting the principal patterns involved in the data. The model is based on a Deep Neuronal Network, which uses different activation functions and several neurons in each hidden layer for getting a stable performance, and was tested in both a validation set and test set, obtaining a satisfactory and reliable accuracy of about 93 % for the survival prediction.

I. Introduction

The COVID-19 virus (SARS-CoV-2) has significantly af-fected people’ s lifestyles worldwide. One of the biggest prob-lems of this virus has been the number of deaths that have been registered because of this infection, so the medical sector is one of the sectors with the highest demand at present even after the vaccination process. The COVID-19 pandemic has been broadly recognized as a public health emergency of global concern as described in [1]. There are several epidemiological models around the world to project the number of people infected and the mortality rates of the COVID-19 outbreak as mentioned in [2]. The progress of accurate predictive models is of extreme importance to take adequate actions and aid in making important decisions about the conditions of infected patients. The COVID-19 virus arrived in Mexico in 2020, but before to the vaccination campaigns the number of infections and deaths was increased considerably. One reason for this problem has been the economic problems suffered by most citizens. This situation drives people to leave behind social isolation irresponsibly and massively, besides the public health policies that governments face, as mentioned in [3]. The number of people admitted to the hospital daily for this disease was higher than the government's estimates. This problem leads to the situation where the necessity to establish priority among the patients arises and identify which are the critical patients and with a high risk of dying. Unfortunately, no method or model helps to identify which are the critical patients or there is not a model that determines the patients with a higher likelihood of dying given their comorbidity conditions. This paper proposes a model based on a Deep Neuronal Network (DNN) or also known as a multilayer perceptrons model (MLP) which is based on the clinical record of patients registered in Mexican hospitals for presenting symptoms of COVID-19 and predicts which are the critical patients with a high likelihood of death or a high possibility of living.