A Grey Combined Prediction Model for Medical Treatment Risk Analysis during Pandemics

Abstract

The role of information systems (IS) were widely discoursed during the spread of the COVID-19 outbreak. We have focused on developing a decision support systems (DSS) based on a combined prediction model, that can essentially be used at the start of any pandemic. Convalescent plasma therapy is generally applied during the spread of a pandemic as a therapy method that transfuses blood plasma from the people, who have recovered from an illness to treat critical cases. We observe, analyse, and predict the risks associated with the treatment effects of convalescent plasma therapy on COVID-19 patients. Based on the secondary data, we build a prediction model to evaluate and predict the trends in the clinical characteristics and laboratory findings for critically ill patients infected with COVID-19 and treated with convalescent plasma. Here, we use a combined prediction model utilizing three models; the grey prediction model (GM (1, 1)), the residual prediction model (residual GM (1, 1)), and a back propagation artificial neural network (BP-ANN) based residual sign prediction model. Also, a validation of the results of the study has been presented at two levels. On analysis of the results from the prediction model, it is observed that the convalescent plasma therapy can show progressive signs on COVID-19 infected patients. Health practitioners can understand, analyze, and predict the potential risks of convalescent plasma therapy based on the proposed model.