The emotional job-stress of COVID-19 on nurses working in isolation centres: a machine learning approach

Abstract

Motivated by the present global coronavirus pandemic with its multiple mental health and adverse psychosocial effects on frontline nurses, enormous research on the ecology and epidemiology of COVID-19 has been carried out. However, there is the sparsity of work done in the psychosocial context and the associated mental health impact of COVID-19. This article provides an overview of the scientific evidence and predicts the emotional health impact of the disease. Qualitative and quantitative data were collected from 543 frontline nurses working in quarantine facilities and were analyzed using machine learning platforms. Two different classifiers (Multinomial Naïve Bayes algorithm and Support Vector Machine) were compared with three human coders for text analysis. The Multinomial Naïve Bayes algorithm was inferior to the Support Vector Machine though both performed better in predicting emotions than humans in this study. The result suggests an increase in levels of fear, anxiety, worry and sadness during the periods of quarantine. Sadness is the most profound emotional impact. The studies promote machine learning to be used to predict social phenomenon. Impliedly, developing a better and more robust digital psychiatry intervention model mechanism to support existing psychological first aid will positively impact on the emotional and mental health disorders.