COVID-19 pandemic’s effect on the mental health among the Tunisian general population: Associated factors mining via machine learning

Abstract

The emergence of COVID-19 pandemic has caused a brutal change in the lifestyle of citizens all around the world and greatly affected the mental health of individuals. In Tunisia, many psychological problems have been triggered during the first peak of the pandemic like anxiety, depression, sleep disturbances, and suicide risk. To overcome such disorders, it is crucial to identify the main factors leading to the mental disorders and then develop preventive strategies if a novel form of pandemic or a traumatic event appears. This paper proposes a novel association rules-based approach to characterize the profiles of citizens highly vulnerable to psychological disorders when confronted to traumatic events. The aim of this work is to use machine learning techniques in order to identify the major factors as well as the clusters of features leading to several psychiatric disorders during the COVID-19 pandemic in Tunisia. Many stressors were found to be associated with some psychiatric disorders. The stronger associations were found between doctor consultation and anxiety, COVID test and depression, quarantine and insomnia, and direct contact with a suspected case and peritraumatic distress and dissociation. In addition, it has been found that some factors, like female gender and regular worker, are not leading to mental disorders when they are treated alone, however, they present a high influence on the mental health when they are associated with other factors. For instance, this work discovered that women who have psychiatric history and who always drink coffee are exposed to depression during the pandemic. Other profile of citizens who are highly vulnerable to peritraumatic dissociation concerns students who are confined and who have recent symptoms. The characterization of such vulnerable profiles can provide considerable decision support for medical staff.

Keywords

• COVID-19
• Pandemic
• Mental health
• Machine learning
• Association rules mining