Abstract. We present a stochastic agent-based model of the COVID-19 pandemic that links the spread of a respiratory viral infection with the dynamics of mental disorders arising under prolonged stress. The model is designed for urban populations. In this study, it is parameterized and calibrated using demographic, social, medical, and epidemiological data for Moscow. The model accounts for heterogeneity in age, contact structure, place of residence, school or workplace, chronic non-communicable diseases, susceptibility to infection, and vulnerability to psychological stress. Transmission of SARS-CoV-2 is simulated through contacts within households, educational groups, and workplaces, with differences between viral variants taken into account. Depressive and anxiety disorders are described as outcomes of the combined effect of individual vulnerability and stress exposure associated with reduced mobility and COVID-19 cases among household members. Health losses due to infection and mental disorders are compared using quality-adjusted life years. For mental disorders, the assessment includes both direct losses caused by reduced quality of life during episodes of depression and anxiety and indirect losses associated with an increased risk of all-cause mortality. The simulations show that, under this approach, the contribution of mental disorders to total population health losses may be comparable with that of the infection itself. Stronger quarantine restrictions may reduce infection-related losses, but this effect can be offset by additional losses related to stress and mental disorders. These results indicate that the evaluation of epidemic-control strategies should include not only infections, hospitalizations, and deaths from COVID-19, but also the broader health consequences of prolonged psychological stress.

 

Key words: agent-based model, COVID-19, depression, anxiety, QALY