3.2 Additional effectiveness analyses
To explore the relations between demographic and clinical characteristics and the observed clinical effectiveness of agomelatine on depression we performed several linear regression analyses. The following characteristics were selected as covariates of interest: age, sex, severity of depressive disorder at baseline according to HAMD-17 total score, COVID-19 severity, and time passed since COVID 19 infection onset. Univariate regression analysis showed that depression severity at baseline (regression coefficient: -0.741; P<0.0001), moderate severity COVID-19 (regression coefficient: -3.395; P=0.0035), and time since onset of infection (regression coefficient: -2.395; P=0.0006) were the parameters with a statistically significant relationship with the observed decrease in HAMD-17 total score at the end of follow up (week 8 visit). All coefficients were negative indicating that the relations between these parameters and the observed improvement in depression severity were reciprocal. However, after multivariate regression analysis the only parameter remaining significantly related to improvement in depression severity at week 8 was baseline depression severity expressed as total HAMD-17 score (regression coefficient: -0.70; P<0.0001). When univariate regression analyses were conducted to explore relations between the covariates of interest and the effectiveness of agomelatine at reducing anxiety severity and improving patients’ QoL, mean total HAMD-17 score at baseline was again the only parameter with a significant relation with the observed improvement in both anxiety and patients’ QoL (Table 2 ).