There were 434 students who completed the questionnaire. SARS-CoV-2 infection was confirmed in 6 respondents (1.38%) and these were excluded from further statistical data processing.
The sample included a statistically relevant difference in the number of female respondents (335, i.e.78.27%) in comparison to male (p < 0.001). The average age of the respondents was 23.81 ±5.25 (range 19−25). The statistically relevant majority, 324 (75.50%) lives with their parents.
Table 1. shows the structure of the respondents by gender, age and the place of residence.
Significantly fewer respondents 71 (16.59%) reported feelings of concern and fear of contracting COVID-19, significantly more respondents 263 (61.45%) had concerns for the health of family members, whereas significantly fewer respondents, 49 (11.45%), participated in any volunteer activities (p < 0.001).
Table 3. shows the values of the descriptive parameters for the PSS-10, CSI and GHQ-28 scores among the non-infected respondents (n = 428).
Based on the average values on the PSS-10 of20.37±7.62 (20.00) higher than 20, it is evident that the average level of perceived stress is high because it is close to the value of 20, which defines it as such based on this scale(Table 3).
Based on the average values of the CSI scores, the value of 25.76±4.61 (26.00) on Problem solving is within the expected average of 26, Seeking social support, 22.15±5.01 (22.00), is slightly lower than the expected average of 23.00, whereas the score on Avoidance of 23.78±4.20 (24.00) is significantly above the expected average of 19 (Table 3).
The findings on the GHQ-28 identified that 48.83% of students scored higher or equal to 24 (Table 3). On the GHQ-28, the highest average values were on the Anxiety/insomnia subscale, with the average value of 8.03±5.91 (7.00), closely followed by the Social dysfunction subscale with 7.99±3.72 (7.00). The score on Somatic complaints was significantly lower, 6.56±4.75 (6.00), while the least pronounced was the Depression subscale, with 3.69±5.04 (1.00)(Table 3).
The Student's t-test (For Independent Samples) revealed a significantly higher PPS in female respondents (p < 0.01), respondents who expressed a concern for the health of family members (p < 0,01) and in students who did not participate in any volunteer activities (p < 0.001)(Table 4)
The Spearman's rank correlation coefficient showed a statistically relevant negative and low correlation of the PSS-10 values with Problem Solving (ρ = -0.21, p < 0.001) as well as a positive and high statistically relevant correlation of PSS with Avoidance (ρ = 0.50, p < 0.001) were found. There was no relevant correlation between the PSS values and Seeking social support (ρ = -0.001, p > 0.05)
The Spearman's rank correlation coefficient showed a statistically relevant positive and extremely high correlation of the PSS-10 values with Somatic complaints (ρ = 0.643, p < 0.001) and Depression (ρ = 0.645, p < 0.001), and an even higher positive correlation with Anxiety/insomnia (ρ = 0.763, p < 0.001).
The results of the univariate linear regression analysis showed that the higher PSS-10 values were significantly influenced by the overall GHQ-28 scores higher or equal to 24, the overall GHQ-28 questionnaire scores alone, and the scores on its three subscales (somatic symptoms, anxiety/insomnia and depression) as well as the Avoidance subscale on the CSI scale, concern for the health of family members, and the female gender (Table 5)
The overall GHQ-28 questionnaire score higher or equal to 24 increased the PSS-10 score for 8.71 (7.521 − 9.898, p < 0.001) in comparison to the respondents with an overall GHQ-28 up to 23. This has the highest influence on the PPS-10 (Table 5).
Female respondents had a 2.659 (0.921 − 4.397, p < 0.01) higher PSS scores in comparison to the male ones, whereas the ones with a concern for the health of family members had a 2.459 (0.989 − 3.929, p < 0.01) higher PPS scores in comparison to the respondents who expressed no concern for the health of family members (Table 5)
A unit increase on the Somatic complains subscale score led to an increase in the PSS score by 0.983 (0.862 − 1.104, p < 0.001); a unit increase on the Anxiety/insomnia subscale score led to an increase in the PSS score by 0.959 (0.877 − 1.040, p < 0.001), and on Depression by 0.871 (0.754 − 0.989, p < 0.001). A unit increase on the Avoidance subscale led to an increase in the PSS score by 0.880 (0.729 − 1.031, p < 0.001) (Table 5).
A unit increase of the overall GHQ-28 score led to an increase in the PSS score by 0.355 (0.319 − 0.392, p < 0.001)(Table 5)
Volunteering – helping thedisabled persons during the COVID-19 pandemic had a statistically relevant influence on lowering the PSS scores by 4.038 (-1.795 − -6.280, p < 0.001) compared with the non-volunteers. Apart from volunteering, a statistically relevant influence on lowering the PSS scores had the sub-score Problem-Solving on the CSI questionnaire (its unit increase lowers the PSS score by 0.393 (-0.240 − -0,546, p < 0.001)) and Social dysfunction, sub-score on the GHQ-28 scale, whose unit increase lowered the PSS-10 score by 0.212 (-0.018 − -0.406, p < 0.05)(Table 5).
The initial model of multivariate linear regression analysis was formed based on the variables that were shown in the univariate analysis as factors with a significant influence on the PSS scores. By applying the stepwise regression in step 5, an optimal model of the combined influence of the variables on the PSS-10 score was obtained, which consisted of anxiety/insomnia, depression, gender, avoidance and social dysfunction (Table 6). The multiple-correlation coefficient R is 0.783, and the multiple-determination coefficient is 0.613, which means that in the 61.3% of the tested sample, the PSS-10 score variance was determined by the variance of the set of predictor variables found in the final model. The female gender showed to be the most significant factor influencing the increase in the PSS score, followed by anxiety/ insomnia, depression and avoidance, whereas social dysfunction showed to decrease the score we examined(Table 6).