Objectives:
To depict overall psychological well-being of a large group of students of different universities in Ukraine three months after the emerge of the full-scale war.

Methods:
A total of 1,142 participants were asked to measure their psychological well-being on a 0–10 scale before and after the onset of full-scale war. Mental health symptoms were measured with questionnaires targeting depression (PHQ-9), anxiety (GAD-7), sleep problems (ISI), eating disorders (SCOFF), alcohol abuse (CAGE), and PTSD symptoms (PC-PTSD-5). To evaluate the connection between variables a χ2 was conducted. Phi and Cramer’s V coefficient were stated to demonstrate the power of the relationships. Additionally, machine learning (the XGBoost regression model) was used to build a predictive model for depressive symptoms.

Results:
Of all respondents, 66% screened positive for PTSD symptoms, 45% – moderate and severe anxiety symptoms, 47% – moderate and severe depressive symptoms. Regarding sleep, alcohol use and eating behavior, 19% of surveyed students had signs of moderate and severe insomnia, 15% reported alcohol abuse and 31% disordered eating. The severity of the aforementioned disorders varied depending on gender, year of study, social status, etc. According to the predictive model, lower initial psychological well-being, female gender, younger age, first years of study and any traumatic experience, including multiple trauma, predicted increases in depression score. Return to home after relocation was a protective factor.

Conclusions:
The study demonstrated the high prevalence of mental health symptoms among university students in Ukraine during the first months of the full-scale war. The psychological well-being pre-war was the strongest predictor of depressive symptoms in the model.