Mental disorders are among the leading causes of years lost due to disability in young people globally. Students are no exception, both mental disorders and mental health problems are common among university students with dramatic increases reported in recent years. The aetiology of mental disorders and mental health problems is generally assumed to be multicausal, with factors at different levels contributing to their development. Given the complexity of the causal network underlying mental health problems, it has been argued that a clear causal framework is needed when studying the aetiology of mental health problems. This thesis aimed to investigate some aspects of the potential aetiology of mental health problems among university students. Specifically, it focuses on four exposures at the psychosocial level presented in four studies: 1) the coronavirus disease (COVID-19) pandemic, 2) poor sleep quality, 3) procrastination and 4) sexual harassment and sexual violence. In addition to the discussion provided in each respective paper, this thesis discusses limitations and possible interpretations of our results from a modern causal inference perspective.

The four studies of this thesis are based on The Sustainable University Life (SUN) cohort. The SUN cohort followed 4262 university students from eight universities in and around Stockholm, Sweden, with web-surveys at five time-points over one year.

In Study I, we aimed to determine the mean trajectories of depression, anxiety, and stress symptoms among university students in Stockholm before and during the first months of the COVID-19 pandemic. For this, we included a subsample of 1836 university students that entered the SUN cohort before the outbreak of the COVID-19 pandemic, and were followed during the months before the pandemic, during the first wave of the pandemic and in the summer, months following the first wave of the pandemic. We found that mean depression, anxiety, and stress symptom levels were largely stable during the first wave compared to the months before the pandemic and decreased slightly during the following summer months. Our results indicate that mean levels of mental health symptoms did not change much during the early phase of the pandemic compared to before the pandemic.

In Study II, we aimed to determine whether sleep quality statistically interacts cross-sectionally with loneliness, risky alcohol use, perfectionistic concerns and/or physical inactivity in relation to depressive symptoms in university students. We conducted a cross-sectional study using baseline-data from all 4262 participants in the SUN cohort. We found that while all factors were associated with depressive symptoms, only perfectionistic concerns interacted with sleep quality in its relation to depression. This interaction was quite weak and explainedonly a small proportion of the variance in depressive symptoms. Overall, we did not find support for our hypothesis that poor sleep quality could interact with several different potential risk factors for depressive symptoms.

In Study III, we aimed to evaluate the associations between procrastination and sixteen subsequent health outcomes (including mental health symptoms, disabling pain, lifestyle behaviours and psychosocial health factors), measured 9 months later, among university students. We used data from all participants responding to the first follow-up in the SUN cohort (n=3525) and found that procrastination was related to several subsequent health outcomes, including symptoms of depression, anxiety and stress, while controlling for multiple potential confounders. Although we cannot rule out non-causal explanations for these associations, the results indicate that procrastination could have an effect on health outcomes among students, but that it is likely to be rather small for any specific health outcome.

In Study IV, we investigated the impact of recent exposure to different forms of sexual harassment and sexual violence; 1) unwanted sexual attention, 2) offensive sexual remarks, 3) presentation or distribution of sexist material, 4) uncomfortable touching, 5) being offered benefits for sex and 6) sex against ones will, along with a wide definition of sexual harassment: sexual harassment (wide subjective definition) on levels of depression and anxiety symptoms three, six and nine months later, for women and men, respectively. We conducted a cohort study using data from all women and men responding to the first follow-up in the SUNcohort (n= 3503). Our results showed that women recently exposed to 1) sexual harassment (wide definition), 2) unwanted sexual attention and 3) sex against ones will showed higher subsequent levels of depression and/or anxiety symptoms. The general trend was that all exposures were related to higher symptom levels at three months, but that this difference between exposed and unexposed diminished over time, although these trends are uncertain with wide confidence intervals. The exception was exposure to sex against ones will, where exposed showed elevated symptom levels throughout the follow-up period. For men, the estimates were uncertain overall, and we refrain from interpreting these results. Our results indicate that recent exposure to different forms of sexual harassment and sexual violence may impact later depression and anxiety symptoms among women, and that there could be differences in the strength and long-term impact on mental health between different forms of sexual harassment and sexual violence. Again, we cannot, with certainty rule out non-causal reasons for these associations.

Interpreting any of these results as causal effects rests on multiple assumptions, which are discussed in the thesis. Determination of causal effects preferably relies on triangulation of results from different studies with different methodology. Overall, however, I believe that this thesis has strengthened the evidence that procrastination and sexual harassment and sexual violence may be causes of mental health problems among university students. An equally important finding, is that mental health seemed rather stable during COVID-19 pandemic, indicating that the pandemic may not have caused increased mental health problems on the group level.