Background: During the COVID-19-pandemic, Swedish higher education institutions shifted to remote teaching to reduce contagion. Among the students, this shift may have involved mental health changes. Such changes in mental health may, in turn, influence academic success.

Purpose: This longitudinal study investigated the effects of mental health changes on academic self-efficacy among university students in Sweden over the course of the pandemic.

Method: Self-reports were collected through online questionnaires. Baseline questionnaires were distributed in May 2020 with follow-ups at 5-months and 10-months post baseline. The longitudinal sample included 2796 students (89,5% of baseline respondents). Taking a Bayesian approach, multilevel multinomial regressions were used to estimate effects, with adaptive intercepts for universities as well as respondents in the longitudinal models.

Results: Reporting worse mental health at baseline, rather than no change, resulted in an observed higher odds of reporting worse academic self-efficacy rather than no change in academic self-efficacy at both follow-ups. Also, reporting a change to worse mental health, rather than no change, at the 5-month follow-up resulted in observed higher odds of reporting worse academic self-efficacy at the 10-month follow-up. A similar pattern was found for students reporting both positive and negative mental health changes, as compared to those reporting no mental health changes.

Conclusions: These findings have implications for the development of easily accessible services that support mental health in order for students, particularly those who experience negative effects on their mental health, to keep up their studies in times of uncertainty.