Background: Cross-sectional studies indicate a high prevalence of mental health problems among college students, but there are fewer longitudinal data on these problems and related help-seeking behavior. Methods: We conducted a baseline web-based survey of students attending a large public university in fall 2005 and a two-year follow-up survey in fall 2007. We used brief screening instruments to measure symptoms of mental disorders (anxiety, depression, eating disorders), as well as self-injury and suicidal ideation. We estimated the persistence of these mental health problems between the two time points, and determined to what extent students with mental health problems perceived a need for or used mental health services (medication or therapy). We conducted logistic regression analyses examining how baseline predictors were associated with mental health and help-seeking two years later. Results: Over half of students suffered from at least one mental health problem at baseline or follow-up. Among students with at least one mental health problem at baseline, 60% had at least one mental health problem two years later. Among students with a mental health problem at both time points, fewer than half received treatment between those time points. Limitations: Mental health problems are based on self-report to brief screens, and the sample is from a single university. Conclusions: These findings indicate that mental disorders are prevalent and persistent in a student population. While the majority of students with probable disorders are aware of the need for treatment, most of these students do not receive treatment, even over a two-year period. Published by Elsevier B.V.