Of the 713 809 members of the cohort (median age at the beginning of the observation period 16.1 [IQR, 15.9-16.4] years; 50.4% were male), 47 433 individuals received a diagnosis of mental disorder up to the ninth grade. Of the remaining 666 376 members of the cohort, 167 227 individuals (25.1%) received a diagnosis of mental disorder during the observation period (7.3 million person-years). A dose-response relationship was found, with no significant increase in subsequent risk in 1 diagnosed classmate (HR, 1.01; 95% CI, 1.00-1.02), but a 5% increase in more than 1 diagnosed classmate (HR, 1.05; 95% CI, 1.04-1.06). Risk was not proportional over time, but was highest in the first year of follow-up, showing a 9% increase for one diagnosed classmate (HR, 1.09; 95% CI, 1.04-1.14) and an 18% increase for more than one diagnosed classmate (HR, 1.18; 95% CI, 1.13-1.24). Of the psychiatric disorders studied, the highest risks were for mood, anxiety and eating disorders. The increased risk was observed after accounting for a number of parental, school and regional confounders.