Objectives: Although there are a number of studies about substance use by medical students, the majority are still cross-sectional. We aimed to investigate prevalence, 2-year incidence, and factors associated with substance use during medical training. Methods: This longitudinal study included medical students in 4 different waves (with each wave equaling 1 semester). Socio-demographic data, Duke Religion Index (DUREL), DASS-21, and the "Alcohol, Smoking and Substance Involvement Screening Test" (ASSIST) were used. Results: A total of 327 (56.2%) medical students were followed for 2 years. Prevalence of lifetime substance use was 89.9% for alcohol, 34.5% for cannabis, and 17.1% for sedatives. Tobacco had the greatest incidence of use over the 2 years (16.4%), followed by alcohol (13.8%) and cannabis (13.8%). At least 24% of the students would need an intervention for alcohol use, 11.4% for tobacco, and 6.5% for cannabis. Alcohol use during wave 4 was associated with organizational religiosity, and alcohol and tobacco use at the baseline; tobacco use during wave 4 was associated with age, non-organizational religiosity, and cannabis and tobacco use at the baseline; and cannabis use during wave 4 was associated with cannabis and tobacco use at the baseline. Conclusions: Our results indicate an increase in the incidence of lifetime substance use during medical training. Prior use of substance was associated with a higher chance of use after 2 years, while age and religiousness seem to reduce the chance of use.