Investigators in mental health research are often interested in examining critical events such as onset, relapse, and recovery from illness, including substance use disorders. As data on these critical events are often collected at discrete-time intervals (e.g., weekly, monthly, or yearly), discrete-time survival models are more appropriate than well-known continuous-time methods. In this paper, we present discrete-time survival analysis methods at an introductory level. Using data collected every 6 months from a 3-year study of assertive community treatment in New Hampshire, we show that discrete-time survival models can be used to analyze patterns of remission from substance use disorder among clients with severe mental illness. The main questions investigated are (1) when are remissions more likely to occur? and (2) what variables predict remission? The results indicate that remission is more likely to occur in the first 6 months and in the 3rd year of the study. Gender, age, baseline use of substances, and diagnosis are strong predictors of remission.