Abstract: Adequate drug treatment for substance users continues to be a challenge for most U.S. cities. To address heroin addiction in Baltimore, the Baltimore Buprenorphine Initiative was implemented as a joint project to promote individualized, patient-centered buprenorphine therapy in conjunction with behavioral treatment to accelerate recovery from opioid addiction. The purpose of this analysis was to explore differences in recovery trajectories predicting length of stay and use this information to predict characteristics that influence an individual’s ability to remain in the Baltimore Buprenorphine Initiative program. The sample consisted of 1,039 subjects enrolled in the program between January 2008 and June 2009. The regression modeling determined that age, income, employment, and higher level of treatment were significant predictors of length of stay in the recovery program. The findings of this study have practical implications for the design and implementation of heroin addiction programs. The research indicates that focusing on these specific predictive variables early in the program design phase could increase recovery success rates as measured by length of stay.