ENGLISH :
While addiction is increasingly recognized as a chronic condition, little information is available on the actual duration of addiction and treatment careers. The purpose of this study was to estimate the duration and correlates of years between (a) first use and at least a year of abstinence and (b) first treatment admission and at least one year of abstinence. Data are from 1,271 of 1,326 (96%) people recruited from a stratified sequential sample of admissions to publicly funded treatment programs in a large metropolitan area. Participants were interviewed at 6, 18, 24, 36, 48, and 60 months post-intake. With an average age at the referent intake of 35 (SD=8) years, the sample is 59% female, 87% African American, 7% Hispanic, and 5% white. The most common dependence diagnoses were for cocaine (64%), alcohol (44%), opioids (41%), and/or marijuana (14%). Using lifetime substance use and treatment histories collected at intake and subsequent treatment utilization recorded during follow-up interviews, we conducted survival analysis to estimate the time from first use and first treatment until people reported 12 months of abstinence or died (with people still using, in treatment, or dead at the last follow-up treated as right censored). During the three years after intake, 47% reached at least 12 months of abstinence. The median time from first to last use was 27 years. The median time from first treatment episode to last use was 9 years. Years to recovery were significantly longer for males, people starting use under the age of 21 (particularly those starting under the age of 15), people who had participated in treatment 3 or more times, and for people high in mental distress. The exploratory results suggest that multiple episodes of care over several years are the norm and that rather than thinking of multiple episodes in terms of "cumulative dosage," it might be better thought of as further evidence of chronicity and that we need to develop and evaluate models of longer term recovery management. (Editor's abstract.)