Longitudinal treatment effects among cocaine users: a growth curve modeling approach.

This study examined longitudinal treatment effects among cocaine users. The study examined a sample of 371 cocaine users screened from arrestees in jails and from patients in sexually transmitted disease clinics and emergency rooms-all in Los Angeles County during 1992-1994. Of the 371 subjects, 121 had never been in treatment and 250 reported a history of participation in drug user treatment (145 subjects' first treatment was for cocaine use and 105 were in treatment for a drug other than cocaine). Data were collected during face-to-face interviews using a natural history interview instrument. We applied a series of growth curve models to investigate treatment effects on cocaine use. For those who had been in treatment for cocaine use, use of cocaine decreased from approximately 70% before treatment to 12% after treatment entry, while no such changes were observed among those who had never been in treatment or those in treatment for other drugs. Relative to nontreated users, cocaine-treated participants showed a greater likelihood of pretreatment use for both initial status (OR = 3.58) and growth rate (OR = 1.05). After treatment entry, cocaine-treated participants as compared to nontreated participants had a lower likelihood of use (OR = 0.27), although their cocaine use after the initial status increased at a greater rate (OR = 1.03). Treated users were five times less likely to use when they were in treatment than when they were out of treatment. Longer treatment retention was related to initially reduced use but not to later rates of change in cocaine use. The study findings support that treatment for cocaine use is effective in reducing cocaine use. Longitudinal models provide opportunities to demonstrate the dynamic relationships between treatment and outcome.