Predictive validity of cocaine, sedative, and alcohol dependence diagnoses.

This study examined the predictive validity of Structured Clinical Interview for DSM-III-R (Spitzer, Williams, Gibbon, & First, 1990) based substance dependence diagnoses (i.e., cocaine, sedative, and alcohol) for 518 opioid-dependent outpatients entering methadone maintenance. Patients were followed over 1 year of treatment, which involved daily methadone substitution supplemented by individual and group counseling. Urine specimens were tested randomly 1-4 times per month. Patients diagnosed with current cocaine, sedative, or alcohol dependence were more likely to use these drugs than were patients with past only or no dependence syndrome. Current cocaine dependence predicted early treatment dropout. The results demonstrate the predictive and discriminant validity of several substance dependence diagnoses common among patients in substance abuse or other psychiatric treatment settings.