Bridging Disciplinary Boundaries (January 11 - 14, 2007)

The purpose of this study was to document differences between adolescent participants who either completed or dropped out of a NIAAA-sponsored randomized, clinical trial of brief motivational intervention conducted with an ethnically diverse sample of substance abusing adolescents. Specific treatment process variables (therapeutic alliance (Bordin, 1975), working alliance (Bordin, 1976; Greenson, 1967; Horvath & Greenberg, 1989), and patient involvement (Garfield, 1994)] were used to document differences between treatment completers and dropouts in the context of Guided Self-Change (GSC; Sobell & Sobell, 1993, 1995) therapy sessions. Specifically, these variables were used to describe significant differences in psychotherapy process ratings in a homogenized sample of Latino males, ages 14-20, referred for alcohol or other drug (AOD) treatment who either (a) completed Alcohol Treatment Targeting Adolescents In Need (ATTAIN) (n=29) or (b) dropped out (n=29) after the first or second GSC therapy session.

Fifteen-minute segments were copied from the midpoint of previously recorded audio-tapes of the manualized GSC therapy sessions. The Working Alliance Inventory-Observer (WAI-O) (á=.98) and the Vanderbilt Psychotherapy Process Scale-Short Form (VPPS-S) (á=.82) were selected for therapeutic process ratings based on their robust psychometric properties, and their use in previously published studies (e.g., Tichenor & Hill, 1989; Smith, Hilsenroth, Baity, & Knowles, 2003). Raters were trained to a criterion level of interrater reliability for both the WAI-O (r=.82) and VPPS-S (r=.74), and among the two raters who reached criterion levels, one rater was blind to the hypotheses of the study.

Results were derived from MANOVA analyses for WAI-O and VPPS-S scores. Treatment completers reported significantly higher mean scores for the WAI-O across the three subscales (i.e., for Goals, Tasks, and Bonds) (p < .001) than adolescents who dropped out of treatment. In addition, treatment completers reported significantly higher alliance scores than dropouts (p < .001) for VPPS-S subscales (i.e., for Therapist Exploration, Therapist Warmth and Friendliness, Patient Participation, and Patient Dependency), but significantly lower scores than dropouts (p < .001) for the VPPS-S Patient Psychic Distress subscale. There was no significant difference between the groups for mean scores for the VPPS-S Negative Relationship subscale.

The study documents that participants who completed treatment exhibited higher therapeutic alliance scores, but that therapists were not differentially negative to either group of clients, regardless of completion status. The implications of the study inform an understudied treatment completion-dropout literature, and are relevant to the delivery of brief motivational interventions (BMIs) in community settings to minority adolescent populations.