African American Adolescent Substance Abuse Treatment: A Multi-Level Analysis Of Selected Socio-Demographic, Client And Treatment Variables That Influence Premature Termination

Abstract

Adolescent substance abuse treatment has continually demonstrated effectiveness within the existing literature. Consequently, individuals who prematurely terminate substance abuse services have demonstrated less favorable chances of experiencing the positive outcomes that have been associated with treatment retention. Despite the well documented positive treatment outcomes, the overall rates of premature termination continue to remain at significantly high levels especially among the African American population. Researcher have advocated for an increase in research that examines treatment related factors and needs of minority substance abusing adolescents (Longshore, Grills, Annon, & Grady, 1998; Wagner, 2003). Using Andersen's (1995) behavioral model of health care use, the current study investigated selected client and treatment related variables which influenced the degree to which African American adolescents complete or do not complete (premature termination non-completers) non-detoxification ambulatory outpatient substance abuse treatment. This study was conducted using a modified multilevel modeling design based on quantitative inferential statistics. Four levels of inferential statistical analysis were conducted: (1) discriminant function model development, (2) alternative confirmation, (3) extended cross validation of the discriminant function model, and (4) post hoc analysis. Research samples were obtained retrospectively from the 2008 Treatment Episode Dataset using a systematic stratified random sampling procedure. Statistical findings indicate that length of stay, substance use severity, education and age are significant variables which resulted in final discriminant model development and can be used to predict treatment completion status among African American adolescents in outpatient substance abuse severity. This finding was confirmed using a two-group K-Means cluster analysis. Moreover, through an extended cross validation approach, the discriminant function model consistently demonstrated effectiveness at accurately classifying treatment completion status to a significant degree. The post-hoc analysis concluded that differences between rapid, median and near-term premature termination non-completers were not significant using the .05 alpha, however results were approaching significance. The clinical implications for policy, treatment, and social work practice were discussed in relationship to substance abuse treatment strategies for prevention and early intervention of premature termination from substance abuse treatment among the African American adolescent population. Additionally future recommendations for research within the area of premature termination are provided.