Estimating Causal Effects When Treatment Is Modeled As a Latent Variable and an Application to Adolescent Drug Treatment

Introduction. There are many treatment options for adolescent substance abuse (e.g., adolescent-centered therapies, family-based therapy, case-management, biological drug testing), any or all of which might be included in a specific intervention protocol. Typical effectiveness evaluations assess outcomes for a specific intervention protocol comprised of discrete treatment components. Yet in practice, adolescents may actually receive significantly more or fewer services than those officially associated with a particular program due to concurrent program enrollment, noncompliance, or lack of program fidelity. We present an application of latent class causal analysis (Schafer & Kang, 2010) to adolescent substance abuse treatment in order to determine the effectiveness of common “clusters” of substance abuse treatment components.

Method. Data are from a multisite, longitudinal observational study of adolescents (ages 12-18) enrolled in various substance abuse treatment programs as part of SAMHSA CSAT funding. This study is restricted to youth reporting only outpatient drug treatment services between baseline and 3-month study visits (N = 4854). The Global Assessment of Individual Needs was the primary study instrument; drug treatment services received were assessed with the Treatment Received Scale and substance use outcomes were assessed with the Substance Frequency Scale and Substance Problems Scale. A propensity score model was fit to equate the different treatment classes on numerous variables including demographics, baseline substance use, and juvenile justice system involvement.  Latent class analysis was performed to identify latent treatment classes; latent class causal analysis was performed to assess the effect of treatment class membership on substance use outcomes.

Results. Results indicate that a 4-class model is optimal. Treatment classes included: High Adolescent, Family, and Case Management services (12%); High Adolescent and Family services (12%); High Adolescent services (39%); and Low Adolescent services (37%). These subgroups represent varying levels of treatment intensity and comprehensiveness, ranging from only adolescent-centered services to a combination of adolescent-centered, family-based and case management services.

Conclusion. Differences in substance abuse outcomes across latent treatment classes will be discussed. Overall, this study provides an illustration of latent class causal analysis for prevention researchers.