Evaluating the comparative effectiveness of promising treatment programs for adolescents in face of differential follow-up
*Beth Ann Griffin, RAND Corporation 

Keywords: lost to follow-up, differential follow-up, sensitivity analyses, propensity score weighting, comparative effectiveness

The purpose of this presentation is to evaluate the comparative effectiveness of an evidence-supported adolescent substance abuse treatment program (MET/CBT-5) as delivered by community-based treatment providers relative to its performance in a controlled research trial. Nonparametric propensity score weights were used to control for pre-treatment differences between two groups of clients receiving the treatment and findings indicated that clients receiving treatment in community settings had similar or improved outcomes compared to the clients receiving treatment in the research trial. However, follow-up rates in the community based settings were extremely low (40%) compared to the experimental setting (98%) and thus we developed a new tool for assessing the sensitivity of our findings to differential follow-up. Sensitivity analyses indicated that the relationship between loss-to-follow-up and negative outcomes among those receiving treatment would have to be strong to account for the observed benefit of receiving MET/CBT-5 in community-based settings versus experimental settings.