Assessing Concurrent Effects of Substance Abuse Treatment Modalities Using Marginal Structural Models

• Activity Number: 293
• Type: Topic Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Health Policy Statistics Section
• Author(s): Beth Ann Griffin and Rajeev Ramchand*+ and Cha-Chi Fan and Daniel Almirall and Daniel F. McCaffrey
• Companies: RAND Corporation and RAND Corporation and RAND Corporation and University of Michigan and RAND Corporation
• Keywords: causal inference ; marginal structural models ; inverse probability of treatment weighting ; substance abuse

Abstract

A compelling case can be made that, for some individuals, substance use disorders may be akin to chronic conditions in which effective treatments may need to help patients when they are actually engaged in treatment (concurrent treatment effect). We develop methods to estimate concurrent effects from longitudinal data building on Marginal Structural Models (MSMs) with Inverse Probability of Treatment Weighting (IPTW). MSMs with IPTW provide a basis for posing scientific questions concerning the causal effects of time-varying treatments. But there is little literature on how best to estimate the IPTW weights and build models of the treatment assignment process necessary for estimating weights. We discuss the estimation of concurrent treatment effect from a sample of over 2000 youth and describe methods for modeling treatment assignments and assessing the quality of the resulting weights.