Evidence for the effectiveness of community-based psychosocial treatments for adolescent substance use disorders is mixed, at best. Researchers'inability to detect strong or replicable treatment effects may result from their focus on the post-treatment effects of a single incident treatment episode. Recent conceptualizations of addiction and treatment suggest that these approaches which compare pre-post treatment effects may obscure some of treatment's most salient effects. For one, treatment effects should be expected to be greatest during treatment or concurrently. Second, from a treatment careers perspective, multiple treatment episodes over time are likely to lead to cumulative effects, which would be greater than those observed for any individual treatment episode. Finally, considerations of client heterogeneity suggest that effects of treatment may be greatest for a subgroup of patients (moderated effects), with such effects being obscured when combined with the smaller effects expected for other patients. With this renewal we propose to estimate the causal effects of treatment on adolescent outcomes by examining these three types of treatment effects which addiction theory suggests have an important role, but which have yet to be satisfactorily measured. To do so, we have assembled a large set of adolescent treatment outcomes data collected by RAND, the Center on Substance Abuse Treatment, and Chestnut Health Systems and which include background information, treatment outcomes and treatment histories for more than 10,000 adolescent treatment admissions. Using this rich data and a powerful new casual modeling technique for time-varying treatments, marginal structural models with inverse probability of treatment weighting, we propose four new aims: (1) Estimate the treatment effects of different levels of care on drug use and other outcomes observed while youths remain in treatment;(2) Estimate the causal effect of cumulative treatment experiences of different levels of care on 1-, 2-, and 6-year drug use and other outcomes;(3) Estimate how baseline and time-varying client characteristics moderate the concurrent and cumulative effects of level of care on recovery and substance use outcomes;and (4) Develop and evaluate statistical methods required for Aims 1 to 3, and compare findings to those produced using conventional treatment research methods.
The goal of this project is to determine causal effects of treatment services for adolescents on drug use and other outcomes using powerful, new causal modeling techniques that allow us to study the concurrent effects of treatment, the cumulative effects of multiple treatment episodes, and differential effects for subgroups of clients. To accomplish this objective, we will use a large set of adolescent treatment outcomes data that has been collected by RAND, the Center on Substance Abuse Treatment, and Chestnut Health Systems, which includes treatment outcomes and treatment histories for more than 15,000 adolescent treatment admissions. By pursuing this research, we will improve the measurement and understanding of the effects of treatment and develop and disseminate more relevant and robust causal modeling approaches to substance abuse treatment services research than are currently standard in the field, thereby addressing key goals set by the 2004 Blue Ribbon Commission Task Force on Health Services Research at the National Institute on Drug Abuse.
|Schuler, Megan S; Griffin, Beth Ann; Ramchand, Rajeev et al. (2014) Effectiveness of treatment for adolescent substance use: is biological drug testing sufficient? J Stud Alcohol Drugs 75:358-70|
|Almirall, Daniel; Griffin, Beth Ann; McCaffrey, Daniel F et al. (2014) Time-varying effect moderation using the structural nested mean model: estimation using inverse-weighted regression with residuals. Stat Med 33:3466-87|
|Ramchand, Rajeev; Griffin, Beth Ann; Slaughter, Mary Ellen et al. (2014) Do improvements in substance use and mental health symptoms during treatment translate to long-term outcomes in the opposite domain? J Subst Abuse Treat 47:339-46|
|Schuler, Megan S; Leoutsakos, Jeannie-Marie S; Stuart, Elizabeth A (2014) ADDRESSING CONFOUNDING WHEN ESTIMATING THE EFFECTS OF LATENT CLASSES ON A DISTAL OUTCOME. Health Serv Outcomes Res Methodol 14:232-254|
|Griffin, Beth Ann; Ramchand, Rajeev; Almirall, Daniel et al. (2014) Estimating the causal effects of cumulative treatment episodes for adolescents using marginal structural models and inverse probability of treatment weighting. Drug Alcohol Depend 136:69-78|
|Almirall, Daniel; McCaffrey, Daniel F; Ramchand, Rajeev et al. (2013) Subgroups analysis when treatment and moderators are time-varying. Prev Sci 14:169-78|
|McCaffrey, Daniel F; Lockwood, J R; Setodji, Claude M (2013) Inverse probability weighting with error-prone covariates. Biometrika 100:671-680|
|McCaffrey, Daniel F; Griffin, Beth Ann; Almirall, Daniel et al. (2013) A tutorial on propensity score estimation for multiple treatments using generalized boosted models. Stat Med 32:3388-414|
|Griffin, Beth Ann; McCaffrey, Daniel; Ramchand, Rajeev et al. (2012) Assessing the Sensitivity of Treatment Effect Estimates to Differential Follow-Up Rates: Implications for Translational Research. Health Serv Outcomes Res Methodol 12:84-103|
|Griffin, Beth Ann; McCaffery, Daniel F; Morral, Andrew R (2008) AN APPLICATION OF PRINCIPAL STRATIFICATION TO CONTROL FOR INSTITUTIONALIZATION AT FOLLOW-UP IN STUDIES OF SUBSTANCE ABUSE TREATMENT PROGRAMS. Ann Appl Stat 2:1034-1055|
Showing the most recent 10 out of 18 publications