|
Activity Number:
|
220
- What Does it Mean to Work on Health Policy?
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, July 30, 2018 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Health Policy Statistics Section
|
|
Abstract #326681
|
|
|
Title:
|
Improving Our Understanding of Dynamic Group-Based Interventions for Treating Depression and Substance Use Disorders
|
|
Author(s):
|
Susan M Paddock* and Bing Han and Lane Burgette
|
|
Companies:
|
RAND Corporation and RAND Corporation and RAND Corporation
|
|
Keywords:
|
open enrollment group;
substance abuse treatment;
dynamic groups;
causal inference;
interference
|
|
Abstract:
|
Group cognitive behavioral therapy (GCBT) for residential substance abuse treatment (SAT) patients with co-occurring depression and substance use disorders is associated with improved patient outcomes. Additional policy-relevant questions remain about non-randomized GCBT factors, such as number of sessions attended. Therapy groups in SAT settings are typically dynamic, with membership changing over time. We extend prior approaches to causal inference from observational studies of non-dynamic groups to the dynamic group setting. Our approach assumes interference may be present - i.e., treatment received by one individual may affect potential outcomes of others. As the networked structure of dynamic group participants implies arbitrary constraints would be required to fit the propensity score model, we employ a prognostic score approach after assuming strong ignorability within covariate levels. Development of such an analytic framework will improve the design and analysis of future group-based SAT interventions and lead to a better understanding of such interventions.
|
Authors who are presenting talks have a * after their name.
