Pre-Conference Workshops - October 14, 2015
Workshop 1: 1:00pm – 5:30pm
Identifying Effective Treatment and Health Services in Substance Abuse Research Using Observational Data
Beth Ann Griffin, RAND Corporation
Donna Coffman, The Methodology Center, Pennsylvania State University
Megan Schuler, Harvard Medical School
Although randomized trials are the gold standard, there are many important substance use health services and policy questions that can be addressed using observational data. Drawing unbiased inferences from observational data relies on the use of appropriate statistical methods, such as causal inference methods, to account for the non-randomized design. In addition to understanding how effective a particular health service or treatment is, researchers, providers and stakeholders also want to answer pertinent questions such as “for whom is treatment most effective?” and “what are the mechanisms/pathways through which treatment works?” To address these questions with observational data, relevant statistical methods (e.g., mediation analysis) must be applied in a causal inference framework.
This workshop aims to promote the use of causal inference methods to improve substance use researchers’ ability to conduct robust analyses of observational data. We first describe the motivation for and an overview of causal inference methods, particularly focusing on propensity score techniques. We then highlight three recent advances in causal inference methodology, specifically in the areas of mediation analysis, latent class modeling, and time-varying treatments, each with substance use applications. First, Dr. Coffman will present a causal mediation analysis identifying mediation pathways in a substance use prevention program. Second, Dr. Schuler will discuss a method to integrate latent class analyses and causal inference and, finally, Dr. Griffin will describe methods for assessing the causal impact of time-varying treatments using marginal structural models and inverse probability of treatment weights. All instructors will provide step-by-step guidelines for implementation of the methods for attendees. The field must adopt more robust statistical methods in order to benefit from the wealth of knowledge available in observational data. Attendees should have some familiarity with running regression models but advanced statistical expertise is not required to use the techniques described in this short course.
Introduction to causal inference and propensity score methods for binary, multinomial and continuous treatments - Drs. Griffin, Schuler, and Coffman (1 hour)
Integrating propensity score methods with latent class models – Dr. Schuler (1 hour)
(Break - 15-30 minutes)
Estimating the causal effects of cumulative treatment episodes using marginal structural models and inverse probability of treatment weighting – Dr. Griffin (1 hour)
Introduction to causal mediation analyses – Dr. Coffman (1 hour)
Workshop 2: 1:00pm – 3:00pm
Using Qualitative Methods to Study Behavioral Health Interventions and Services in Diverse Settings
Shannon Gwin Mitchell, Ph.D., Friends Research Institute
Cathleen E. Willging, Ph.D., Pacific Institute for Research and Evaluation, Behavioral Health Research Center of the Southwest, Albuquerque, NM
Alison B. Hamilton, Ph.D., M.P.H., VA Health Services Research & Development Center for the Study of Healthcare Innovation, Implementation & Policy, VA Greater Los Angeles Healthcare System; UCLA Dept. of Psychiatry and Biobehavioral Sciences
This workshop will provide an overview of the use of qualitative research methods in behavioral health services research (HSR). Specifically, we will discuss study design options, formulation of HSR specific aims, and methodologic decision-making in the context of project resources (e.g., time, expertise, funding). Presenters will describe examples of qualitative behavioral HSR in diverse settings such as the Veterans Administration, clinical and community settings, and criminal justice programs. Participants are encouraged to bring their qualitative research questions and concerns for group discussion and problem-solving.
Workshop 3: 3:30pm– 5:00pm
Strategies for Publishing in Addiction Health Services Research: A Workshop for Early Career (and any other) Investigators
Deborah Garnick, ScD, Brandeis University
Carmen Masson, PhD, University of California, San Francisco
Mark McGovern, PhD, Dartmouth
Richard Saitz, MD, MPH, Boston University
Jeffrey Samet, MD, MPH, Boston University
James Sorensen, PhD, University of California, San Francisco
This workshop will provide an opportunity for participants to engage with seasoned researchers in the field of addiction health services, several of whom are also journal editors, and to hear their suggestions for strategies to help increase one’s success in publishing in peer-reviewed journals. Come prepared with your questions and suggestions for a lively session!