Online Program

Return to main conference page

Thursday, January 11
Thu, Jan 11, 2:00 PM - 3:45 PM
Crystal Ballroom E
Generalizability

Estimating population effects: Case study of Generalizing Results of Methamphetamine Dependence Trials (304200)

*Benjamin Ackerman, Johns Hopkins Bloomberg School of Public Health 
Ramin Mojtabai, Johns Hopkins Bloomberg School of Public Health 
Kara Rudolph, University of California Berkeley 
Elizabeth A Stuart, Johns Hopkins Bloomberg School of Public Health 
Ryoko Susukida, Johns Hopkins Bloomberg School of Public Health 

Keywords: generalizability, randomized trial, methamphetamine, substance use, external validity

Results from randomized trials are often utilized in the formulation of health policy. However, there is growing concern that results from trials may not generalize well to the populations for which policies or decisions are being made. In trials related to substance use disorders (SUDs), strict exclusion criteria make it challenging to obtain study samples that are fully “representative” of the populations that researchers wish to generalize their results to, such as individuals actually seeking treatment for substance use across the US. Recently developed methods have proposed ways to estimate effects in target populations, using data from randomized trials and the target population. This work illustrates these methods in a case study of three trials of treatments for methamphetamine dependence by generalizing the trial results to well-defined target populations obtained from the TEDS-A, an annual census dataset collecting information on admissions to SUD treatment facilities. Our aim is to implement these methods in order to obtain more accurate target population-level treatment effect estimates, providing policymakers with greater insight on the potential impact of a treatment.