ISR Awards

Statistical Methods for Sequential Treatment Decisions in Alcoholism

This project involves the evaluation of existing methods (developed and published by Nahum-Shani and colleagues: Psychological Methods, 2012, 7; 4) for constructing high quality adaptive interventions based on SMART (Sequential Multiple Assignment Randomized Trial) study data. Dr. Nahum-Shani and Dr. Efrati will work closely with the University of Pennsylvania investigator (Dr. Kevin Lynch) to build a simulation model mimicking the ExTENd data set and evaluating the method developed by Nahum-Shani and colleagues to compare adaptive interventions that are embedded in a SMART design. The U-M investigators will use the simulations to challenge the method in ways that may occur with real data such as ExTENd (e.g., when people dropout early, the distribution of the primary outcome is non-normal). The U-M investigators will also take an active role in disseminating the results through publications.