SAS code to compare embedded adaptive interventions.
What does this software do?
This software identifies the sequence of intervention options in a SMART that produces the best outcome on average.
Which SMART experimental designs can this software analyze?
This software is designed to analyze data from SMARTs with a continuous outcome, two decision stages, up to two treatment options at each decision stage.
How does this software work?
This software applies Q-learning, an extension of the standard regression method designed for use with longitudinal data, to multiple stages of decision making. Q-learning implements two regressions. The first selects the best second-stage treatment. The second selects the best first-stage treatment given the best second-stage treatment. The backward ordering of regressions avoids selecting treatment options that appear optimal in the short term but may lead to a less desirable outcome in the long run.
Key Terms
Q-Learning
Q-learning, a method developed in computer science, is an extension of the standard regression method that can be used with longitudinal data in which treatments and covariates vary over time.
SMART
A sequential, multiple assignment, randomized trial (SMART) is an experimental design that enables scientists to build better adaptive interventions.
Adaptive Intervention
An adaptive intervention is a course of time-varying treatment designed to adapt to an individual’s changing life circumstances, response to treatment, or other designated indicator.