July 30, 2012
We are pleased to announce the release of PROC QLEARN, a SAS procedure for building adaptive health interventions. PROC QLEARN was developed by the Methodology Center to analyze data from sequential, multiple assignment, randomized trials (SMARTs). This software will help researchers build interventions that adapt at the right times to improve participant outcomes (e.g., intensifying for people who do not respond to initial treatment) while decreasing the cost and burden of the intervention (e.g., stepping down treatment for responsive participants).
SMART designs were developed by Methodology Center Investigator Susan Murphy and her collaborators to provide the information needed to construct high-quality adaptive interventions. The procedure uses a generalization of Q-learning, a method developed in computer science, which allows treatments and covariates to vary over time.