Sequential Multiple Assignment Randomized Trials: Learning and Adaptation through Biostatistics

Patients’ responses to treatment often depend on their individual features, and the best next step generally depends on how they responded to the last treatment. Sequential multiple assignment randomized trials (SMARTs) are the gold-standard design for developing and evaluating adaptive treatment strategies, also known as dynamic treatment regimes: sequences of decision rules that tailor treatment over time to each patient's evolving response.

SMART-LAB develops state-of-the-art statistical methods for the design and analysis of SMARTs. Our research includes estimating and comparing treatment regimes, adaptive randomization, interim monitoring, sample size planning, and identifying optimal regimes. We also work directly with clinical researchers to design SMARTs for their studies, and we build statistical software that makes these methods practical to use in biomedical research.

SMART-LAB is a collaboration among the University of Rochester, University of Pittsburgh, MD Anderson Cancer Center, University of Arizona, and Konkuk University in South Korea.

For questions or collaboration inquiries, contact us at abdus_wahed@urmc.rochester.edu and/or yucheng@pitt.edu.

widget-bc087844-617