Biostatistics Weekly Seminar


Improving Individual Outcomes: Optimizing Stage Duration in a Sequential Multiple Assignment Randomized Trial (SMART)

Hayley Belli, PhD
New York University Langone School of Medicine

In the era of precision medicine, there is a clear need to design more patient-focused, pragmatic clinical trials. In this talk, I will introduce a statistical method for selecting individualized treatment duration in a Sequential Multiple Assignment Randomized Trial (SMART). In this type of adaptive design, patients move through a series of stages with the option to continue or switch interventions at the end of each stage based on treatment response. Over time, patients will be assigned to more effective interventions. However, under the standard SMART framework, the length of the treatment period is usually fixed, selected by investigators in advance (without much guiding data), and applied uniformly to all participants in the study. We address these short comings of SMARTs by introducing a likelihood-based method to determine when an individual patient should stay on a treatment or switch interventions prior to the stage end. This new methodology uses data and design principles to provide patients with treatment options that are advantageous both clinically and ethically, all within the rigorous, experimental framework of a randomized controlled trial. Following derivation, we illustrate the performance of this algorithm using data from the Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care (EMBARC) Study, a two-stage SMART design measuring the effectiveness of sertraline in patients with Major Depressive Disorder.


MRBIII, Room 1220
23 January 2019
1:30pm


Speaker Itinerary

Topic revision: r1 - 08 Jan 2019, TawannaPeters
 

This site is powered by FoswikiCopyright © 2013-2017 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback