Online Program

Saturday, February 20
PS3 Poster Session 3 & Continental Breakfast sponsored by Capital One Sat, Feb 20, 8:00 AM - 9:15 AM
Ballroom Foyer

A Web-Based System for Randomized Assignment in Clinical Trials Using Minimization (303230)

*Sophie Yu-Pu Chen, University of Michigan 
Brenda W Gillespie, University of Michigan 
Kerby Shedden, University of Michigan 

Keywords: randomization, clinical trials, minimization

“Minimization” is a common way to balance potential confounders when assigning subjects sequentially to treatments in a trial, but several practical challenges result from the need to randomize subjects in real time at their point of enrollment. Therefore, we have implemented a web-based minimization and trials management system that is open source, and is designed to be hosted on Google Appengine, a low cost "platform as service." Our system allows the user to specify the number treatment arms, the number of stratification variables, the number of levels within each variable, and the randomization ratios. Various forms of logging, data retention, and data export are supported, and the system can be configured to allow multiple users to perform assignments for each trial. Google accounts are used to securely restrict access to each project to its authorized personnel, as defined by the project owner. In several years of use our application has never approached the resource limit for free Appengine accounts, allowing us to run it without incurring charges from Google.