Unpublished and misreported studies make it difficult to determine the true value of a treatment. Peter Doshi and colleagues call for sponsors and investigators of abandoned studies to publish (or republish) and propose a system for independent publishing if sponsors fail to respond