Many clinical trials and medical experiments involve an end event, such as the end of pain, relief of symptoms, rejection of the transplant, or death. The analysis of data from such studies has been criticized for not considering when the event happened, and for the inappropriate use of information about test subjects who left before the awaited event did or did not happen. So new approaches to trials and analysis have been developed, and some of these are described here by researchers in statistics and various medical specialties. Among the topics are parametric methods, inferential methods for categorical time-to-event endpoints, an efficient alternative to the Cox model for small trials, the design and analysis of analgesic trials, antiviral trials, parametric likelihoods for multiple nonfatal competing risks and death, and chronic carcinogenicity studies of pharmaceuticals in rodents. Annotation ©2009 Book News, Inc., Portland, OR (booknews.com)