ABSTRACT
Most Phase 2 or 3 clinical trials are designed to study the efcacy of a new treatment or, for a treatment already in use, its utility for a new application. People rarely enroll in a trial designed to demonstrate that an intervention under study is not harmful; rather, rational participants who understand the informed consent document they have signed trust that the investigator will tell them if the ongoing data show more risks than anticipated at the time they agreed to enter the study. A well-designed protocol species the outcomes for measuring efcacy. In safety monitoring, on the other hand, important outcomes are often unknown. A taxonomy of harms provides a structure for thinking about how to monitor for safety. A drug or other intervention has some risks that are known, some that are unexpected but nonserious, others that are unexpected but serious, those that are unexpected and life threatening, and some that though not biologically credible are frightening, if true (Proschan et al. 2006; Crowe et al. 2009; Xia et al. 2011).
