During drug development we focus formidable resources on defining the overall efficacy and safety of a chemical entity. This process moves from basic, theoretical research, through development of compound screening paradigms, into in vitro and animal testing systems and finally into clinical studies where we define the indications and patient populations where the compound has a favorable benefit-risk profile. The process of clinical development is extremely demanding of resources, expertise and time and it generates a great deal of data from which we make an overall benefit-risk assessment for a specific indication and patient population. We then must select from that data in communicating the efficacy and safety profile of the new compound to medical professionals and patients so that they can make individual patient-based decisions on whether or not the medication is appropriate for a given patient. Despite the time and resources devoted to this process there remain uncertainties both on the side of the data on a compound and its interpretation as well as on the side of the individual benefit-risk decision for patients. In the clinical development program we will usually be able to show that a certain - hopefully substantial—proportion of patients with the condition being treated will benefit from the drug but it is never the case that efficacy is demonstrated for all patients. We will also note that a variety of adverse events occur in patients exposed to the drug, but there is a great deal of uncertainty regarding both which patients will experience adverse events and the extent to which the medication has caused those adverse events.