Regulators require less evidence to take action on drug safety concerns than would be required to grant drug efficacy claims. Such an asymmetrical view of risk points to the need for large studies to judge the safety of novel medicines compared with standard therapy as smaller sample sizes might produce unreliable and misleading toxicity signals. Large studies conducted in the classical manner are expensive and difficult to conduct. In addition, tight entry criteria and strictly protocolized care results in such trials having poor external validity. If the safety of medicines is to be judged efficiently, novel, easier to conduct, and less-expensive solutions are required. Such trials could have improved external validity were they incorporated into normal care. This paper discusses the safety of cyclooxygenase-2 inhibitors and describes the sort of studies that could be carried out to gather better safety information on their use versus standard nonsteroidal anti-inflammatory drugs.