Adjudication of clinical events is often used as a quality assurance method in clinical research. During the design of the Viral Activation Transfusion Study (a clinical trial in patients with advanced HIV disease), a set of study endpoints was defined (primarily AIDS-defining conditions), criteria for confirmation of each event type were developed, and an adjudication procedure was established. The adjudication process included 1) an initial review of documentation of each event by two independent reviewers, 2) the opportunity to request additional information, 3) a second review either of additional documentation or of cases in which there was disagreement on first review, and 4) the consultation of a third reviewer if there was still disagreement. Overall, of 288 reported endpoints, 30% required additional documentation or more than one review, and 16% were not confirmed at the end of the adjudication process. However, these percentages varied widely over different types of events. For example, of 30 reported nonophthalmalogic cytomegalovirus events, 37% required additional documentation and 40% were not confirmed. In contrast, every one of 17 reported Pneumocystis cariini pneumonias were confirmed with no requirement for additional documentation. The results can be used to help design endpoint documentation and adjudication procedures for other studies, thereby improving data quality and reducing costs.