Abstract 1530: Efficient Adjudication of Endpoints in an International Trial: Strategic Lessons from the Clinical Events Committee (CEC) of the Assessment of Pexelizumab in Acute Myocardial Infarction (APEX AMI) Trial
Background: Independent adjudication of clinical trial events is important but requires substantial resources.
Methods: A CEC adjudicated all suspected CHF and cardiogenic shock (CS) events through 90 days in APEX AMI. Detailed information was collected on CHF/CS events by electronic database capture (EDC) with limited source documents. Two subsequent strategies were used to adjudicate each potential event:
a computer algorithm (followed by physician confirmation) analyzed EDC data to determine whether events met CEC definitions; or
a physician(s) review, if data were inadequate to allow classification by the algorithm.
Results: Of 5745 patients, 282 suspected CS and 465 suspected CHF events were found. The CEC confirmed 196 (3.41%) CS and 277 (4.82%) CHF endpoints. The nature of adjudication required and disagreements between sites and the CEC are shown (Table⇓). Overall, 242/727 (32.4%) of suspected events were classified by computer algorithm. Of events not resolved by computer algorithm, the CEC agreed with site assessments in 126/277 (45%) of CHF and 151/196 (77%) of CS events. Ninety-day mortality was: 2.96% if CEC and site agreed no CS occurred; 47.68% if CEC and site agreed CS occurred; and 77.78% if CEC said no CS and site said yes CS occurred.
Conclusions: A computerized approach categorized about 1/3 of suspected CHF/CS events. Of the CS or CHF events not resolved by computer algorithm, 37% had disagreement between the site and CEC. Many events that occurred prior to randomization were inappropriately reported as endpoints by the sites. The APEX AMI CEC experience shows that an efficient CEC effort can enhance the quality of endpoint data in clinical trials and indicates the need for improvement of site evaluation of endpoints.