Evaluation of benefits derived from a computerized data management system for clinical trials data.

Identification and reporting of adverse events is an important responsibility of investigators conducting clinical trials of new pharmaceutical compounds. We have designed and implemented a computerized data management system for clinical trials data and have evaluated its performance in managing adverse event information from clinical studies. All cases in which adverse event occurred and a random sample of cases without adverse events were selected from two clinical trials managed with the new computerized system and compared with identically selected cases from two clinical trials managed with the paper system. Rates of transcription errors, data selection errors and failure to identify adverse events were examined by logistic regression. Analysis of variance of the time required to evaluate clinical information or report adverse events to sponsors was also conducted. Implementation of the computer system resulted in significantly fewer errors in data transcription (17% vs 0%; P less than 0.001) and fewer cases in which adverse events were overlooked (35% vs 3%; P less than 0.001). However, data selection errors were more common using the computer system (2% vs 8%; P = 0.03) because the items evaluated were transferred electronically, rather than selected for appropriateness by a clinician. The average time required to evaluate clinical information and to report adverse event information to sponsoring agencies has been reduced by 2-4 months (P less than 0.001) depending on the severity of the event. This reduction has improved compliance with federal regulations governing the responsibilities of clinical investigators. It has also permitted early identification of toxicities and appropriate amendment of research protocols, thus reducing risk to patients enrolled in clinical trials.