Accuracy and performance evaluation of a clinical decision support system for laboratory result alerts

This study was designed to evaluate the accuracy and performance of a clinical decision support (CDS) architecture consisting of a knowledge engine and data interface adapter (DIA), which were developed as critical components for a clinical decision support system (CDSS). Accuracy and performance are the most essential quality attributes. CDSSs are becoming increasingly important in efforts to increase quality control in hospitals, but if a CDSS adds a new burden to the legacy system, the manager is unlikely to adopt it in the hospital information system. A laboratory alerting system was selected for evaluation, because of complex rules and the large amount of the data interface. Evaluation cases (N=323,455) were captured from a tertiary care and teaching hospital over a 3-month period. We were able to confirm the accuracy, stability, and feasibility of the CDS application architecture and the knowledge engine, but we identified areas of the DIA that require improvement in order to obtain satisfactory performance.