Early warning scores in the perioperative period: applications and clinical operating characteristics.

Abstract

Early warnings scores are designed to detect clinical deterioration and promote intervention at the earliest possible moment. Although the ultimate effects on patient outcomes are unclear, early warning scores are now legally mandated in several countries. Here, we review the performance of early warning scores in surgical and perioperative populations. Early warning scores can be used to screen for postoperative deterioration and surgical complications. We describe a framework to evaluate the balance between missed events and warning signals that are not followed by an adverse event (nonevents). In large surgical cohort studies, the missed event rates ranged between 19 and 69% and the nonevent rates ranged between 72 and 99% for 'optimal' threshold early warning sores. Recent investigations have shown that there may be a substantial discrepancy between the theoretical benefits shown in validation studies and the practical clinical implementation of early warning scores, which may partly explain the absence of measurable benefit from these systems. Early warning scores may facilitate protocolized escalation of care for patients at risk of adverse events and can be used in surgical and postoperative patients, but high nonevent rates and practical implementation problems can restrict their usefulness.