Analysis and actions after laboratory errors in a Chinese university hospital.

Background: Diagnostic errors pose a critical threat to patient safety, heavily relying on accurate laboratory medicine. However, research specifically addressing laboratory errors (LEs) remains limited globally. This study aimed to categorize LEs, identify their root causes, and develop targeted interventions within a large, specialized hospital in China, where systemic factors amplify their potential impact.

Methods: A retrospective quality improvement study was conducted in the ISO 15,189 and CAP-accredited Department of Medical Laboratory at a women and children's hospital. Eighty-three documented LEs (51 general, 32 transfusion-specific) from March 2016 to April 2023 were analyzed. Errors were captured via internal incident reporting and hospital risk management systems. LEs were classified using five criteria: responsibility attribution (exclusively lab, extra-lab, conjoint, undetermined), testing phase (preanalytical, analytical, postanalytical), error type, preventability (using a cognitive psychology framework: cognitive vs. noncognitive), and patient impact. Root cause analysis and corrective actions were tracked.

Results: Among the 51 general LEs, the preanalytical phase was most error-prone (51.0%), primarily due to specimen collection (29%) and request procedure errors (22%). Analytical (4%) and postanalytical (18%) phases had fewer errors. Responsibility analysis showed 20% exclusively lab-originated, 60% extra-lab-originated, and 16% conjoint. Cognitive errors dominated preventable incidents. Environmental/infrastructure (6%) and Laboratory Information System (LIS) errors (14%) were significant concerns. Separately, among 32 transfusion-related errors, clinical physicians bore primary responsibility in 51%, with common issues being improper specimen collection (22%) and non-evidence-based orders (16%). Corrective actions (e.g., workflow optimization, staff training, improved communication, LIS upgrades like an electronic critical value notification system, facility relocation) led to significant reductions in preanalytical errors over time. Improvements were achieved cost-effectively.

Conclusion: Preanalytical errors are the most prevalent LEs, often originating outside the laboratory. Cognitive errors are highly preventable. Implementing targeted interventions based on systematic error classification and root cause analysis-including technological solutions (e.g., electronic alerts, LIS improvements), workflow simplification, enhanced training (especially for non-laboratory personnel in transfusion contexts), and interdepartmental communication-significantly reduces LEs and enhances laboratory quality management. Continuous monitoring and context-specific strategies are crucial, especially in large healthcare systems. Study limitations include potential underreporting and limited generalizability beyond specialized women and children's hospital settings.