Development and measurement properties of the Safety Threats and Adverse Events in Trauma (STAT) Taxonomy: A critical review

AbstractBackground Adverse events are common in trauma care, with most occurring during the initial resuscitation phase. The Safety Threats and Adverse Events in Trauma (STAT) taxonomy is a comprehensive tool that was created to evaluate both technical and non-technical errors in trauma resuscitations. The objective of this study is to critically appraise the development and measurement properties, including sensibility, reliability, and validity of the STAT taxonomy. Methods A literature search was conducted in MEDLINE, Embase, and CINAHL, supplemented by citation searching. Studies describing the development, evaluation, or adaptation of the STAT taxonomy were included. Evidence was synthesized narratively across tool development, sensibility, reliability, and validity. Results Seven studies met inclusion criteria, encompassing tool development, pilot testing, reliability assessment, and adaptation. The STAT taxonomy was developed using a systematic review and a RAND-modified Delphi process involving 22 trauma experts, resulting in 67 adverse events across nine domains. Early evidence supports strong content validity, with comprehensive coverage of both technical and non-technical domains. Feasibility was demonstrated across simulation and clinical trauma video review settings. Inter-rater reliability was consistently high, with agreement rates of 90.1% in simulation studies and Gwet's AC1 of 0.94 (95% CI: 0.93–0.95) in real-world trauma resuscitations. Preliminary concurrent validity was supported by significant correlations between STAT-identified adverse events and T-NOTECHS domains, including communication and situational awareness. Conclusion The STAT taxonomy demonstrates strong early evidence of methodological rigor, sensibility, and inter-rater reliability, with preliminary support for validity. It appears promising as a standardized tool for identifying adverse events during trauma resuscitation, but further independent multicenter validation is needed before broader implementation.