Reliability and reproducibility of subaxial cervical injury description system: a standardized nomenclature schema.

STUDY DESIGN: Radiographic measurement study. OBJECTIVE: To develop a standardized cervical injury nomenclature system to facilitate description, communication, and classification among health care providers. The reliability and reproducibility of this system was then examined. SUMMARY OF BACKGROUND DATA: Description of subaxial cervical injuries is critical for treatment decision making and comparing scientific reports of outcomes. Despite a number of available classification systems, surgeons, and researchers continue to use descriptive nomenclature, such as "burst" and "teardrop" fractures, to describe injuries. However, there is considerable inconsistency with use of such terms in the literature. METHODS: Eleven distinct injury types and associated definitions were established for the subaxial cervical spine and subsequently refined by members of the Spine Trauma Study Group. A series of 18 cases of patients with a broad spectrum of subaxial cervical spine injuries was prepared and distributed to surgeon raters. Each rater was provided with the full nomenclature document and asked to select primary and secondary injury types for each case. After receipt of the raters' first round of classifications, the cases were resorted and returned to the raters for a second round of review. Interrater and intrarater reliabilities were calculated as percent agreement and Cohen kappa (κ) values. Intrarater reliability was assessed by comparing a given rater's diagnosis from the first and second rounds. RESULTS: Nineteen surgeons completed the first and second rounds of the study. Overall, the system demonstrated 56.4% interrater agreement and 72.8% intrarater agreement. Overall, interrater κ values demonstrated moderate agreement while intrarater κ values showed substantial agreement. Analyzed by injury types, only four (burst fractures, lateral mass fractures, flexion teardrop fractures, and anterior distraction injuries) demonstrated greater than 50% interrater agreement. CONCLUSION: This study demonstrated that, even in ideal circumstances, there is only moderate agreement among raters regarding cervical injury nomenclature. It is hoped that more familiarity with the proposed system will increase reproducibility in the future. Additional research is required to establish the clinical utility of this novel nomenclature schema.

Duke Scholars

Author Christopher Ignatius Shaffrey Orthopaedic Surgery