Modified Clavien-Dindo-Sink Classification System for operative complications in adult spine surgery.

Abstract

Currently there is no standardized mechanism to describe or compare complications in adult spine surgery. Thus, the purpose of the present study was to modify and validate the Clavien-Dindo-Sink complication classification system for applications in spine surgery. The Clavien-Dindo-Sink complication classification system was evaluated and modified for spine surgery by four fellowship-trained spine surgeons using a consensus process. A distinct group of three fellowship-trained spine surgeons completed a randomized electronic survey grading 71 real-life clinical case scenarios. The survey was repeated 2 weeks after its initial completion. Fleiss' and Cohen's kappa (κ) statistics were used to evaluate interrater and intrarater reliabilities, respectively. Overall, interobserver reliability during the first and second rounds of grading was excellent with a κ of 0.847 (95% CI 0.785-0.908) and 0.852 (95% CI 0.791-0.913), respectively. In the first round, interrater reliability ranged from good to excellent with a κ of 0.778 for grade I (95% CI 0.644-0.912), 0.698 for grade II (95% CI 0.564-0.832), 0.861 for grade III (95% CI 0.727-0.996), 0.845 for grade IV-A (95% CI 0.711-0.979), 0.962 for grade IV-B (95% CI 0.828-1.097), and 0.960 for grade V (95% CI 0.826-1.094). Intraobserver reliability testing for all three independent observers was excellent with a κ of 0.971 (95% CI 0.944-0.999) for rater 1, 0.963 (95% CI 0.926-1.001) for rater 2, and 0.926 (95% CI 0.869-0.982) for rater 3. The Modified Clavien-Dindo-Sink Classification System demonstrates excellent interrater and intrarater reliability in adult spine surgery cases. This system provides a useful framework to better communicate the severity of spine-related complications.