Reproducibility of preventable death judgments and problem identification in 60 consecutive road trauma fatalities in Victoria, Australia

Background: Since 1992, the Consultative Committee on Road Traffic Fatalities in Victoria has identified problems in the management of traffic fatalities. Its two evaluative committees have additionally assessed the potential preventability of death. Previous studies have shown only poor to fair reproducibility of death judgments. Methods: Problems in the management of 60 consecutive road traffic fatalities and the potential preventability of death were independently evaluated by the two committees. Inter-rater and inter-committee concordance were analyzed using the κ statistic. Results: Reproducibility was high. Inter-committee agreement on nonpreventable, potentially preventable, and preventable death judgments was high (κ = 0.73, 95% confidence interval = 0.57-0.89). Agreement within the two evaluative committees was also high (average weighted κ = 0.73 and 0.74). There was good agreement between committees on problems identified, including those contributing to death. Conclusion: The high κ concordance on preventable death judgments and the agreement on problem identification supports the reproducibility of the methodology used.