Occupational
Counting the Carnage: How Sensitive are State Data Systems for Occupational Fatality? Chelsea Martin* Chelsea Martin Catherine Wolff Maryalice Nocera Elizabeth McClure Shabbar Ranapurwala Morgan Richey Stephen Marshall
Background: Occupational fatalities are a significant public health concern. Attaining accurate records of occupational fatalities is essential to discern the true number of occupational fatalities, and better target prevention programs. The purpose of this study was to estimate the sensitivity of fatal occupational injury ascertainment based on flags that indicate death “on the job” and “at-work” as recorded in the Chief Medical Examiner (OCME) data system and on death certificates in North Carolina (NC) from 1992 to 2017.
Design: Validation Study
Methods: NC fatal occupational injuries were identified between 1992-2017 from flags in the NC OCME and death certificates data and reviewed for accuracy. Next, we sampled 5% of non-occupational deaths from alternating calendar years between 1992-2017 to identify occupational fatalities that had been inaccurately flagged. We estimated misclassified non-occupational fatalities using the inverse probability of the sampling fraction. Sensitivity of the classification of fatal occupational injuries based on the data systems’ flags was calculated.
Results: From 1992-2017 there were 3,202 occupational fatalities originally classified in the combined database and verified by the study investigative team. We extracted and adjudicated 2,129 non-occupational deaths. Following adjudication, 41 non-occupational deaths were determined to have been misclassified in the sample, suggesting an estimated 1,500 misclassified occupational deaths and an overall sensitivity of 66.0% (95% CI: 64.6%, 67.4%). Sensitivity of the OCME data for civilian adult occupational fatality differed by intent and means [unintentional driving fatality: 61.2%, (58.5%, 63.8%); unintentional non-driving fatality: 71.4%, (69.5%, 73.3%); intentional fatality: 59.6%, (56.2%, 63.0%)].
Conclusion: An estimated 28-41% of occupational fatalities may have been misclassified in OCME systems, with unintentional driving and intentional fatalities being most affected.