Occupational Fatality Data Systems: A Sensitivity Analysis of North Carolina Data Systems for Occupational Fatality from 1992-2017
Catherine S Wolff, Chelsea Leonard Martin, Maryalice Nocera, Shabbar I Ranapurwala, David B Richardson, Elizabeth S McClure, Amelia Martin, Stephen W MarshallBackground:
Accurate surveillance of occupational fatalities is essential to determine their true burden and target prevention programs. The purpose of this study was to estimate the sensitivity of fatal occupational injuries abstracted from the Office of the Chief Medical Examiner (OCME) data system and death certificates in North Carolina (NC) from 1992-2017 overall, by intent, and by race.
Methods:
We identified NC occupational fatalities from 1992-2017 through NC OCME data abstractions. From the non-occupational deaths, we sampled 5% of deaths from OCME data even calendar years with estimations applied to odd years. We estimated misclassified non-occupational fatalities using inverse probability of the sampling fraction and calculated sensitivity of the occupational fatalities identified from NC OCME and NC death certificates data by intent and race.
Results:
Following adjudication of the 5% sample, 41 non-occupational deaths were determined to have been misclassified, suggesting an estimated 1,520 misclassified occupational deaths and an overall sensitivity of 67.8% (66.5%, 69.1%). Sensitivity differed by intent and means [unintentional driving: 62.1%; unintentional non-driving: 73.5%; intentional: 61.3% (58.0%, 64.6%)], and by race [black decedents 63.2%, and white decedents 66.9%].
Conclusions:
Sensitivity ranged from 61-73% for occupational fatalities, with lower sensitivity observed for unintentional driving and intentional fatalities events among black decedents, indicating higher misclassification of these individuals in the OCME data systems.