Occupation and Industry Data Quality Among Select Notifiable Conditions in Washington State

Abstract

Context: Occupation and industry are basic data elements that, when collected during public health investigations, can be key to understanding patterns of disease transmission and developing effective prevention measures. Objective: To assess the completeness and quality of occupation and industry data among select notifiable conditions in Washington and discuss potential improvements to current data collection efforts. Design: We evaluated occupation and industry data, collected by local health departments during routine case investigations, for 11 notifiable conditions, selected for inclusion based on an established or plausible link to occupational exposure. Setting and Participants: Confirmed cases of select notifiable conditions among Washington residents aged 16 to 64 years, for years 2019-2021. Main Outcome Measures: We calculated the percentage of cases among working-age adults reported as employed, the percentage with occupation and industry data collected, and the percentage assigned standard occupation and industry codes. We identified the most common responses for occupation and industry and challenges of assigning codes to those responses. Results: Among the 11 conditions evaluated, one-third of cases aged 16 to 64 years were reported as employed. Among the cases reported as employed, 91.5% reported occupation data and 30.5% reported industry data. “Self-employed” was among the top responses for occupation, a response that does not describe a specific job and could not be assigned an occupation code. In the absence of additional information, 4 of the most common responses for industry could not be coded: “health care,” “technology,” “tech,” and “food.” Conclusion: Routine collection of informative occupation and industry data among working-age adults is largely absent from case investigations in Washington. Methods of data collection that improve quality while minimizing the burden of collection should be pursued. Suggestions for improving data quality are discussed.