Occupational
Validating a Task-Based Asbestos Job Exposure Matrix for Screening Pulmonary Disease in Trade Workers: A Calibrated Approach for Surveillance and Prevention Sundus Siddique* Sundus Siddique Siddique Siddique Siddique Siddique Siddique Siddique Northwell, New Hyde Park, Department of Occupational Medicine, Epidemiology and Prevention, Feinstein Institutes for Medical Research, NY, USA
Background: Cumulative asbestos exposure causes respiratory impairment and malignancies, highlighting the need for industry specific job-exposure matrices (JEMs) for surveillance and risk assessment.
Methods: We constructed a task-based JEM for 492 unionized workers using Northwell Occupational Lung Registry data. Three indices (linear & two nonlinear frequency weightings) integrated 29 tasks, exposure frequency, duration, and employment era. Cronbach’s α, principal component analysis and kappa were used to examine reliability, dimensions and agreement. Associations with fibrosis, pleural plaques and impaired pulmonary function were estimated via logistic regression. We examined discrimination (AUC), sensitivity, specificity, and calibration; analyses adjusted for trade, stratified by smoking.
Results: Median age was 61 years; ~60% had smoked. Rates of FEV1 < 80%, FVC < 70%, fibrosis, and pleural plaques were 22%, 7%, 15%, and 17%. JEM indices were reliable (Cronbach’s α > 0.87), unidimesional and high in agreement (kappa ≥ 0.7). High scores were associated with radiological and functional impairment (p < 0.01); the highest JEM quartile accounted for most adverse outcomes. Associations between JEM scores and radiological outcomes were significant after adjusting for trade and stratifying by smoking. Discrimination for fibrosis was moderate (AUC 0.68-0.70); calibration was excellent ((Brier 0.119-0.124), nonsignificant Hosmer-Lemeshow test, decile calibration plots showed agreement between observed and predicted outcomes). Using optimized cut-points, 35% of workers were classified as high-risk (sensitivity 63-67%, specificity 68-69%). Risk assignment was concordant across JEM types.
Conclusions: We validated that our task-based asbestos JEM was associated with adverse pulmonary outcomes and was a reliable and accurately calibrated tool. Both linear and nonlinear task scoring provided consistent risk classification, supporting the use of this approach in surveillance.