Objectives To provide insight into the contributions of exposure measurements to job exposure matrices (JEMs), we examined the sensitivity of exposure-response associations between occupational benzene exposure and non-Hodgkin lymphoma (NHL) using benzene JEM-based estimates with and without measurement calibration.
Methods NHL risk was examined in a prospective population-based cohort of 73,087 Shanghai women with no prevalent cancer at baseline and a valid occupational history. An ‘uncalibrated’ benzene JEM was developed using expert judgment to assign ordinal (0 to 3) probability and intensity ratings to each occupation and industry group. JEM intensity estimates were combined with >60,000 short-term, area benzene inspection measurements using a mixed-effects model framework that incorporated fixed effects for year and intensity rating and random effects for occupation, industry, and measurement occurrence. The model was used to derive ‘calibrated JEM’ estimates from the fixed effect model parameters and ‘job/industry-specific’ estimates from both the fixed effect model parameters and the best unbiased linear prediction estimates from the occupation and industry random effect terms. Cumulative exposure for each subject was calculated for each approach using varying exposure definition criteria based on the JEM’s probability ratings. We examined the agreement between the cumulative metrics and evaluated changes in the benzene-NHL associations.
Results For our primary exposure definition, the job/industry-specific estimates were moderately to highly correlated with the calibrated and uncalibrated JEM estimates (Spearman correlation amongst exposed subjects: 0.64–0.87). The uncalibrated, calibrated, and job/industry-specific metrics all resulted in statistically significant exposure-response associations for NHL, with similar model fit. This similarity occurred because the measurement- and ordinal-based weights across the occupation intensity ratings were nearly identical in this study.
Conclusions The measurement-based JEM estimates had similar model fit to the uncalibrated JEM estimates in this study; however, the measurement-based estimates allowed us to evaluate quantitative exposure-response curves.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.