Article Text

PDF

7 Statistical modelling of occupational exposure to polycyclic aromatic hydrocarbons using osha data
  1. D L Lee1,
  2. Lavoué2,
  3. Aronson3,
  4. Spinelli4,
  5. Burstyn5
  1. 1University of British Columbia, Vancouver, Canada
  2. 2Université de Montréal, Montreal, Canada
  3. 3Queen’s University, Kingston, Canada
  4. 4British Columbia Cancer Research Centre, Vancouver, Canada
  5. 5Drexel University, Philadelphia, United States of America

Abstract

Objectives Polycyclic aromatic hydrocarbons (PAHs) are a group of chemicals consisting primarily of fused aromatic rings. As environmental pollutants, PAHs are of concern because some variants are carcinogenic. Our objective is to predict probabilities of PAH exposure based on industry to allow assessment of individual PAH exposure through occupational history.

Methods The Occupational Safety and Health Administration (OSHA) provided access to two PAH exposure databanks of U. S. workplace compliance testing compiled between 1979 and 2010. Multivariable logistic mixed-effects models were used to predict, for each industry, the probability of a PAH measurement exceeding OSHA’s permissible exposure level (exceedance fraction, PEL = 0.200 mg/m3). Time, databank, and industry were included as fixed-effects while inspection number, i.e. an identifier for the compliancy inspector, was included as a random-effect. Industry codes, represented by Standard Industrial Classification or North American Industry Classification System, were maintained or collapsed based on the number of measurements per cell to ensure sufficient sample size.

Results Databank records were amalgamated to yield 2,509 day-specific personal measurements representing 756 companies across 45 states. Regardless of industry code used, analysis revealed that for 1980 less than 5% of industry codes had an exceedance fraction (EF) greater than 0.8. The remaining industry codes were equally distributed between an EF range of 0.2–0.8 and an EF less than 0.2. Overall, more than 80% of industry codes had an EF less than 0.5, databank indicator was marginally significant (p < 0.10), and there was an inverse temporal trend of exceeding the PEL, with lower risk in most recent years (albeit not statistically significant).

Conclusions These statistical models allow identification of industries with different risks of elevated PAH exposure. However, because exposure may not be homogeneous within industry codes, future work will involve incorporating information on jobs/occupations with industries to more accurately identify PAH exposure.

Statistics from Altmetric.com

Request permissions

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.