A Population-Based Job Exposure Matrix for Power-Frequency Magnetic Fields

Abstract

A population-based job exposure matrix (JEM) was developed to assess personal exposures to power-frequency magnetic fields (MF) for epidemiologic studies. The JEM compiled 2317 MF measurements taken on or near workers by 10 studies in the United States, Sweden, New Zealand, Finland, and Italy. A database was assembled from the original data for six studies plus summary statistics grouped by occupation from four other published studies. The job descriptions were coded into the 1980 Standard Occupational Classification system (SOC) and then translated to the 1980 job categories of the U.S. Bureau of the Census (BOC). For each job category, the JEM database calculated the arithmetic mean, standard deviation, geometric mean, and geometric standard deviation of the workday-average MF magnitude from the combined data. Analysis of variance demonstrated that the combining of MF data from the different sources was justified, and that the homogeneity of MF exposures in the SOC occupations was comparable to JEMs for solvents and particulates. BOC occupation accounted for 30% of the MF variance (p ≪ 10⁻⁶), and the contrast (ratio of the between-job variance to the total of within- and between-job variances) was 88%. Jobs lacking data had their exposures inferred from measurements on similar occupations. The JEM provided MF exposures for 97% of the person-months in a population-based case-control study and 95% of the jobs on death certificates in a registry study covering 22 states. Therefore, we expect this JEM to be useful in other population-based epidemiologic studies.

[Supplementary materials are available for this article. Go to the publisher's online edition of the Journal of Occupational and Environmental Hygiene for the following free Supplemental resources: Appendices A–D and Formula Proofs.]

Keywords:

• ELF
• EMF
• exposure assessment
• extremely low frequency
• JEM

ACKNOWLEDGMENTS

We gratefully acknowledge the following contributors: Birgitta Floderus (National Institute for the Working Life) who helped us translate the Nordic occupational codes; Fabriziomaria Gobba (University of Modena and Reggio Emilia) who gave us the complete Italian data set; and Bryan Goldman (University of Washington) who programmed the JEM database. We also thank Dr. Margaret Wrensch of the University of California at San Francisco who directed the SFBF adult glioma study that provided the initial motivation and support for this project.

Notes

^A Electricians working in residences, commerce, and construction.

^B Fitters are electricians working in electric utilities and other industries.

^C Data not included in JEM, only U.S. data used for this category.

^A Originally coded to a miscellaneous BOC category.

^B Data not included in JEM as the exposures of clericals working at a hydro generating plant were unusually high.