EFFECTS FUNCTION-ANALYSIS OF ELF MAGNETIC-FIELD EXPOSURE IN THE ELECTRIC UTILITY WORK-ENVIRONMENT

The incomplete understanding of the relation between power-frequency f ields and biological responses raises problems in defining an appropri ate metric for exposure assessment and epidemiological studies. Based on evidence from biological experiments, one can define alternative me trics or effects functions that embody the relationship between field exposure patterns and hypothetical health effects. In this paper, we e xplore the application of the ''effects function'' approach to occupat ional exposure data. Our analysis provides examples of exposure assess ments based on a range of plausible effects functions. An EMDEX time s eries data set of ELF frequency (40-800 Hz) magnetic field exposure me asurements for electric utility workers was analyzed with several stat istical measures and effects functions: average field strength, combin ation of threshold and exposure duration, and field strength changes. Results were compared for eight job categories: electrician, substatio n operator, machinist. welder, plant operator, lineman/splicer, meter reader, and clerical. Average field strength yields a different rankin g for these job categories than the ranks obtained using other biologi cally plausible effects functions. Whereas the group of electricians h as the highest exposure by average field strength, the group of substa tion operators has the highest ranking for most of the other effects f unctions. Plant operators rank highest in the total number of field st rength changes greater than 1 mu T per hour. The clerical group remain s at the lowest end for all of these effects functions. Our analysis s uggests that, although average field strength could be used as a surro gate of field exposure for simply classifying exposure into ''low'' an d ''high,'' this summary measure may be misleading in the relative ran king of job categories in which workers are in ''high'' fields. These results indicate the relevance of metrics other than average field str ength in occupational exposure assessment and in the design and analys is of epidemiological studies. (C) 1997 Wiley-Liss, Inc.