Simplification of the Wertheimer-Leeper wire code

Wire codes, introduced by Wertheimer and Leeper as a surrogate for residential magnetic fields, divide imputed exposure into several categories based on the configurations of electrical wiring within 40 m of homes. Using the data set gathered by Savitz et al. in the Denver, Colorado, area of the United States, we determined which of the wiring attributes that contribute to Wertheimer-Leeper coding are actually associated with low-power magnetic fields measured in bedrooms of subjects. The results led us to propose a considerably simplified three-category form of the Wertheimer-Leeper code that 1) drops the distinctions between thick and thin primary wires and between first-span and other secondary lines, 2) adds a new distinction between open (i.e., conductors not in physical contact) and spun secondaries, and 3) explains as much of the between-home variability in log-transformed bedroom fields as does the five categories of the original Wertheimer-Leeper code. The data necessary to classify residences using the modified code are considerably simpler to obtain and should lead to more reliable results. A separate reanalysis of the Denver data set of Savitz et al, shows that the modified code yields risk estimates that are both precise and markedly elevated for the highest exposure category, suggesting that this code may be useful in other studies.