This research addressed the question of how well measurement data collected during a single visit, made at an arbitrary hour of day, day of week, and season, estimate longer term residential 60 Hz magnetic field levels. We made repeat spot and 24 h measurements in 51 children's home, located in the Detroit, MI, and the Minneapolis-St. Paul, MN metropolitan areas, on a regular bimonthly schedule over a 1 year period, as well as a single 2 week measurement, for total of eight visits, producing 21 days of data for each residence. We defined the long term estimate (LTE) as the geometric mean of all available 24 h geometric means from the first six bimonthly visits. The LTE served as the reference level for assessing seasonal, day of week, and diurnal effects, as well as the potential for misclassification. We found a small, but statistically significant (P<.05), seasonal effect, with levels approximately 3% lower than the LTE in the spring and about 4% greater during the summer. No effect was found for day of week. However, we did find a systematic and appreciable diurnal effect, suggesting that, for example, an evening spot measurement may overestimate the LTE by 20% or more. We also assessed how well the 24 h measurement from the last visit, which was not used in calculation of the LTE, estimated the LTE. We found a high degree of correlation (r=.92) and fair to good agreement using four exposure categories (kappa=.53). Thus, the 24 h measurement appears to be a satisfactory LTE estimator. However, this finding must be interpreted with caution since considerable unexplained variability was present among the repeat 24 h measurements in about one-third of the homes. While the 2 week measurement does somewhat decrease exposure misclassification, its added intrusiveness and cost are likely to outweigh the improved precision.

No abstract

Bowman et al. used epidemiologic data to test a model in which subjects were classified as being "in-resonance" or "not-in-resonance" for 60-Hz magnetic-field exposures depending on single static magnetic-field measurements at the centers of their bedrooms. A second paper by Swanson concluded that a single static magnetic-field measurement is insufficient to meaningfully characterize a residential environment. The main objective of this study was to investigate exposure-related questions raised by these two papers in two U.S. data sets, one containing single spot measurements of static magnetic fields at two locations in homes located in eight states, and the other repeated spot measurements (seven times during the course of one year) of the static magnetic fields at the centers of bedrooms and family rooms and on the surfaces of beds in 51 single-family homes in two metropolitan areas. Using Bowman's criterion, bedrooms were first classified as being in-resonance or not-in-resonance based on the average of repeated measurements of the static magnetic field measured on the bed where the presumed important exposure actually occurred. Bedrooms were then classified a second time using single spot measurements taken at the centers of bedrooms, centers of family rooms, or on the surfaces of beds, as would be done in the typical epidemiologic study. The kappa statistics characterizing the degree of concordance between the first (on-bed averages) and second (spot measurements) methods of assessing resonance status were 0.44, 0.33, and 0.67, respectively. This level of misclassification could significantly affect the results of studies involving the determination of resonance status.

No abstract