ID Number		1586
Study Type		Epidemiology
Model		Exposure assessment at base stations [900 MHz (GSM) and 2 GHz (UMTS)], correlations with headache, etc.
Details		Residents in Germany (n = 40,000) assessed by questionnaire for correlations between mobile phone base station exposure and headache and other subjective symptoms. Exposures determined by geo-coding the locations of base stations and residences,as well as information on living-environment and characteristics of the antenna used. Personal dosimeters are also used to verify certain individual exposure assessments in the residence. The study was initiated to support a nationwide survey by the same group that looked at correlations between base station proximity and subjective complaints including headaches and dizziness (Interim Report of German Mobile Telecommunications Research Program (2006) available on http://www.emf-forschungsprogramm.de/forschung/epidemiologie/epidemiologie_verg/epi_020_ZwB_2.pdf. Unfortunately a significant level of uncertainty is associated with the model. Although more precise data on the exact position and other detailed parameters of base station antenna and households results in better agreement with measurements, such information may not be routinely available on a larger scale unless detailed and individual interviews are conducted. The authors conclude the calculation of exposures on the basis of the normally available imprecise data is associated with a relatively high degree of uncertainty. Thus, the model can only be applied in epidemiological studies, when the uncertainty of the input data can be considerably reduced. In a subsequent dosimetry study, the authors used 20 Antennessa EME personal dosimeters to measure ambient RF fields from GSM, DCS, and UMTS base stations and report reasonably consistent emissions measurements that can be correlated with individual exposure.
Findings		No Effects
Status		Completed With Publication
Principal Investigator		University of Bielefeld, Germany - gabriele.berg-beckhoff@uni-bielefeld.de
Funding Agency		BfS, Germany, DMF, Germany
Country		GERMANY
References		Breckenkamp, J et al. Rad Prot Dos, (2008) 131:474-481 Bornkessel, C et al. Radiat Prot Dosimetry, (2008) 124:40-47 Bornkessel, C et al. Radiat Prot Dosimetry., (2010) 140:287-293 Bornkessel, C Wiener medizinische Wochenschrift (1946)., (2011) 161:233-239
Comments		Following the collection of questionnaire data on subjective complaints from 30,047 individuals spread across Germany, a subset of the respondents (n = 1322) was selected by the authors (from predominantly urban regions where RF exposure had previously been demonstrated by the authors to be stronger than in rural areas) to evaluate the performance of a predictive model for exposure calculation. For comparison with the model calculations, measurements of RF fields were performed over each subjects bed using an EME Spy 120 from Antennessa®. Measurements were all performed between March and August 2006 and between the hours of 8:00 am and 9:00 pm at four different positions (center of the pillow, 10 cm to the left, 10 cm to the right, and 50 cm downward from the center of the pillow). The measurements detected several RF sources including FM radio (88108 MHz), analogue TV and DVB-T (174233MHz), TETRA (380400MHz), analogue TV and DVB-T (470830 MHz), DECT (1.881.90 GHz), wireless LAN and Bluetooth (2.402.50 GHz), GSM900 downlink (925960MHz) and uplink (880915MHz), GSM1800 downlink (1805 1880 GHz) and uplink (17101785 GHz) and UMTS downlink (2.1102.170 GHz) and uplink (1.9201.980 GHz). The Antennessa dosimeters (19 in all) were reported to be relatively stable and showed a maximal variation of +/- 1 dB. For the exposure calculation model, several sources of input data were required. For each of 54,000 mobile phone base station sites (> 350,000 antennas) in the study area, the main beam direction, mounting height, and operating frequencies (GSM900, GSM1800, UMTS) for each antenna were obtained from the Federal Network Agency (as of December 2006). In addition, geo-coordinates for each antenna site were obtained from certification material submitted to the Federal Network Agency by each of the major network operators (T-Mobile, E-Plus, Vodafone, and O2). Finally, the spatial orientation (height, direction, surrounding properties, vegetation) of the bedroom and window(s) and the degree of urbanization of the household was obtained via questionnaire and confirmed via interview. Information was not routinely available for base station antenna transmit power, or down-tilt angle, or other detailed characteristics and thus standard values were used instead. The geo-coordinates of the households (used to determine the distance to nearby base stations) was performed by entering the address into either commercial programs or the local land survey office when possible. Finally, the number of base station antennas within a 500 meter radius of each household was determined (the authors demonstrate that at distances greater than 500 meters the exposures are < 10 mW/m2 or 0.061 V/m and thus inconsequential) and the azimuth angle of each window relative to the beam of nearby antenna was estimated. Of 1132 households that fulfilled the criteria of having at least one base station within a 500 meter radius (average across the study population of 3.4 base stations and 17 antenna), the measured level of RF fields in 61% of the households was below the Antennessa dosimeter sensitivity of detection (0.05 V/m). In the rest of the households, values ranged from 0.05 to 1.1 V/m. When the measurements from all households were combined (using 0.05 V/m as the measured value for the 61% of households below detection), an average value of 0.72 V/m was obtained - far below ICNIRP reference values. The corresponding calculated values tended to significantly over-estimated the measured values. When only households with values above the 0.05 V/m sensitivity threshold were compared (n = 440), the corresponding average measured (0.107 V/m) and calculated (0.103 V/m) values correlated better (spearman correlation coefficient 0.194; p < 0.01). When only the most accurate geo-coordinate data was used from 343 households, the precision improved further.