..................................................................................................................................................................................................................
..................................................................................................................................................................................................................
In re-examining the Brazilina cell phone base station study, Dr. Ken Foster came up with the following additional observation concerning Fig 16 from the article. We have already noted that this chart does not make sense because cancer mortality dropped for each additional year of exposure. This study was done during a period in which the number of base stations was growing rapidly. Given the fact that the study only "counted" deaths within 1 km of a base station, Fig 16 may simply reflect the number of base stations over time.
To check this, Dr. Foster found the web site containing the data used by Dode et al. on the installation history of base stations in the city. In 1996 there would have been relatively few base stations, and relatively few residents would be counted in the mortality statistics. This corresponds to the 9 - 10 year period in Fig 16 and explains why so few deaths were "counted" in that time slot. As the number of base stations increased, the number of deaths that were counted increased.
Dr Foster plotted Fig. A below in which the number of base stations over time is plotted against the deaths reported in Fig 16. Both results are normalized to their respective values in 2005. The match is extremely good, which indicates a key flaw in the study design. The selection criteria for cancer deaths that the authors used are merely a surrogate for the number of base stations. This is further evidence that selection bias skewed the results - only 7191 cancer deaths out of a total of 22,493 were "selected" using dubious criteria. This precludes making any conclusions about the death rate as a function of distance from such stations. Fig. A and the additional images were provided by Dr. Foster.
To check this, Dr. Foster found the web site containing the data used by Dode et al. on the installation history of base stations in the city. In 1996 there would have been relatively few base stations, and relatively few residents would be counted in the mortality statistics. This corresponds to the 9 - 10 year period in Fig 16 and explains why so few deaths were "counted" in that time slot. As the number of base stations increased, the number of deaths that were counted increased.
Dr Foster plotted Fig. A below in which the number of base stations over time is plotted against the deaths reported in Fig 16. Both results are normalized to their respective values in 2005. The match is extremely good, which indicates a key flaw in the study design. The selection criteria for cancer deaths that the authors used are merely a surrogate for the number of base stations. This is further evidence that selection bias skewed the results - only 7191 cancer deaths out of a total of 22,493 were "selected" using dubious criteria. This precludes making any conclusions about the death rate as a function of distance from such stations. Fig. A and the additional images were provided by Dr. Foster.
Fig A. (l) Increase in total number of base stations in Belo Horizonte Brazil from mortality data (Fig. 16 of Dode et al); (n) cumulative number of base stations from the Anatel database. Both results normalized to their respective values in 2005.
Fig B Locations of 8 cellular base stations in the Centro neighborhood in 1999. Source: Google Maps. Centro was the district with the largest number of base stations at that time. Centro was the district with the highest population density, and with the wealthiest (and probably oldest) population. From Google Maps
Fig C Street view (Google Street View) from Av. Alfonso Pena, showing high rise apartment buildings with cell base station antennas on their roofs. The antennas are panel antenna, and direct their beams away from the building, resulting in no significant RF exposure to the residents of the building. This shows the difficulty of using distance to base station as a proxy measure of exposure as noted in the previous section on Exposure Assessment.
From Google Maps
From Google Maps
Fig D R. Cajuru, 49 - Calafate
Belo Horizonte - MG, 30480-160, Brazil
Randomly selected base station in an outlying district (the only station in that district in 1999). In contrast to base stations in the busy urban center, the antennas are mounted on a tower, in a mixed use (commercial and residential) neighborhood. From Google Maps
Belo Horizonte - MG, 30480-160, Brazil
Randomly selected base station in an outlying district (the only station in that district in 1999). In contrast to base stations in the busy urban center, the antennas are mounted on a tower, in a mixed use (commercial and residential) neighborhood. From Google Maps
Except where noted all images on this web site are taken from the Wikipedia commons
All trademarks are the property of their respective owners Copyright 2009 EMF & Health
All trademarks are the property of their respective owners Copyright 2009 EMF & Health