Using meta-analysis and public domain cancer incidence data to identify the smoking gun. Is it RF (radio frequency) technologies or is it other features of modern living? By Dr Chris Barnes, Bangor Scientific and Educational Consultants e-mail






There are some 5000 or so publications in the public domain documenting the effects of electromagnetic radiation on biological systems from the sub-cellular to whole organism level, including human kind. Excitation frequencies range from Hz to THz.   Over 3000 of them published up until 1981  are available at NSCEP EPA 1981 [1]. 

Further interest has come in recently years because of the explosion in radio frequency technologies and because we live in a more and more safety and blame conscious age.



Despite such a large number of publications overall and with a few hundred of these suggesting that RF might be either geno-toxic, see for instance  Lai (2104) [2]  a co-initiator, a promoter or a co-promoter of cancer there have only been a very limited number of epidemiological studies attempting to associate RF and specific cancers. RF emissions from TV transmitters have pervaded the space of British homes since about the 1950’s. Since, presumably, more data was available, Radio and TV Broadcast transmitters are the subject of most of these available epidemiological studies with the exception of one German study which covers more recent mobile phone technology.  The WHO has concluded that although evidence in unclear, further studies are needed [3].



Perhaps because Leukaemia has been associated with low frequency fields, even though this was a correct epidemiological association in about 7%  of studies, see for example Pellisari et al [4] at mains frequencies (50 Hz and 60 Hz), those studying RF attempted to make the same or similar associations, see for example Havas [5]. Other cancers such as Brain cancers, especially gliomas have been associated with RF fields. An association has also been made with malignant melanoma. Some of the studies have suggested that AM and Pulse Modulated RF emissions are more dangerous than CW and FM. Until recently (2010 in Wales) TV transmitters were analogue emitters of vestigial sideband, a sort of AM and also transmitted frame synchronisation pulses at 25 Hz. GSM pulses at about 8 Hz and at 217 Hz. TETRA has some pulse amplitude content at 17 and 70 Hz. DECT pulses at 100 Hz.


Very recently indeed, I have suggested there be an association with breast and prostate cancer on the basis of these body parts and organs being able to act as dielectric antennas. Furthermore I have shown that within certain geographic locations in Gwynedd, the victims of almost all newly diagnosed cancers that have been drawn to my attention live in houses which are exposed to higher than the average combined VHF, UHF, Microwave electric field for that particular area. Of course this does not prove that these are the only diagnosed cancers in the area not does it prove cause.  The sample is far too small.  It does however signal caution and ring alarm bells.   These works (Barnes 2012, 2013, 2014) can be at  [6].




Perhaps because mechanisms for the geno-toxicity of pulsed RF are very complicated requiring interdisciplinary Physics/Engineering and Biological Sciences knowledge to comprehend and because they have only recently been demonstrated experientially this has led to scepticism as to the implication of RF in cancer biology at all. Also the epidemiological studies have only produced weak and variable associations and have thus been similarly criticised.



Recently, Barnes (2013) [6] has shown that far more sense can be made of the epidemiological studies if the effect of co-promoters such as environmental pollution is factored in. Also a special statistical evaluation using a derivative of the electromagnetic AB (  Aaronhnov Bohm) effect gives a much fuller explanation of the results. When such factors are taken into account much higher relative risks commensurate with those of Cherry’s Sutro study [7] are seen.



An additional or alternative mechanism linking RF and cancer biology is down to that of melatonin deficiency as a result of sleep disturbance either directly caused by RF or by some other facet of modern living which has proceeded at the same or very similar rate to that of the expansion of RF technology. For example the proliferation of nightlight can be highly relevant to breast and prostate cancer, see Barnes (2014) [8].  





It is proposed to use linear correlative meta-analysis to test two hypotheses arising from the above.


Firstly, if certain cancers are associated with RF (or equally with some other common factor), they ought also to be closely correlated with each other.


Secondly, the historic incidence of ‘RF’ type cancers will be correlated with historic rate or coverage development of RF technologies such as TV, digital TV or mobile phone. Also the correlation will be inspected of these cancers with other modern living factors such as night shift working or night noise or night light pollution all of which can reduce melatonin and increase certain cancer rates. The meta-analysis showing the best correlation will be significant.






Strikingly all the cancers which have previously been associated with RF by others or by me give very good or excellent correlations against each other in the period 1971-2002 (which is the most available period of data in the public domain).


A few examples are given below:
Above: Figures 1-3


All three of the above examples produce very good linear regression factors in excess of 0.91 and some approach .97. If attempts are made to correlate other 'non-RF cancers' such as, for example, stomach or lung they fail.
Figure 4


Data on areal coverage of UK TV was not readily available. If one makes the assumption that the broadcasters progressively increased coverage terrestrial to make sure all users of TV could receive a signal at a rate comparable with the uptake of TV sales then the above meta-analysis is valid.


The linear regression factor is .96 and is very similar for the other cancers described herein as 'RF cancers' malignant melanoma has been previously reported by others to be an RF cancer. It too gives a regression factor in XS of .96 when treated in this manner.


From 2001 onwards the incidence rate of malignant melanoma seems even steeper as it does for prostate cancer. One needs to enquire if this is the result of mobile phone and /or wifi and/or TETRA and/or DECT type technologies?
Figure 5


Per capita energy consumption was chosen as a possible factor that might affect cancer by factors like increased all night working, night noise etc. It appears there is only a very weak positive correlation here.

Figure 6


The above data is taken from a combination of 'RF' cancers Prostate and Breast since 1947.

It was decided to see if 'RF sensitive' cancers could be seen to correlate any further back in time. Cancer cohort rates had to be taken from a compatible European country, namely Sweden since UK data could not be readily found. These were used to extrapolate UK data in proportionality backwards from 1971 to 1947. However, it is considered valid to do this since TV broadcasting has developed more or less at parallel rates in the two countries. These 'RF' rates were correlated against the total number of TV licences issued over time, see figure 6.

A virtually perfect linear correlation is obtained. The correlated cancer data is taken from a combination of 'RF' cancers Prostate and Breast since 1947.

Again one is making the reasonable assumption that the more TV licences granted the greater is the areal signal coverage. An interesting complication is that in the period 1947 -1969 all VHF broadcasting was used with a 405 line standard.

It has previously been postulated that it is the synchronisation pulses in a TV transmission which are associated with cancer, not its frequency per se. These present results seem to support that finding.


Conclusions and Discussion

The historic incidence rate of several different types of cancer in the UK including: breast, prostate, leukaemia and brain have been shown to be very closely correlated with each other in the period 1971-2002. These are all cancers which have been correlated by various authors with electromagnetic fields of various frequencies from ELF to microwave. Not all the observed correlations have been convincing. It is believed these present results show that at least for UHF TV frequencies the results show a very convincing correlation of cancer (leukaemia) incidence r~0.9 with penetration of UHF TV signal coverage if fairly assumed to be based on number of TV sets.


Furthermore in the period 1947-2002 for combined incidence of Breast and Prostate cancer a near perfect correlation R>.99 is obtained for cancer incidence versus cumulative number of TV licenses issued. Assuming all the change in risk is due to TV and TV frequency broadcast technology a relative risk (RR) of approximately 5 results. Furthermore for R> .99 and even with only the four degrees of freedom employed, the two-tailed P value equals 0.0001 which by conventional criteria, this difference is considered to be extremely statistically significant.    The RR  is substantially greater than the RR in the Dolk 1 study (1997|) [9]  but comparable with some of the RR's quoted by Cherry for the Sutro TV tower study [7].


One possible criticism of this type of analysis is that the TV sets used between 1947 and the introduction of plasma and LCD technologies contained CRTS (cathode ray tubes) which were sources of X-rays. They also contained very high voltages between 11 KV and 27 KV which could concentrate particulates from the air, including any potential radon and daughters. They could also produce electromagnetic radiation in the Hz and KHz frequency ranges, which too has been associated with cancer in its own right. Exposure to these scenarios might also be expected to impact cancer rates. However, one would expect concentrations to be within the TV set itself and X -rays mainly to be radiated backwards towards the internal protective screen shield.  Nevertheless, there is evidence of the bio-damaging effect of CRT MONITORS in the literature, see for example, Carbonari et al 2005 [10] who found increased micronucleated cell frequency related to exposure to radiation emitted by computer cathode ray tube video display monitors. See also from more recent literature Lakshmi (2010) who discuss DNA damage and TV/CRT monitors [11].

The flat screen LCD TV was invented in 1983 but was so expensive that it did not penetrate many UK households until the mid to late 1990's. Worldwide LCD use has overtaken CRT use since 2007.

But adding to the complexity, from 1998 onwards several new technologies began to emerge in parallel including analogue mobile phones and by the early 2000's, digital TV broadcasting and digital mobile phones.


If CRT TV's had been impacting badly on cancer incidence, one would expect to see a decline in incidence rates after the mid 1990's with a possible lag of up-to five years. On the other hand if new RF (radio frequency) technologies impacted cancer incidence rates over and above those due to TV transmission signals then one would expect an increase.


It is possible to search for such effects by applying a polynomial data analysis. This has been done for

Leukaemia figure 7 and Breast cancer figure 8.

Figure 7 above.
Figure 8 above.



The Leukaemia incidence polynomial predicts a peak in incidence in 2005 and now a decreasing trend.

Similarly Breast cancer appears to saturate in 2012 with a possible hint that this too will start to diminish.

Since Breast cancer has a longer latency period than Leukaemia, 8-15 years as opposed to 5 years it would seem the change leading to lower incidence may have occurred between 1998 and 2004 for breast cancer and around 2000 for leukaemia.


Possibly the same process or multiple processes are at work here. In the UK most people would own a LCD TV by the early 2000's and CRT monitors in offices and industry are now almost exclusively phased out.

It is tempting to suggest that part of the increase in cancer rates over the period was due to CRT use and part due to RF technologies. Thus that the rate of change  has become a slowing increase or possibly even reversed to become a decline  since the use of CRTS has died out.

Destaillts et al (2008) [12] has suggested that CRT devices can also seriously increase the load of indoor hazardous atmospheric pollutants which can of course also encourage tumouri- genesis. Furthermore, Kumar et al  (2012)  [13] have found harmful effects in blood tissues at distances as much as 50 cm from CRT Monitors and TV sets.


There is one other possible explanation and that is that as multiple RF technologies have appeared since the late 1990's some combination of all these different frequencies, modulation types and rates may actually be starting in some way to counter the adverse effect of just TV and radio broadcast emissions. For instance CDMA emissions might not be as biologically damaging as GSM, see Moskowitz(2011) [14], nor is CDMA as disruptive to the operation of the brain as GSM, see Tyagi et al (2011) [15].


 It is my present suggestion that as use of wireless technologies absolutely starts to reach saturation point, there will be so many different frequencies and modulation schemes bombarding the human body it may begin to simply interpret them as noise. Electromagnetic noise superimposed on microwave radiation has been shown to block DNA damage and ROS formation, see Yao et al (2008) [16].  


Whichever of the two above is the correct explanation remains wholly to be seen.  However it is highly instructive and perhaps critically imperative to note that a similar stagnation in cancer rates has been noticed in the USA since 2003, [17]. The USA as well as the UK will have phased out CRT TV and now leads Europe in fast 3G technology by a wide margin, .

  In the USA both CDMA and QAM modulation schemes are used but the QAM bit rates are so fast that they would in my personal opinion  probably not be expected to provoke as much biological response as early GSM systems.



  1. Breast, prostate, leukaemia and brain cancers have been shown to be very closely correlated with each other and with the growth of TV use/broadcasting in the period 1971-2002, but not with general per capita growth.
  2. The sum total of incidence of prostate and breast cancer combined in the period 1947 to 2002 is extremely closely correlated with the number of TV licenses issues in the UK R>.99. This is suggestive of an extremely close association with TV use and/or TV broadcast emissions.
  3. Use of polynomial expansion to predict cancer incidence beyond 2010 suggests a long terms saturation for Leukaemia incidence trend at circa 2005 and a fall thereafter, similar results are obtained for breast cancer in the medium term with about a 6 year lag.
  4. Reasons for (3) above may be use of CRT TVS and monitors coming to an end or many competing wireless technologies actually reducing risk -further research is urgently needed but will not be possible until relevant data falls into public domain.


Further work

To confirm the association between various cancers, the ultimate cohort is to employ World statistics.

A significant number of developing countries in the world did not begin TV broadcasting until much later than the UK and some have very poor terrestrial signal penetration and use satellite systems instead. However, what is known very accurately is what the mobile telephone coverage is in developing countries and all are investing heavily at present. It is prosed to explore the effect of this technology in the near future. Further with the exception of liver and cervical cancers many cancer rates are much lower in developing countries. Geographic meta-analysis of factors which could protect from cancer and from RF associated cancer will also be explored.









1.      NSCEP EPA INDEX  EPA-600/9-81-011\ZYFILES\INDEX%20DATA\81THRU85\TXT\00000024\9101FEXP.txt&User=ANONYMOUS&Password=anonymous&SortMethod=h|-&MaximumDocuments=1&FuzzyDegree=0&ImageQuality=r75g8/r75g8/x150y150g16/i425&Display=p|f&DefSeekPage=x&SearchBack=ZyActionL&Back=ZyActionS&BackDesc=Results%20page&MaximumPages=1&ZyEntry=2