Mobile Phones and Cancer - George Carlo
The Latest Reassurance Ruse About Cell Phones and Cancer
After a number of prominent, peer reviewed studies indicating that cell phone
radiation can cause genetic damage, brain and blood cell dysfunction and a host
of health problems including cancer, now comes a Danish study that appears to
say there is no danger at all. And, that is exactly what cell phone users want
to hear - if it is true.
The problem is that the new Danish cohort study does not support the
reassurances that have been ascribed to it. It is a ruse based upon a program
initiated by the telecommunications industry more than a decade ago to control
the global scientific research agenda concerning cell phones and health effects.
The industry strategy has been to fund low-risk studies that will assure a
positive result - and then use it to convince the news media and the public that
it is proof that cell phones are safe. Even though the actual science proved
nothing of the sort. It is against this backdrop that the Danish study provides
an illustrative case history.
In my training as a professional epidemiologist, I learned early in my career
how to properly design a study to produce a valuable finding. In the process it
became evident that it was also possible to skew a study in order to produce a
pre-ordained outcome. And this is what appears to have happened in this Danish
epidemiological study. This study, funded by the telecommunications industry,
was clearly created in order to produce a positive, low-risk finding.
A low risk study is one that has minimal opportunity to conclude with results
contrary to the interests of the industry. Epidemiological studies are targets
because they are observational in nature instead of experimental (the subjects
are not manipulated according to a study protocol) and there are methodological
shortcomings in all study designs. It is therefore possible to design studies
with pre-determined outcomes that still fall within the range of acceptable
science. Thus, even highly flawed epidemiological studies can be published in
peer-reviewed journals because they are judged against a pragmatic set of
standards that assume the highest integrity among the investigators.
Experienced investigators clearly understand the influences of various
decisions in the study process, and that is why epidemiological studies contain
very extensive discussions about methodological assumptions and shortcomings.
Because this in-depth knowledge of methods is a necessary skill for all
epidemiologists, epidemiological studies can be relatively easy to manipulate -
in design, implementation and reporting.
The Danish Cohort Study Was Epidemiologically Designed To Produce A
Pre-Ordained Positive Outcome
The study, although published in the Journal of the National Cancer
Institute, followed a pre-determined methodology that gave it little chance to
find any increase in the risk of cancer among cell phone users in Denmark. The
basis for this assertion is described below. The study was also
industry-friendly from a public relations perspective: several hundred thousand
people could be included in the study, and thus it could be packaged as a very
large study that would seem meaningful to the media. Prior to the onset of the
study, it would have been clear that, if managed properly, the study could
provide very useful data for the industry's position that cell phones are safe.
The Working Hypothesis And Statistics Were Weighted Toward Finding No Risk
The working hypothesis involved looking at a large group of cell phone users and
comparing their estimated cancer expectancy with the general population in
Denmark. The cell phone user population was heterogeneous in terms of gender,
age, socio-economic status, personal health habits and occupation. There was
nothing in the definition of who was included and who was not that was actually
relevant to cell phone radiation exposure - a cell phone user was defined as
someone who made only one call per week over a period of six months or more.
Finding a cell phone related cancer risk among this group would be akin to
identifying excess lung cancer risk among people who smoked one cigarette a week
- similar to inding a needle in a haystack.
The investigators failed to include statistical power calculations in the
report.These data would have detailed specifically how large a risk the study
could address as well as how large a statistical risk could be excluded with
this study design. A doubling or tripling in risk could still be in the data but
not revealed because the statistics were not robust enough to find it. Leaving
the power calculations out of the report opened the door for un-challenged,
openendedassurances about the safety of cell phones - which were not based at
all on data within the study.
The Danish investigators failed to explain that, even if taken at face value,
the study can only rule out statistical risks that are dramatically large. In
fact the study did not contain even one statistically significant finding that
would attest to the veracity of the headlines that it spawned. The only
statistical finding in the study had to do with latency of more than ten years
from the time people began using the cell phone to the development of their
tumor. Latency is a very poor measure of cumulative exposure to cell phone
radiation as evidenced by a dozen other published studies.
The Study Included No Reasonable Exposure Variables And Thus Created Bias
Toward A Finding Of No Risk
The study covered phones used from 1982 through 1995 - antique technology that
bears little resemblance to the phones in use today. Early phones operated with
very different power profiles compared to modern phones and thus exposure plumes
with the ability to affect biological cells were very different. In addition,
background levels of information-carrying radio waves are orders of magnitude
higher in today's environment than in the past; synergies between ambient
exposures and phone related near-fields are more severe today than during the
timeframe of the study. The investigators failed to point out these caveats and
the probability that the study has little relevance to the dangers of phones
that are being used currently by more than 2 billion people around the globe.
Without quantitative exposure variables, it is not possible to obtain a
reasonable assessment of cause and effect. In this study, all 420,000 people
were assumed to use the cell phone exactly the same way. No account is made for
time on the cell phone at all. No measures of minutes or hours per day of use
were included. The underlying premise, therefore, is that everyone in the study
used the phone for the same amount of minutes. Common sense and experience tells
us that would not be the case. This assumption of similar use produces a
dramatic misclassification of exposure. Statistically, this bias leads to a
significant underestimation of true risk because the exposure variable is
imprecise. In this case, the exposure variable has no precision at all because
it is non-existent. Thus, statistical findings of no risk would be the primary
expectation from this study design.
Further, in the early days of cell phones, time on the phone was very expensive.
People did not use a cell phone as a major source of communication. Thus, people
in the Danish study were unlikely to have used cell phones for very long each
day. It is therefore reasonable to assume that exposures across the entire
cohort were on average very low. Low exposures coupled with no precision in the
distinguishing high from low exposures compounds bias in this study design
toward the null. Thus, at the outset, it would be clear to both the
investigators and the sponsors of the study that the chances of finding any
increase in risk were very minimal.
The Study Was Not Based on Any Previous Biological Findings, so its Chances
of Finding a Biological Risk Were Exceedingly Low
A basic tenet of epidemiology is that hypotheses in studies should be linked to
biological mechanisms that will help explain the observations in the research.
Studies intended to assess risk therefore include rationale that would address,
for example, brain tissue exposed within the near-field plume emanating from a
cell phone and subsequent biological responses. That is why studies that have
provided the most information about risks in humans from cell phones have
focused on such epidemiological variables as concordance between the side of the
head where a phone is used and the side of the head where a tumor occurs. No
such rationale was included in the Danish study design so the study had no
biological basis. Thus, any finding of risk increase would be due to chance and
therefore extremely unlikely.
Further, the investigators took no account in their writing of findings already
confirmed in the peer-reviewed literature. Studies show that radio frequency
radiation causes biological cell membrane dysfunction that leads to genetic
damage, cellular dysfunction including blood-brain barrier leakage and
disruption of intercellular communication - all harbingers of serious disease,
including cancer. The working hypothesis of the Danish study remained purely
statistical and by design stayed away from anything to do with mechanisms of
disease. It is noteworthy that disease mechanism data which portend disease are
the weak point in the industry's argument that cell phones are safe.
The Investigators Limited Their Access to Data That Would Have Made the
Study More Robust, Thus Limiting the Chances of Finding Any Risk Increase
The most important data in this study would have come from people who used the
phones most. The enrollment criteria for the study, one phone call per week for
six months is likely to have created a very large group of study subjects who
had very little exposure to cell phone radiation. In the a priori study design,
all commercial phone subscribers were excluded from the study. Considering the
high expense of phone minutes during the 1980s and 1990s where the study was
focused, it is likely that commercial subscribers would have been the heaviest
users. Therefore, a selected group who would have provided important information
on any disease risks actually related to cell phone radiation was eliminated by
choice of the investigators.
Another important subset would have been those people who used their phones on
the same side of the head where their tumors were located. This would be the
only subset of phone users with tumors who would definitively have been exposed
to measurable levels of cell phone radiation in the only part of the head where
the radiation could have an effect. Those people were not culled out and studied
and thus an opportunity to learn something important about brain cancer risk was
missed. Because the investigators had access to subscriber information and user
lists obtained from the phone companies, it is reasonable to assume that type of
in-depth study could have been completed if the investigators so desired.
Internal Inconsistencies Raise Red Flags About the Methods
The most fundamental portion of a scientific report is the review of prior
literature and a clear statement of the hypothesis the study is intended to
address based on that literature. The balance and fairness with which the report
is written isreflected in the review of the literature. It is noteworthy that on
the first page of the Danish report virtually all research that indicated any
risk from cell phones was summarily dismissed. One specific quote by the Danish
investigators on page one is illustrative: "Most studies have not found a
statistically significant overall association with the risk of brain tumors for
use of 10 or more years, except two, for which methodologic issues have been
raised." That quote does not square with the data presented here in the appended
Table 1. In the peer reviewed published literature, there are more than 300
statistically significant findings indicating an increase in risk. It is
noteworthy that most of the same studies cited under Table 1 are also cited in
the Danish paper.
In this study the risks of all types of cancer in the Danish men and women were
in women. Those findings would suggest that the population of Denmark has a
comparatively low cancer risk overall. However, international cancer statistics
show that Denmark is among the top 15 countries in the world for cancer deaths
among both men and women. Some tallies show Denmark to be number one in cancer
mortality among women. The bottom line is that this study shows Danes have low
cancer risk but the world data say they have high cancer risk. The low Danish
cancer risk hypothesis does not hold up and it suggests something is wrong
within the study's data.
The apparently low site-specific cancer risk findings could be the result of a
systematic overestimate of the expected numbers of cancer cases that were
calculated in the study. That would mean that the analytical algorithm, or the
method used by the investigators for calculation, was somehow altered or flawed.
If expected numbers are artificially high, then risk estimate numbers would be
artificially low. Such a phenomenon would result in an artificial finding of no
risk in the study. Only the investigators themselves have control over that
Dissemination of the Study Results
It is not likely an accident that immediately after the Danish study was
released media outlets from around the world were somehow made aware of the new
study and the findings. It is noteworthy that the bland title of the study,
"Cellular Telephone Use and Cancer Risk: Update of a Nationwide Danish Cohort,"
bears little resemblance to the headlines it produced:
- 'Cell Phones Don't Cause Brain Cancer' - The Toronto Daily News
- 'Cell Phones Don't Raise Cancer Risk' - Reuters
- 'Big Study Finds No Link Between Cell Phones, Cancer' - SJ Mercury News
- 'Study: Cell Phones Cause Cancer' - Albuquerque Tribune
- 'Study: Cell Phones Safe' - Newsday
- 'Cell Phones Do Not Cause Cancer' - Techtree.com, India
It is also not likely to be an accident that science groups immediately and
aggressively weighed in to validate the conclusion:
- 'Cellular Telephones Not Associated With Cancer Risk' - JNCI Spectrum
- 'Cell Phone-y Scare' - American Council on Science and Health
It would strain the credulity of even the most casual observer to believe
that these media and science groups were anxiously awaiting the release of the
Danish cohort update.
That the information release was highly coordinated would be a reasonable
assumption based on the facts. However, given the weaknesses in the study and
the obvious limitations that follow from those weaknesses in terms of ruling out
a cell phone related cancer risk, it is remarkable that so many groups would
blindly run with the headlines.
However, when a study like this Danish report is released, the media need a
reason not to cover it, especially against the frontal public relations assault
that comes from the industry with the release of new data that they believe
supports their position. No one is providing those reasons to the media, so the
stories run around the globe. But, there is more going on here.
In 1993, I was asked by the cell phone industry to run a $28.5 million dollar
research effort, funded by the industry and overseen by the federal government.
The program, called the WTR, was aimed at addressing the cell phone and brain
When the initial studies seemed to indicate no problem, the industry was
pleased and supportive; but when the subsequent and more thorough studies
indicated that cell phone radiation caused biological changes, the industry
became displeased and sought to close down the program. Completed in 2000, the
WTR program remains the largest ever addressing cell phones and cancer.
Our WTR program also had a tie with the present Danish study. Back in the
1990s, two of the authors of the 2006 Danish study, John Boice and Joe
McLaughlin, applied to the WTR program for funding to do the same epidemiology
study that was released this week. When they made the proposal on behalf of
their company, the International Epidemiology Institute, both were employees of
the National Cancer Institute. That affiliation was an important part of how
they presented their credentials. After consideration of their proposal, we
denied them funding because we were not convinced they would provide meaningful
findings. We also were not comfortable with the study design that was presented
to us. The investigators put too much emphasis on the probability that the study
would not find risk increases. Because the program was funded by the industry,
they might have thought the low-risk pitch was what we wanted to hear.
When we refused to give them funding to do the work, Boice and McLaughlin
went directly to the industry with the same pitch - and they were hired.
The Danish study released this week is one of many studies from this group of
investigators - all concluding with similar findings of no tumor risk from cell
phones. In 2001, they released what they then lauded as one of the largest
studies to date, and Boice went on a television tour to blunt the effects of a
book I co-wrote with renowned Washington syndicated columnist, Martin Schram.
The book, Cell Phones: Invisible Hazards in the Wireless Age told of the clash
of science and politics within the research program and concluded that by 2001,
science had raised serious red flags about the health of millions of cell phone
The American Cancer Society is also in on the support bandwagon for the
Danish study. Now circulating to attest to the findings of the study are
comments from Michael Thun, a vice-president at the society. He has taken the
position that the Danish study confirms no risk from cell phones. It is
noteworthy that in 2002, scientists from the American Cancer Society testified
on behalf of the cell phone industry in brain cancer litigation. The case was
brought in Federal Court in Baltimore, Maryland by surgeon Christopher Newman,
who had claimed that his terminal brain tumor was the result of his cell phone
use. The American Cancer Society testified that the tumor had not been caused by
his cell phone.
Within a year of that testimony, a report was released by the American Cancer
Society that included cell phones as one of that year's greatest cancer myths.
The subsequent connection between the American Cancer Society and the cell phone
industry was arrogantly blatant. In 2005, Dr. Sanjay Gupta of CNN ran a story
that included Mt. Sinai Medical Center surgeon Keith Black. Dr. Black believed
that the tumor that took the life of his patient, famous attorney Johnnie
Cochran, was due to cell phone use. The industry did not even reply to the
questions raised in the story about the link between the tumor and Johnnie
Cochran's cell phone use. Instead, the industry submitted a written statement
that simply referred to and quoted the American Cancer Society's report
indicating that cell phones were a cancer myth. Thus, the industry was able to
use the American Cancer Society paper as a public relations shield to manage the
The world wants to know that the cell phones that they have used and will
continue to use are safe. But the Danish study was epidemiologically constructed
to produce a finding of reassurance that may well not have been supported even
by a more professionally conceived study designed to really assess risk. The
study has been trumpeted far beyond any reasonable reading of the data as proof
that cell phones are safe. This is a disservice to consumers who want to believe
that scientists and doctors can be trusted to be honest with data and to keep
them safe. The Danish registry remains a valuable resource. But, we are still
awaiting a proper epidemiological study that will be able to use that resource
to help tell the world what it needs to know.
The cell phone industry indeed has a sophisticated program in place to guard
their financial interests. That is reasonable so long as harm to an unwitting
public is not part of the result. One symptom of the effectiveness of the
program is that industry-funded studies in many cases now produce
industry-desired outcomes. But tampering with the integrity of scientists,
scientific systems and public information steps over the lines of propriety that
are appropriate for protecting business interests - especially when the casualty
of the interference is public health and safety.
Dr. George L. Carlo
Science and Public Policy Institute
Safe Wireless Initiative
Jill Unger, Milt Bowling and Martin
Schram contributed to this work
- Auvenin, Anssi; Hietanen, Maila; Luukonen, Ritva; Koskela, Riitta-Sisko.
May 2002. Brain Tumors and Salivary Gland Cancers Among Cellular Telephone
Users. Epidemiology 13(3):356-359
- Hardell, Lennart; Mild, Kjell Hansson; Sandstrom, Monica; Carlberg,
Michael; Hallquist, Arne; Pahlson, Anneli. 2003. Vestibular Schwannoma,
Tinnitus and Cellular Telephones. Neuroepidemiology 22:124-129 DOI:
- Hardell, L., Carlberg, M., Mild, K. Hansson. 2005. Case-Control Study on
Cellular and Cordless Telephones and the Risk for Acoustic Neuroma or
Meningioma in Patients Diagnosed 2002-2003. Neuroepidemiology 25:120-128.
- Hardell, L., Carlberg, M., Mild, K. Hansson. 2006 Pooled analysis of two
case-control studies on the use of cellular and cordless telephones and the
risk of benign brain tumours diagnosed during 1997-2003. International
Journal of Oncology 28:509-518
- Hardell, L., Carlberg, M., Mild, K. Hansson. 2006 Pooled analysis of two
case-control studies on use of cellular and cordless telephones and the risk
for malignant brain tumours diagnosed in 1997-2003. International Archives
of Occupational and Environmental Health DOI: 10.1007/s00420-006-0088-5
- Hardell, Lennart; Carlberg, Michael; Mild, Kjell Hansson. 2006.
Case-control study of the association between the use of cellular and
cordless telephones and malignant brain tumors diagnosed during 2002-2003.
Environmental Research. 100:232
- Hepworth, J. Sarah; Schoemaker, J. Minouk; Muir, R. Kenneth; Swerdlow,
J. Anthony; van Tongeren, Martie J.A.; McKinney, Patricia A. November 14,
2005. Mobile phone use and risk of glioma in adults: case-control study. BMJ
Online First www.bmj.com 5 pages.
- Inskip, D. Peter, Sc.D., Tarone, E. Robert, Ph.D., Hatch, E. Elizabeth,
Ph.D., Wilcoxky, C. Timothy, Ph.D., Shapiro, R. William, M.S., Selker, G.
Robert, M.D., Finje, A. Howard, M.D.,Black, M. Peter, M.D., Loeffler, S.
Jay, M.D., Linet, S. Martha, M.D. January 11, 2001.Cellular-Telephone Use
and Brain Tumors. New England Journal of Medicine. www.nejm.org 344(2):79-86
- Kundi, Michael; Mild, K. Hansson; Hardell, Lennart; Mattsson, Mats-Olaf.
2004. Mobile Telephones and Cancer - A Review of Epidemiological Evidence.
Journal of Toxicology and Environmental Health, Part B. 7:351-384 DOI:
10.1080/10937400490486258 ISSN: 1093-7404 print/ 1521-6950 online
- Lonn, S; Ahlbom, A; Hall, P; Feychting, M; Swedish Interphone Study
Group. March 15, 2005. Long-term Mobile Phone Use and Brain Tumor Risk. Am.
J Epidem. 16(6):526 DOI:10.1093/aje/kwi091
- Muscat, Joshua E., MPH; Malkin, Mark G., MD, FRCPC; Thompson, Seth, PhD;
Shore, Roy E., PhD; Stellman, Steven D., PhD; McRee, Don, PhD., Neugut,
Alfred I., MD, PhD; Wynder, Ernst I., MD. December 20, 2000. Handheld
Cellular Telephone Use and Risk of Brain Cancer. Journal of the American
Medical Association 284(23):3001-3007
- Muscat, J.E.,Ph.D.; Malkin, M.G., M.D., FRCPC; Shore, R.E., Ph.D.;
Thompson, S., Ph.D.; Neugut, A.I., M.D., Ph.D.; Stellman, S.D., Ph.D.;
Bruce, J., M.D. 2002. Handheld cellular telephones and risk of acoustic
neuroma. Neurology 58:1304-1306
- Shuz, Joachim; Jacobsen, Rune; Olsen, Jorgen H.; Boice, John D. Jr.;
McLaughlin, Joseph, K. and Johansen, Christoffer 2006 Cellular Telephone Use
and Cancer Risk: Update of a Nationwide Danish Cohort Journal of the
National Cancer Institute 98(23):1707-1713http://www.canceractive.com/page.php?n=1541