In epidemiological studies, the distribution and determinants of health
events of populations are studied, to try to determine the cause of
are two main types of epidemiological studies:
Case control: This type of study starts with the identification
of a group of persons with the disease of interest, and uses for
comparison a suitable control group of persons without the disease.
The two groups are then compared with respect to possible causative
Cohort: In this study a population (i.e. a cohort) is defined according
to the presence or absence of a factor that might influence the
probability of occurrence of a given disease. The cohort is then
followed to determine if those exposed to the factor are indeed
at greater risk of the disease.
The design, execution and interpretation of epidemiological studies
of the association between human health, or disease, and EMF exposures
are difficult for a number of reasons (Phillips, 1998; Savitz, 1993;
UNEP/WHO/IRPA, 1993, Swerdlow, 1999, Ahlbom, 2004).
RF far-field radiation emitted from a mobile phone base station
can be affected by many factors. These include the type of mobile
phone service, the antenna characteristics, the height at which
the antenna is sited, the number of proximal antenna sites, the
number of users in the area, and structures or buildings that may
impede the energy. An example of the variability that can exist
in different locations was seen in the Health Canada (1997) study
of RF field measurements of Vancouver-area schools that had either
a PCS or an analog cellular base station antenna nearby. Health
Canada found that maximum outdoor power density recorded varied
from 230 to 6,200 times below the exposure limit specified by Safety
Code 6, and that the maximum indoor measurements varied from 4,900
to 59,000 times below the limit. To obtain accurate far-field exposure
data for an epidemiological study, multiple individual measurements
of each antenna site in question should be made.
It is also very difficult to obtain accurate exposure assessments
from near-field RF sources such as a hand-held unit. Power absorption
from the antenna of a hand-held wireless telephone is very heterogeneous,
and is dependent on a number of physical factors related to the
power level of the RF signal. These include the distance of the
user from the base station, the interference of the signal by buildings
or other structures, and the direction the user is moving in relation
to the antenna (ICNIRP, 1996, Ahlbom, 2004). The amount of RF absorption
is also dependent on the duration of use, the number and length
of individual calls, and any other 'individual habits of use' (Rothman,
1996). The latter include the angle at which the phone is held and
the tendency to use one or other side of the head.
Another challenge in observational studies is the selection of appropriate
control groups. Most individuals are exposed to varying amounts
of background EMF depending on their history of use of electrical
devices at home and work, and proximal location to telecommunications
transmitters or electrical power distribution sources.
It is very difficult to establish EMF exposure in individuals over
a meaningful period of time.
Reconstruction of an exposure history without having direct measurements
requires a number of assumptions, which may or may not be valid.
bias is common in case-control studies. This occurs when there is
a systematic error due to differences in accuracy or completeness
of recall of prior experiences. For example, brain tumour patients
may be more likely than controls to misrepresent their use of a
bias may also be present (Lahkola, 2005). This occurs when there
is a systematic difference in characteristics of those agreeing
to participate in a study and those who do not.
A major problem is the control of confounding exposures (other factors
that may influence an outcome and distort an apparent effect).
The temporal relationship between exposure and disease is often
The relatively recent advent of wireless technologies makes it more
difficult to detect an adverse health outcome with a long latent
period, such as most forms of cancer.
Many of the cancers studied are rare diseases and therefore it is
large sample sizes,
prospective cohort studies because of the need to observe a
large number of individuals over time to obtain a statistically
adequate number of cases,
specific subtypes of cancers (e.g., astrocytomas, gliomas) that
might have different etiologies in relation to electromagnetic
A detailed examination of the issues surrounding epidemiological studies
of health effects of base stations can be found in:
Feasibility of future epidemiological studies on possible health effects
of mobile telephone base stations (Neubauer, 2007).
this section we review studies on the association between RF radiation
and health effects, mainly cancer.
All the studies suffer from the major weakness that exposure assessment
is imprecise. The difficulties with exposure assessment in radiofrequency
radiation studies are discussed in another section. The striking
finding in the cancer studies is the lack of consistency with regard
to cancer types or strength of association.
Reviews of epidemiological studies were conducted by Elwood (1999,
2003) and by Moulder et al. (1999). Breckenkamp (2003) reviewed
cohort studies involving potentially high-exposed occupational groups.
The references to these papers are given in the "Bibliography"
section. Other reviews were undertaken by the Expert Panel of the
Royal Society of Canada (www.rsc.ca)
and by the Independent Expert Group on Mobile Phones (www.iegmp.org.uk).
A review of primary brain tumours (Wrensch, 2002) stated that "studies
of cell phone use
. have found little to support a causal
connection with brain tumours; caveats remain". Other reviews
have been conducted by Jauchem (2003), Kundi (2004), Ahlbom (2004),
and Moulder (2005) ), and Ahlbom (2005) . Kan (2007) conducted a meta-analysis of cell phone use and brain tumour.
Elwood in his latest review (2003) offers this overall assessment:
epidemiological results fall short of the strength and consistency
of evidence that is required to come to a conclusion that RF emissions
are a cause of cancer. Although the epidemiological evidence in
total suggests no increased risk of cancer, the results cannot
be unequivocally interpreted in terms of cause and effect. The
results are inconsistent, and most studies are limited by lack
of detail on actual exposures, short follow-up periods, and the
limited ability to deal with other relevant factors. In some studies,
there may be substantial biases in the data used. For these same
reasons, the studies are unable to confidently exclude any possibility
of an increased risk of cancer. Further research to clarify the
situation is justified."
Kundi et al. (2004) agree that all studies have some methodological
deficiencies, but their overall conclusions are somewhat different:
"...all studies approaching reasonable latency found an increased
cancer risk associated with mobile phone use".
The studies reviewed in this section can be grouped under the headings
"Studies examining cell phone use", and "other epidemiological
studies". The latter is divided into "studies of general
populations", "Studies of military personnel", and
"studies of other occupations". All the research papers
are listed in the reference section at the end of each subsection,
in the alphabetical order of the first authors.