physicians need an understanding of the funda-
mentals of this field, in order to understand
and harness the value that epidemiology, and
epidemiologists, can bring to drug development
and product surveillance programs. Epidemiology
is taught in all schools of public health and, in
varying depth and quality, in schools of medicine.
Epidemiological techniques are used by many
people who would not describe themselves as
epidemiologists. Board certification in preventive
medicine requires a Master’s degree with a large
epidemiology component, and further tough
examinations.
Such epidemiology training emphasizes obser-
vational research methodology as the core
approach of the field. However, emphasis is also
given on building expertise in clinical trials design
and biostatistics. These disciplines require exper-
tise and experience in the management of huge
quantities of data and the attendant expertise in
scientific computing/informatics. These are skills
that find natural places in phase III and phase IV
clinical study design and conduct within industry,
and evaluation within the regulatory environment.
However, it is in the understanding of the applica-
tions (and often more importantly the limitations)
of the nonexperimental/observational method that
the epidemiologist brings special value addition to
the pharmaceutical sector.
It is important to remember that epidemiology
represents another set of tactics to address the same
underlying motive as others working in and with
the development enterprise. Just as much as a
molecular biologist or clinical pharmacologist,
the epidemiologist is trying to find out which set
of conditions causes a particular disease or benefit
or adverse event (AE). The additional perspective
of the impact on actual populations (the actual
effectiveness) complements the emphasis on the
experimental subject (the efficacy) of much of
clinical research. The epidemiologist is faced
with the substantial challenge of observational
approaches. Without the benefits (comforts) of
randomization and blinding afforded by the experi-
mental method, only rarely can the epidemiologist
imitate the pharmacologist, who can premeditate
an intervention in a confined population, and then
prospectively observe its effect. However, even
when constrained by the observational approach,
the epidemiologist is like other scientists in that
findings are in the context of comparison among
various structured observational groups, differing
in their known exposures or outcomes (Strom,
2005; Hartzema, 2006).^124.2 Epidemiological
methodologiesProspective cohort studies
A prospective cohort epidemiological study appro-
ximates to a parallel-group clinical trial in its
scientific basis, and epidemiologists will be as
aware as clinical trialists of the bias that can be
introduced if the study groups do not contain com-
parable, well-balanced and homogeneous groups
of people. Although the experimentalist uses
exclusions, randomization and blinding as tools
to control for unseen biases, the epidemiologist
is, rather, required to measure and document attri-
butes and control for those that may lead to skewed
results, by selection in ascertainment and stratifi-
cation in analysis. Furthermore, like others calling
themselves drug surveillance specialists, the epi-
demiologist will be well aware that the size of
the groups that must be studied increases with the
rarity of the phenomenon that is sought. The
latency of the effect (e.g. the duration between
exposure to an unsuspected atheromatous stimulus
and coronary artery disease) can define the desir-
able duration of follow-up, in a manner analogous
to the study of the probability of AEs arising only
after prolonged multiple-dose drug exposure.
Often, rather, the size of the available population
and the duration for which it has already been
followed (e.g. for other, administrative or clinical
purposes) will dictate the extent to which an obser-
vational study is able to state the level of certainty
of its observations.(^1) For a more extensive discussion of the field of pharmacoe-
pidemiology, the reader is referred to the two most widely
cited textbooks in the field.
304 CH24 PHARMACOEPIDEMIOLOGY AND THE PHARMACEUTICAL PHYSICIAN