24.7 Epidemiology in drug
registration and licensingDuring the registration process, there is typically
repeated interaction between an NDA/PLA spon-
sor and the regulatory authority. Often these inter-
actions revolve around whether the tolerability of
the new drug is sufficiently well characterized, and
the criteria for the inclusion and exclusion of par-
ticular observed intolerabilities in labeling, as well
as the weight that should be applied to each (e.g.
AE list, warning, contraindication or, rarely, pre-
cluding drug approval). Often the question ‘What
level of risk is acceptable?’ becomes quickly
answered with ‘Acceptable, compared to what?’
These considerations are clearly based on an
insight into public health decision making and
are often well served by inclusion of epidemiolo-
gists working with and/or consulting with regula-
tors and sponsors.
Furthermore, the scientific proof of a negative is
virtually impossible when fewer than an infinite
number of patients have been studied. The more
elusive problem, then, is to define acceptable un-
certainty. Some of these decisions are based on
precedent rather than observation. Pharmacoepi-
demiology is often a useful source of precedents, as
well as available for further, future study in the
post-approval period to clarify residual questions
and/or reduce the remaining uncertainty.
24.8 The emerging world of risk
managementAlthough many of the activities described can then
be orientated towards support of registration with
fair and balanced labeling or education that is
useful to the prescriber, recently regulators have
requested and sponsors have proposed programs of
‘risk management’ under circumstances in which
an unacceptably serious risk has been identified or
can reasonably be anticipated and risk factors or
situations can be specified which, if addressed, can
reduce that risk to a level which, balanced against
anticipated benefits, could be accepted for clinical
use. To manage such situations, sponsors are asked
to develop, test and field interventions with the
provider, distributor or consumer to accompany
the marketplace activities with a drug with a resi-
dual safety concern. Recently, FDA has issued
official guidance regarding action programs to
minimize such risks (RiskMaps) (US FDA, 2005)
and theEuropean drugregulatory authority, EMEA
has followed suit (EMEA, 2005). Expectations
regarding these programs likewise include docu-
mentation of the effectiveness of the intervention,
once again a challenge for the public health
researcher, that is, the pharmacoepidemiologist.Post-marketing surveillance studies
Approval (especially in the United States) to intro-
duce a drug into the market is now often contingent
on the agreement of the sponsor to conduct one or
more post-marketing surveillance studies. Typi-
cally conducted ona scale of5000 or more patients,
the design of such studies poses classic epidemio-
logical challenges: the choice of control cohorts (if
any), appropriateness of historical controls, power
calculations, the nature and range of confounding
variables among others. Many of these may be
addressed using the databases described above. It
should be noted that post-marketing surveillance
studies are often implemented by companies with-
out any imposed regulatory requirement, simply
due to the value that they bring in understanding a
new product that may formerly only have been
tested in several hundred patients. The strategic
or forward thinking company will initiate such
studies well in advance of the appearance of a
signal of a potential problem, recognizing that
large-scale and long-term studies need to be in
placebeforea problem emerges if they are to be
of use in clarifying the extent and nature of that
problem in time to be of use, particularly in a
closely regulated and often adversary environment.General pharmacovigilance
Whether or not a post-marketing surveillance
study is used, all drugs undergo pharmacovigilance
when in the marketplace. This can be especially308 CH24 PHARMACOEPIDEMIOLOGY AND THE PHARMACEUTICAL PHYSICIAN