Biomedical Engineering Reference
In-Depth Information
may also vary across multiple attributes, for example, a physician's uncertainty
about a new drug's effectiveness might be higher or lower than his uncertainty on its
side effects; as a result, the impact of uncertainty reduction on consumers' choice
may be very different across these attributes. A pharmaceutical fi rm may want to
focus its marketing communication on an attribute that physicians are most uncer-
tain about and/or it has an advantage over others. This is an important strategic
question in product positioning. To address this question, we need to understand the
relative effi ciency of alternative information sources in facilitating physician's
learning of a drug's effectiveness and side effects, which calls for the use of pre-
scription choice data under the real market environment rather than the clinical trial
data alone.
Lastly, while pre-marketing clinical trials may be able to provide adequate evi-
dence about the effi cacy of a new drug relative a placebo or an existing drug, the full
safety profi le of a new drug is rarely known at the time of approval by the FDA
(Hiatt
2006
). This is mostly due to the limited number of participants and short
duration in clinical trials. For example, even though the larger Phase III clinical tri-
als typically involve several thousands of participants over a few years period, it
may still be diffi cult to detect rare side effects when treating chronic conditions.
Furthermore, clinical trials have become increasingly expensive because of the
growing competition for participants and reliable contract research organizations
(CROs). According to a recent survey from Cutting Edge Information (
2011
), in the
3-year period from 2008 to 2011, the average per-patient trial costs have risen 70 %
across all development phases, with the largest increase of 88 % (from $25,280 per
patient in 2008 to $47,523 in 2011) in Phase IIIa trials. Under this situation, expand-
ing the scope of clinical trials may be fi nancially prohibitive for pharmaceutical
companies. In contrast, post-marketing prescription data provides an ample, effi -
cient, and economical alternative for researchers to study rare side effects after the
drug is being marketed (Okie
2005
).
From a researcher's perspective, the main issue of using post-marketing pre-
scription data is that treatment conditions are not controlled and patients who use
different drugs are not randomly assigned. Consequently, one has to make model
assumptions to infer the true effectiveness and side effects of prescription drugs on
patients, and how these treatment outcomes are evaluated by physicians and patients.
6.3.2
A Review of the Learning Literature
in Pharmaceutical Studies
A large number of studies in marketing and economics have investigated how physi-
cians and patients learn about product attributes in the context of prescription phar-
maceutical market. Ching (
2010
) constructed a model where physicians learn about
the quality of a generic drug through patients' feedback in the presence of consumer
heterogeneity in price sensitivity. He then estimated the model using aggregate sales
data for 14 drugs with patent expired during the 4-year period from 1984 through 1987.
Search WWH ::
Custom Search