Biomedical Engineering Reference
In-Depth Information
controlling for various drug quality dimensions, he finds significant evidence for the
impact of clinical results on prescription behavior. In addition to the demand model,
he specifies a model of detailing supply. It turns out that the level of detailing
increases with the level of useful clinical research output. Accounting for the medi-
ating effect via detailing in the demand model, drug sales respond with a total elas-
ticity of 0.3-0.5 to the level of scientific evidence.
Venkataraman and Stremersch (
2007
) use physician-level data to analyze the
moderating effects of drug effectiveness and drug side effects with respect to mar-
keting efforts on drug prescription and sample dispensing. Their prescription model
is a truncated regression model that assumes a negative binomial distribution (NBD)
of prescriptions and allows for heterogeneity in physician responsiveness.
Specifically,
a
k
l
al
Γ
ΓΓ1
(
a
+
k
)
a
al
(
)
=
pjt
Pr
RX
=
k
|
l
(19.1)
pjt
pjt
()(
a
k
+
)
+
+
pjt
pjt
with
(
)
ln
l
=
ln
ERX
pjt
|
Z
pjt
x
pjt
=
b
+
b
SE
+
b bb
Eff
+
+
SE
t
+
b
EffDet
pjt
,
p
jt
j
p
j
j
pjt
p
1
p
2
p
1
p
2
(
)
+
bb b
+
SE
+
EffMeet
+
x
pjt
Q
+
x
p
p
1
jt
p
2
j
pjt
pjt
where Pr(
RX
pjt
=
k
) is the probability of
k
new prescriptions of brand
j
written by
physician
p
in period
t
. λ measures the mean prescription rate and α the degree of
overdispersion. The mean description rate depends on λ, the drug's number of side
effects, SE, and its effectiveness, Eff, which both moderate the influence of details,
Det, and physician meeting attendance, Meet, on prescription behavior. Other vari-
ables are summarized in the vector
x
; β and Θ are parameter vectors to be estimated
and ξ represents an error term. The sample-dispensing model is specified in a simi-
lar fashion (p. 1694).
The study finds that marketing responsiveness in terms of new prescriptions and
sample dispenses is higher for drugs that demonstrate a greater effectiveness. This
suggests that sales reps may use scientific evidence to improve the impact of a sales
call. The study also shows a positive moderating effect of the number of side effects
with respect to marketing efforts. The authors explain this finding with the informa-
tion effect that a detailing call helps reduce the physician's uncertainty about the
side-effect profile of a drug. While estimation results inform about the moderating
role of drug effectiveness for marketing responsiveness, Albers (
2012
) remarks that
the exponential conditional mean function does not allow optimizing marketing
expenditures. Hence, using this model to derive normative budget implications is
not possible.
Competitive interaction
. Many firms compete for market share and profits in the
global pharmaceutical industry. It seems reasonable to assume that marketing
activity levels and market outcomes are not independent of competitive activity. As
Search WWH ::
Custom Search