The Successful Launch and Diffusion of New Therapies - Innovation and Marketing in the Pharmaceutical Industry - page 199

component is a function of a vector of physician and/or market covariates that affect
the adoption hazard rate. Thus, y (

X pt ) adjusts h p 0 t up or down proportionally to

reflect the effect of the covariates.
The post-adoption stage can be modeled as a physician-level choice process,
where P pjt represents the probability that a physician p chooses drug j ( j = 1, …, J )
at time t , conditional on drug j 's adoption by physician p by t . This probability can
be specified as a multinomial logit model:

V

e

pjt

P pjt = (7.9)

∑

J

V

e

pjt

j = 1

where V pjt is the deterministic part of the utility obtained from choosing drug j at
time t . V pjt can be specified as a function of a set of covariates that can characterize
the drug, the physician, or the combination of both, and a time dynamic element
affecting physician choice of the new drug (Coscelli and Shum 2004 ). To explain
the dynamic adoption process, at least some covariates in the two model functions
must vary over time.

7.1.6 Conjoint Analysis

The methods we have reviewed so far only use observed data either from the new
treatment's own prescribing behavior or from the past prescribing behavior of other
drugs that have been available in the market for a longer time. They do not use so-
called primary data. Nonetheless, the use of primary data in the estimation of the
commercial potential of a new pharmaceutical may yield valuable insights. Conjoint
analysis is a particularly useful method to assess physicians' and patients' prefer-
ences and unmet needs before the launch of a new drug. Conjoint analysis is a
method to estimate the structure of consumers' preferences, given their overall eval-
uations of a set of alternatives that differ with respect to several attributes. The main
advantage of this research tool is that it can be used before a new product enters the
market. Since its introduction (Green and Rao 1971 ), conjoint analysis has been
widely adopted by marketing scientists and practitioners as a method for preference
measurement. Conjoint analysis can assist firms in developing and launching new
products, as it can be used to integrate knowledge on potential adopters' expected
reactions to these products. This ability facilitates prelaunch sales forecasts for a
new product, thus avoiding the high costs and time investments required for the use
of test markets. Conjoint analysis is most appropriate when new levels of attributes
are being introduced or when new attributes can be described well to potential cus-
tomers (Urban et al. 1996 ). Figure 7.2 describes the different steps in setting up a
conjoint study.
The basic premise of conjoint analysis is to present physicians or patients with
several variations of attribute levels for a new product and to assess their choices,
rankings, or ratings. This is typically done in a survey setting (Cattin and Wittink