Modeling the Effects of Promotional Efforts on Aggregate Pharmaceutical Demand: What We Know and Challenges for the Future - Innovation and Marketing in the Pharmaceutical Industry - page 592

appropriate communication tools to generate awareness for under-treated ailments,
and also discourage the use of certain instruments for categories where pharmaceuti-
cal treatment is less effective or where additional treatment is deemed unnecessary,
i.e., to prevent overprescription. Unfortunately, there is no consensus about which
marketing instrument is the main driver for category demand in the pharmaceutical
market. Most studies appear to fi nd that detailing is most effective, followed by jour-
nal advertising and direct-to-consumer advertising (DTCA) (Berndt et al. 1995 ,
1997 ; Fischer and Albers 2010 ), but there are notable exceptions. Narayanan et al.
( 2004 ) fi nd that detailing is not signifi cant and that DTCA is the main category
expansion driver. This fi nding is confi rmed by Rosenthal et al. ( 2003 ), but is contra-
dicted by Calfee et al. ( 2002 ), who discover no effect of either DTCA or other phar-
maceutical marketing instruments on new prescriptions or renewals in a disease
category. Vakratsas and Kolsarici ( 2008 ) fi nd that journal advertising is the most
effective instrument, followed by DTCA and that detailing is not signifi cant. There
is a need for reconciling these different fi ndings, for example, by determining which
variables moderate the effectiveness of the different instruments on category demand.
The type of analysis that is used to determine the effects of promotional efforts
on category demand may moderate promotional effectiveness in several ways. As we
discuss in Sect. 20.4.4 , Fischer and Albers ( 2010 ) derive category sales effects from
brand-sales models. They argue that this method is more accurate than the more
common approach of inferring category sales effects from category sales models
and this might explain some of the differences observed across studies. Similarly,
analyses that account for endogeneity tend to produce different results than those
which do not (Kremer et al. 2008 ). In a number of studies category demand data are
pooled over categories. Leefl ang and Wieringa ( 2010 ) demonstrate that pooling
across categories should be avoided as such studies (e.g., Iizuka and Jin 2005 ) are
more likely to generate biased results than those studies which do not.
Another possible set of moderators may be found in market characteristics. For
example, Fischer and Albers ( 2010 ) fi nd considerable differences between short-
term and long-term elasticities. This might indicate that the age of the product cat-
egory moderates promotional effectiveness, which possibly affects some instruments
more than others. Further research here is needed as, despite the growth in the num-
ber of studies on category level demand, the body of knowledge is still not suffi -
ciently large to identify empirically the relevance of possible moderators.

20.2.2 Brand Level Demand Effects of Pharmaceutical

Promotion

The number of studies that examine pharmaceutical promotional effectiveness at the
brand level is considerably larger than at the category level. Compared to category
level outcomes most brand level studies report somewhat stronger promotional effects,
but the effects are still reported to be moderate (Azoulay 2002 ; Berndt et al. 2003b ;
De Laat et al. 2002 ; Kolsarici and Vakratsas 2010 ; Windmeijer et al. 2006 ), or small