Spillovers and Other Externalities in Pharmaceutical Marketing - Innovation and Marketing in the Pharmaceutical Industry - page 677

DTCA changed in the USA. For both the US and Canadian markets, the data contain
monthly information for the period of January 1997 through June 2002. In each
country, the data set tracks the total prescriptions written, the average price of the
prescription, detailing expenditures, DTCA expenditures (in TV and magazines) in
the US market, and other marketing efforts (denoted as “ome”) such as professional
journal advertising, meetings, and events. IMS Health Canada provided the
Canadian data, while the US data on sales, prices, detailing, and ome are from
Verispan, and data on DTCA expenditures are from TNS Media Intelligence-CMR
( www.tnsmi-cmr.com ) .

23.3.2 The Model

Here, monthly category sales denote the total number of prescriptions for nasal
steroid medications in that time period. Category sales depend on the prices, detailing,
DTCA, and other marketing activities of all brands in the category; further, sales are
also influenced by seasonality. As marketing variables such as detailing, DTCA,
and ome can have carryover effects, we allow for such effects in the specification.
The category sales relationships are captured via country-specific linear regression
models. The DTCA expenditures via TV and magazines are combined together (and
denoted “dtc”); this allows us to compare our analysis of the US data with extant
research on DTCA effects that has investigated only the combined effects of all
DTCA. In period t and country j , category sales, S jt , are defined by:

Ln () S =+ + aa a SD SD + a SD + a P +

jt 0 j 1 j 1 2 j 2 3 j 3 4 j jt

a D + a OME + a DTC US, + e

(23.1)

5 j jt 6 j jt 7 j t jt

where SD k ( k = 1, 2, 3) refers to the seasonal dummy variable, P jt the share-weighted
category price in the country under consideration, 6 D jt the goodwill level associated
with detailing (det), OME jt the goodwill associated with other marketing efforts
(ome), and DTC US, t the goodwill arising from the combined (i.e., TV and magazines
dtc) expenditures. The goodwill terms will reflect carryover effects of marketing
activities (see below). Notice that DTCA expenditures are incurred only in the USA.
However, they affect sales in both countries, and our specification allows the impact
of these expenditures to vary across countries, as indicated by the subscript j . The
α 's are the estimated country-specific parameters, and e jt is the mean-zero random
error term.

6 A log-log formulation for price is another possible specification that can be employed here; we
use the current approach to allow for easier comparison with some of the other empirical explora-
tions of DTCA.