Pharmaceutical Detailing Elasticities: A Meta-Analysis - Innovation and Marketing in the Pharmaceutical Industry - page 534

dimensions offered by Stremersch and Lemmens ( 2009 ), European pharma mar-
kets, especially France, Italy, Spain, and UK are far more regulated than the USA.
Greater regulation implies that persuading physicians to prescribe a new drug is a
much more complex or diffi cult sale in Europe than in the USA. Buyers will need
more answers and convincing to try something new when their actions are restricted
by regulation. In such circumstances requiring more information provision, personal
selling tends to be the most effective marketing instrument. An indicator of this is
the relative length of detailing calls in the USA vs. Europe. In the USA, detailing
visits last on average between 3 and 4 min; in Europe (UK) around 9 min (Heutschi
et al. 2003 ), suggesting that European physicians fi nd detailing visits to be relatively
more valuable than US physicians.
The next three fi ndings pertain to dataset characteristics .
Year of data collection . As stated in the introduction, there appears to be a trend of
growing resistance of physicians to detailing as pharma product lines mature and
there is more oversight of prescription decisions by managed care organizations.
Therefore, we expect that within the span of our database, detailing elasticities will
decrease over the years of data collection. Towards this end, we operationalized
year of data collection as mean year in the panel from which elasticities were esti-
mated. Again, we found no main effect for this argument ( b = 0.002, ns).
However, we found a signifi cant interaction effect between the year of data collec-
tion and stage in the PLC. Specifi cally, we found that detailing elasticities in the early
stage of the PLC signifi cantly decline over the years of data collection ( b = −0.023,
p < 0.01). One possible reason is that since it became permissible in 1997, DTC adver-
tising has been frequently used in addition to detailing in early phases of product life
cycles in the USA (e.g., Donohue et al. 2007 ; Liang and Mackey 2011 ) and recent
econometric models that include DTC advertising in addition to detailing have found
such advertising has a signifi cant positive effect on total class sales (e.g., Donohue
and Berndt 2004 ), as well as a main effect on brand share (Narayanan et al. 2004 ).
We reason that in the presence of such signifi cant effects of DTC advertising, more
recent econometric models involving newer products that include DTC advertising
will yield lower detailing elasticities than older models from the pre-DTC era.
Relative vs. absolute output measure . Elasticities based on absolute sales measures
capture changes in sales due to both primary (market expansion) and secondary
(share expansion) changes resulting from varying selling effort (see, e.g., Hagerty
et al. 1988 ). In contrast, share-based elasticities classify only a portion of the market
expansion as primary demand (Steenburgh 2007 ). Accordingly, we expected that
detailing elasticities from models using relative (share) output measures will be
smaller than those from models using absolute output measures.
In our analysis, we dummy coded the output measure variable as 1 if measured
in relative terms and 0 if measured in absolute terms. From the main effect in
Table 18.3 , we found no evidence that detailing elasticities are higher in settings that
use relative output measures than those that use absolute measures ( b = 0.073, ns).
Temporal aggregation of data . Previous meta-analyses (e.g., Tellis 1988 ; Assmus
et al. 1984 ; Bijmolt et al. 2005 ) have suggested that estimates of marketing