Session: The effect of direct-to-consumer pharmaceutical advertising on search, utilization, and welfare
Room: Phillips 101
Time: Wed 10:15-11:45

Presenter: Jonathan Cantor (Cornell University. )

Discussant: Julie Donohue (University of Pittsburgh)

Abstract

The US is one of only two countries in which Direct-to-consumer (DTC) advertising is currently permitted. There are ongoing calls in Congress to ban the practice, because of concerns that it may lead to the inappropriate diagnosis and treatment of diseases, and to higher prices for medicines. In this paper, we study the influence of DTCA for antidepressants by studying how they affect the decision to seek physician care, and whether the impacts differ by TV v.s. print media. We add to the literature because prior work in this area has only been able to link market level (or national) advertising to individual level physician visits, whereas our exposure comes from individual measures of advertising exposure.
Our main data set is the Simmons National Consumer Survey (NCS), a nationally representative cross-sectional survey covering 2003-2007. We create our independent variable by linking reports in the NCS of TV viewing and magazine reading habits with data from other databases that tell us what ads appeared in those shows or magazines. NCS asks how often the respondent read a specific magazine and watched either a specific television program or channel. Respondents were asked about their viewing habits for a list of broadcast and cable television programs, whether they watched broadcast television on a typical weekday and typical weekend during a specific timeslot, and their viewing of specific cable networks during different time slots. They also reported how many issues of each magazine they read on of the past four issues; we extrapolate for the entire year from this, assuming that the reading pattern of the last 4 issues is indicative of behavior during that past year.
The data on television DTC advertisements for antidepressants came from TNS Media Intelligence, which gives the exact time, date and program each antidepressant advertisement aired during, for national networks, cable and spot markets. Data on DTC print advertisements were taken from a database created at Cornell University containing a digital representation of every pharmaceutical advertisement that appeared between our time period in 26 of the top read consumer magazines in the United States, which in turn represents approximately 60% of readership data in the NCS.
Among the 105,299 respondents to the NCS on whom we have data, the average exposure to antidepressant print ads is 5.43 and 174.5 for TV ads. We examine the causal impact of antidepressant advertising exposure to number of health care provider visits, ( general/family practitioners, nurse practitioners, and internist) in the last 12 months in the NCS. We will estimate a falsification test using a dependent variable that should not be affected by antidepressant ads, visits to podiatrists. The richness of our data allows us to specify a model in which the variation in ad exposure used to estimate the key parameters are orthogonal to unobservables about the individual that may be correlated with both ad exposure and with care seeking behavior, controlling for factors used by marketers in targeting ads.

Key Terms

Advertising, prescription drugs, antidepressants

Authors:

Jonathan Cantor (Cornell University) , Kosali Simon (Cornell University) and Rosemary Avery (Cornell University)