The Effects of Direct-to-Consumer Advertising in the Prescription Drug Market

The year of 1997 witnessed an important regulatory change in direct-to-consumer (DTC) advertising of ethical drugs. For the first time, the Food and Drug Administration (FDA) permitted brand-specific DTC ads on TV without a ”brief summary” of comprehensive risk information. This led to a three-fold growth of DTC advertising expenditure in four years, followed by an intensive debate about the effects of DTC advertising on patient and doctor behaviors. This paper empirically examines the effects of DTC ads on ethical drugs by combining 1996-1999 DTC advertising data with the annual National Ambulatory Medical Care Survey (NAMCS). We find that DTC advertising leads to a large increase in the number of outpatient drug visits, a moderate increase in the time spent with doctors, but no effect on doctors’ specific choice among prescription drugs within a therapeutic class. Consistent with the proponents’ claim, this finding suggests that DTC ads encourage patient visits but do not challenge doctors’ authority in the specific choice of prescription drugs. We cannot rule out the possibilities, however, that DTC ads may induce doctors to use prescription drugs over alternative treatments and doctors may spend extra time clarifying DTC ads if they do not prescribe the most advertised drug(s). Our results also find DTC advertising to be a public good for all prescription drugs in the same therapeutic class. This implies that DTC advertising has a market-expanding effect but not a business-stealing effect. ∗†Owen Graduate School of Management, Vanderbilt University. Email: [email protected] ‡Department of Economics, University of Maryland. Email: [email protected]. We wish to thank Thomas Hubbard, Judy Hellerstein, Bill Evans, Seth Sanders and the participants in workshop at the University of Maryland for constructive comments and suggestions. We are very grateful for Competitive Media Reporting and IMS HEALTH for generously providing the advertising data for this study. All errors remain ours.