The effect of direct-to-consumer pharmaceutical advertising on health choices: an empirical investigation using panel data

In 1997, the FDA relaxed regulations on direct-to-consumer (DTC) advertising for pharmaceutical products. This has led to a substantial increase in such expenditures which in turn raises some important public policy questions. For example, has this increase in DTC advertising affected the doctor-patient relationship in a way that distorts optimal choice of care? The benign view of such advertising is that DTC ads are simply informative and thus lead consumers to obtain better care than would have occurred otherwise. An opposing view is that DTC advertising entices consumers to obtain treatments that are unnecessary or more expensive than “optimal” treatment (For general discussions, see Berndt et al, AER 1995; Iizuka, T. and G. Jin, working paper 2002). This paper will investigate the effects of DTC advertising on a relatively rich panel data of roughly 50,000 individuals obtained from an insurance firm. Included in this data are all prescription drugs purchases for covered individuals from 1997-2001, as well as all other health services provided for the individuals through the insurance company. Demographic variables are also available for the individuals. We will use monthly DTC expenditures in specific geographic areas for a set of pharmaceutical products (provided by a major pharmaceutical firm) to measure the impact of expenditures on 1) physician visits, and 2) prescriptions written for the product, controlling for other observable characteristics. We will also examine the impact on brand switching within a therapeutic class. We build on the work of others by exploiting the panel nature of the data, using recently-developed estimators for non-linear panel data models (Honore, B. and E. Kyriazidou, ECM, 2000).