Panel Paper: Spillover Effects from Direct-to-Consumer Advertising (DTCA) of Prescription Drugs

Friday, November 8, 2019
I.M Pei Tower: Majestic Level, Vail (Sheraton Denver Downtown)

*Names in bold indicate Presenter

Matthew Eisenberg1, Brendan Rabideau1, Neeraj Sood2, Abby Alpert3, Jeff Niederdeppe4 and Rosemary Avery4, (1)Johns Hopkins University, (2)University of Southern California, (3)University of Pennsylvania, (4)Cornell University

Background: Detractors of direct-to-consumer-advertising (DTCA) argue that DTCA encourages consumers to try prescription drugs first rather than making lifestyle changes (such as increased exercise, healthier diets, and smoking cessation), and promotes inappropriate drug prescribing behavior (such as patients pressuring physicians to prescribe drugs even when the drug is not medically indicated). Those in favor of DTCA argue that it provides important information to consumers, increases disease awareness and diagnosis, and increases initiation of both treatments and healthy lifestyles.

Objective: To examine the effect of prescription drug DTCA on outpatient utilization for five common chronic conditions (hypertension, hyperlipidemia, diabetes, depression, and osteoporosis). We further decompose this effect into office visits that result in a drug claim, office visits that do not result in a drug claim, and office visits that result in a more sustained engagement with the health care system.

Data: We combine data on pharmaceutical advertising from Nielsen (Nielsen Ad*Views) and pharmaceutical and insurance claims data from 40 large national employers, covering 18 million person-years (2004-2010). We collapse data on outpatient utilization to the three-digit ZIP code level and merge this to the Nielsen Ad data to create a three-digit by quarter dataset (N=7,705) for those aged 40-60.

Methods: Building on prior research, we examine how DTCA impacts office visit and outpatient utilization by exploiting a large and plausibly exogenous shock to DTCA driven by the introduction of Medicare Part D. Specifically, we use an instrumental variable strategy which exploits variation across geographic areas in the population share that is covered by Medicare (aged 65+) to predict changes in advertising exposure across areas. Our analysis focuses on those aged 40-60. Since advertising cannot be perfectly targeted to the elderly, we exploit the sudden differential increase in advertising exposure for non-elderly (under age 65) that live in elderly-dominated areas to estimate the effect of DTCA.

Results: Our reduced form analysis finds significant effects. When comparing those living in elderly-dominated areas to those who do not, before and after the introduction of Medicare Part D, we find that 2.8 more people per 1,000 had an office visit for one of the chronic conditions. Our two-stage least squares analysis found larger effects, with 9.7 more people per 1,000 having an office visit. These results were primarily driven by individuals with diabetes, hypertension, hyperlipidemia, and osteoporosis. Our preliminary decomposition and trajectory analysis suggest substantial spillovers; we find that a large share of the increased office visits from advertising do not result in a short-term drug claim.

Conclusions: In the context of a rapidly changing policy environment (e.g. AMA proposed ban, the President’s proposed regulation to mandate price transparency in television advertising), understanding the total effect of DTCA is crucial. While prior studies answer the business question of whether DTCA increases product sales, this study sheds light on the policy-relevant question of how DTCA effects a broader range of health care inputs. Indirect or spillover effects must be taken into consideration when analyzing the full costs and benefits of proposed regulations.