Direct-to-Consumer Advertising: Title and subTitle BreakA Haphazard Approach to Health Promotion

In JAMA 6 years ago, Holmer (then president of the Pharmaceutical Research and Manufacturers of America) and I published paired opinion pieces about the public health benefit of direct-to-consumer advertising (DTCA). Holmer wrote, “[DTCA] is an excellent way to meet the growing demand for medical information, empowering consumers by educating them about health conditions and possible treatments.”¹ I countered that, “unlike the truly valuable contributions of the pharmaceutical industry, [DTCA] is not good for patients, physicians, or the public’s health.”² Neither of us had much evidence to substantiate our arguments. We agreed that policymakers needed studies to determine more definitively the impact on public health.

In the intervening years, researchers have published a substantial body of literature on DTCA. The bulk of this work has been observational analysis and survey data of the $3.2-billion-per-year³ uncontrolled experiment with DTCA in the United States brought on by the decision by the US Food and Drug Administration (FDA) in 1997 to relax the rules governing mass media advertising for prescription drugs.

Findings from the literature show that DTCA drives sales of newer, more expensive products for symptomatic relief of chronic conditions or conditions with huge market potential; indeed, just 20 prescription drugs account for about 60% of the total industry spending on DTCA.^{4 - 7} Perhaps not surprisingly, retail sales of the 50 drugs most heavily advertised to consumers increased an aggregate 32% from 1999 to 2000 compared with 13.6% for all other drugs combined.⁸ Analysts estimate that each 10% increase in DTCA spending within a drug class increases sales in that class by 1%.⁴ Physicians have identified erectile dysfunction, arthritis, and allergies as the 3 most common conditions for which DTCA prompted care.⁹

Relying on emotional appeals, most advertisements provide a minimal amount of health information, describe the benefits in vague, qualitative terms, and rarely offer evidence to support claims.^{5 ,10 - 12} Physicians’ opinions support these findings. More than 80% of physicians believe that DTCA does not provide balanced information.⁹ Despite these shortcomings, only a small proportion of advertisements prompt the FDA to write regulatory letters.⁴ In 2002, the US General Accounting Office (GAO) highlighted the limitations on FDA oversight of DTCA.⁴ Subsequent articles pointing to declining regulatory action in the face of a steady number of complaints to the FDA about DTCA have confirmed that such limitations have resulted in millions of patients being exposed to misleading advertisements.^{13 - 14}

Do advertisements stimulate discussions between patients and physicians? According to consumer surveys, between one quarter and one third of adults annually have talked to a physician about a health issue after seeing an advertisement.^{15 - 17} Between 40% and 70% of physicians say that a patient seeing an advertisement helps the discussion about the condition and its treatment.^{9 ,18 - 19} Nearly 80% of physicians, however, think that DTCA encourages patients to seek treatments they do not need.⁹ Less than 10% of physicians consider DTCA a positive trend in health care.¹⁹ Thus, physician opinion suggests that DCTA may not be prompting the most important health discussions.

Does DTCA stimulate diagnosis? In 2003, a retrospective, cross-sectional telephone survey (n = 3000; 53% response rate) cited by proponents of DTCA and funded partly by a consortium of pharmaceutical companies looked beyond opinions and attitudes about this marketing strategy by cataloging the types of conditions discussed with physicians.¹⁷ The researchers examined only visits in which DTCA prompted clinical discussion. About 25% of patients with DTCA visits were given a new diagnosis, and 43% of these diagnoses were for conditions with “high-priority,” as identified by the Agency for Healthcare Research and Quality. Surprisingly, the new diagnosis rate was higher for visits in which the patient deemed DTCA to have “low influence” compared with “high influence.” Despite the large size of the study, researchers did not collect data on the rate of new diagnoses at non-DTCA visits.

Direct-to-consumer advertising has been associated with health service utilization for some conditions. Women familiar with 1 of 2 osteoporosis drugs due to self-reported exposure to advertisements had 9 times the odds of having bone densitometry performed.²⁰ Zachry et al²¹ demonstrated an association between DTCA expenditures and health services for certain conditions, including seasonal allergies and hyperlipidemia, but not for others, such as hypertension. Indeed, physician visits for allergy symptoms were relatively stable from 1990 to 1998 but increased by 5 million in 1999 following heavy DTCA for prescription allergy medication.¹⁵ During a period of sizable growth in DTCA for cholesterol-lowering therapy, a study²² found that the proportion of the population in 22 states using statins increased from 5% to 8%. The percentage of treated patients in the lowest categories of cardiovascular risk did not increase disproportionately, suggesting an appropriate expansion in the use of these drugs. That this expansion was due solely to DTCA, however, is unlikely.

Patient requests for prescription drugs appear to influence prescribing decisions. Survey data suggest that approximately 40% of visits in which a DTCA discussion occurs result in a prescription for the advertised drug.^{9 ,17} In more than half the cases, a physician prescribes the drug partly to accommodate a patient’s request.⁹ In a prospective survey comparing physician prescribing in Canada (which prohibits DTCA) and the United States, Mintzes et al²³ found that patients in the United States were twice as likely to request advertised drugs and that those who requested DTCA drugs were nearly 17 times as likely to receive a new prescription. The US General Accounting Office has concluded that about 8.5 million consumers annually, prompted by DTCA, request and receive an advertised drug.⁴ There is no compelling evidence that this is inappropriate prescribing according to FDA administrators.²⁴ In addition, it is possible that DTCA improves adherence to prescribed regimens. Based on a study funded by Pfizer, patients who asked their physician about an advertised medication were more than twice as likely to continue taking the medication.¹⁶

The available literature suggests that DTCA increases sales of prescription drugs primarily for symptomatic treatment of chronic conditions. It does so by stimulating patients with advertising messages of limited informational value to discuss therapy with physicians, who then prescribe these medications at a relatively high rate. Patients who receive care prompted by DTCA may be more adherent to these treatments. However, from the physician, patient, and public health perspectives, issues of the safety and net benefit of DTCA remain controversial. Whether DTCA helps physicians and patients achieve their joint goals of improved health through shared decision making by facilitating optimal therapeutic decisions and bringing the most appropriate patients to care is entirely uncertain. Compelling evidence is difficult to obtain as it requires teasing out an independent effect for DTCA amid a range of other forces. As Findlay¹⁵ noted, “survey-based assessments cannot be used to assess . . . the public health impact of DTCA.” In the face of this massive and uncontrolled experiment, it is very helpful to finally have the controlled study by Kravitz et al²⁵ in this issue of JAMA evaluating how DTCA influences therapeutic decisions.

Kravitz et al take an important new step in DTCA research, presenting a cleverly designed randomized controlled trial that plumbs the impact of this marketing strategy. These investigators examined what happens when physicians in primary care clinical settings are faced with specific requests for drug therapy. Standardized patients (SPs) portrayed major depression or adjustment disorder with depressed mood and made 1 of 3 request types for medication (brand-specific, general, or none). The study generated several key findings. First, if SPs made brand-specific requests, physicians’ prescribing rates of antidepressant medication for major depression were lower than if SPs made general requests for medication (53% vs 76%). Second, any request for prescription drugs in the setting of major depression helped avert undertreatment. Among SPs portraying major depression, “minimally acceptable initial care” was received by 98% of those making a general request, 90% of those making a brand-specific request, and 56% of those making no request. Third, when the condition was less severe with less clear indication for drug treatment (adjustment disorder), physicians were more likely to prescribe an antidepressant in response to a brand-name request rather than a general request (55% vs 39%). Thus, in facilitating treatment for major depression, general requests for medication were as good as, if not better than, brand-specific requests and, in adjustment disorder, brand-specific requests stimulated excessive care.

The study has several limitations. Obviously, it is difficult to generalize the results to other health conditions. More critically, the extent to which patient behavior is appropriately or inappropriately influenced by DTCA (differential activation) cannot be determined from this study. Put differently, the study leaves unanswered whether DTCA motivates the right patients to seek the right care or, on balance, inordinately influences patients to seek unnecessary care.

The recent debacle about cyclooxygenase 2 (COX-2) inhibitors may be instructive. Rofecoxib was the most heavily advertised drug to consumers in 2000 and retail sales quadrupled from 1999 to 2000 (rofecoxib’s advertising campaign by Merck was bigger than campaigns for either Pepsi or Budweiser in 2000).⁸ In 2003, Pfizer spent $87.6 million promoting celecoxib directly to consumers.²⁶ Recent data highlight that marked increases in COX-2 inhibitor use occurred primarily among patients at low risk of adverse events from less expensive nonsteroidal anti-inflammatory drugs (NSAIDs).²⁷ This inappropriate increase in COX-2 inhibitor use among patients for whom NSAIDs could be used accounted for more than 63% of the growth between 1999 and 2002. That this growth was due solely to DTCA is, again, unlikely, but Dai et al²⁷ describe succinctly the important role that DTCA probably played in this trend—a trend that may have resulted in as many as 140 000 serious adverse cardiovascular events.²⁸

The critical policy question remains. Do the benefits of DTCA outweigh the danger that consumers will demand and use medicines inappropriately? On balance, the answer appears equivocal and awaits further research. For conditions for which there is ample evidence that a particular therapy works, has limited adverse effects, and is underused, any mechanism that raises awareness is probably beneficial. Depression may be such a condition.²⁹ However, contrary to the claim that DTCA “is an excellent way to . . . [empower] consumers,”¹ the data from Kravitz et al suggest that while DTCA may guard against initial undertreatment, it simultaneously undermines optimal prescribing and stimulates overtreatment for lesser conditions on the continuum of affective disorders.²⁵ Additionally, for conditions such as osteoarthritis, for which there is evidence that alternative, less costly therapies work equally well and the risk of therapy is not completely understood or is substantial, DTCA is probably harmful, suggesting that such marketing is one of the least valuable contributions of the pharmaceutical industry to the public’s health.

In response to the concerns about COX-2 inhibitors, leading pharmaceutical companies voluntarily withdrew DTCA for these medications. Experts, however, consider it unlikely in the current US sociopolitical environment that the federal government will restrict DTCA and limit patient access to product information by “arbitrarily” excluding pharmaceutical companies from advertising their products directly to consumers.³⁰ Moreover, the court system appears willing to uphold the right to truthfully advertise prescription drugs to the public as protected commercial speech.³⁰ If New Zealand passes a ban on DTCA in 2005 as anticipated, the United States will be the only industrialized country permitting such practice.³¹

Given that DTCA in the United States is likely here to stay, how can the practice be regulated to most effectively achieve public health goals? Direct-to-consumer advertising, by simultaneously advancing consumerism over shared decision making and providing pseudoeducational material to patients, can undermine choices patients and physicians make together and negatively impact the public’s health.³² Decisions to advertise a specific product to the public do not necessarily reflect superior safety, efficacy, or the interest of the public’s health but rather calculations of return on investment.^{33 - 34} Driven by this financial interest, the domination of health-related communication by companies with drugs to sell will, by its very nature, lead to problems unless balanced by educational efforts geared to public rather than private good.³⁵

As Bodenheimer⁷ asked, “What if millions of parents were prompted by a TV spot to ask their physicians about asthma action plans to help in the self-management of their children’s asthma, an intervention shown in a Cochrane review to improve outcomes?” The answer seems clear. Concise, coherent, evidence-based messages, delivered using the most sophisticated techniques of Madison Avenue, unbiased by the motivation to turn a profit, and funded by either a tax on DTCA or an alternative financing scheme will benefit the public’s health. These public service messages could supplement the haphazard approach to health promotion that relies on occasionally helpful and intermittently harmful advertisements that play to a patient’s vulnerability and appeal to the desire to assert control over a potential outcome—advertisements that, first and foremost, sell prescription drugs.

If such a strategy proves untenable, then at a minimum, the FDA should expect that pharmaceutical companies improve the educational nature of DTCA by further developing and adhering to appropriate standards to facilitate regulatory efforts.^{12 ,36 - 38} In addition, the FDA should be adequately funded to accomplish its regulatory function.³⁹ Last, because the safety of a new drug cannot be known with certainty until it has been on the market for several years, and since drug withdrawals occur more than two thirds of the time within 3 years of release,⁴⁰ the FDA should consider a moratorium on advertisement of drugs directly to consumers for 3 years after initial market release.