Consumers’ Understanding of FDA Approval Requirements and Composite Scores in Direct-to-Consumer Prescription Drug Print Ads

Abstract

In two studies, we investigated how laypersons perceive the FDA approval process, FDA authority, and the presentation of composite scores in direct-to-consumer (DTC) prescription drug print ads. The first study consisted of four focus groups (N=38) in two cities. Using a semi-structured guide, a moderator led participants through the viewing of three existing DTC print ads that differed in the presence or absence of composite score information, and participants discussed their views of the ads and their understanding of composite scores. The second study surveyed a nationally representative sample of 1,629 individuals from the general population who saw a fictitious DTC print ad and answered closed-ended questions about the same topics. Results showed that knowledge of FDA approval and authority was mixed, with several misconceptions apparent. Many consumers were not familiar with the use of composite scores in a medical context or in advertising and, in the first study, expressed distrust of the product and the ad after learning about how composite scores are used. In the second study, receiving composite score information changed the perceived clarity of the ad, but not the perceived risk or benefits. Implications for the presentation of complex medical information are discussed.

In 1997, the Food and Drug Administration (FDA) issued draft guidance that clarified regulations for television advertising of prescription drugs, leading to a surge of advertising to consumers, both in television and print formats (U.S. Food and Drug Administration, 1999). According to Nielson data, pharmaceutical companies spent $4.3 billion on direct-to-consumer (DTC) promotion in 2013 (Mack, 2014). Thus, medical concepts previously limited to health care professionals are now presented to laypersons regularly. Consequently, laypersons navigate these medical concepts without the medical education of health care professionals.

For many years, drug benefit information was presented only as vague statements in DTC advertising, most often derived from FDA-approved product labeling for the drug (Kaphingst, DeJong, Rudd, & Daltroy, 2005). Clinical trial information, which forms the basis of marketing claims in prescription drug advertising, has been difficult for both laypersons and health care professionals to understand (Biedrzycki, 2011). Recent research, however, has shown that many people value the presentation of clinical trial information in ads (Schwartz, Woloshin, & Welch, 2009), and several studies have shown that adding quantitative information from clinical trials to DTC advertising may be helpful for laypersons (O’Donoghue et al., 2014; Schwartz & Woloshin, 2011; Woloshin & Schwartz, 2011). Quantifying particular aspects of benefits, such as how many people in the drug and placebo conditions of a clinical trial had symptom relief, provides additional information for laypersons who may then have more informed discussions with their health care professionals.

For a product to be approved and later advertised, a pharmaceutical company must conduct clinical trials to ensure that the product is safe and effective. In some cases, the outcomes of the trial are singular, such as a blood sugar reading. In these cases, it is relatively easy to understand what makes a treatment effective or not: if the product brings levels closer to acceptable readings by some margin, the product is effective. In other cases, however, the efficacy of the drug is measured by multiple items, and the items are combined to form a composite endpoint or score. In fact, of the ten most advertised-to-consumer prescription drugs in 2014 (Millman, 2015), eight utilized outcomes involving composite scores. When composite outcomes are used, not all items need be effective to be scored as efficacious overall. For instance, the Hamilton Depression Inventory (Ham-D; Kobak & Reynolds, 1999) is a commonly used measure of 21 items. Patients do not have to score high on every item to be considered depressed if the overall combined score is high (e.g., the patient’s score on the “loss of weight” item could be zero but the overall score could be high enough to indicate depression). In terms of prescription drugs, this means that not all items need to reach statistical significance compared to placebo when combined into a composite score. Thus, a drug evaluated for its efficacy using the Ham-D could show a decrease in overall depression without necessarily affecting weight loss. As laypersons view more quantitative information, it is important to determine how well they understand composite scores.

Another central issue in DTC advertising is the FDA approval process. Bell, Kravitz, & Wilkies (1999) and Schwartz and Woloshin (2011) asked nationally representative samples questions about their knowledge of the FDA approval process and found that a substantial number of people have misconceptions that could influence their consideration of prescription drugs, such as (1) FDA only approves drugs that are extremely effective; (2) only extremely effective prescription drugs are allowed to be advertised to consumers; (3) FDA only approves drugs that do not have serious side effects; and 4) drugs that have serious side effects cannot be advertised to consumers. In reality, FDA approves drugs based on an analysis of the benefits and the risks of the drug in the context of the disease. Over 50% of participants had at least one of these misconceptions, and people with lower versus higher levels of education were more likely to have a misconception. The approval process influences the marketing claims that pharmaceutical companies can make, and knowledge of this process may inform a layperson’s understanding of clinical trial information included in DTC advertising.

To investigate these issues, we conducted two studies using an exploratory sequential mixed method design. The first study involved a qualitative examination (through focus groups) of consumer understanding and familiarity with the concept of composite scores. Focus groups are qualitative in nature, allowing for in-depth, but not generalizable, exploration of particular topics. Focus groups have the advantages and disadvantages of capturing social interactions that arise from conversations (Dodds, Tseelon, & Weitkamp, 2008). We explicitly chose focus groups so that we could examine discussions among participants. The three main goals of this qualitative research were to determine:

1. participants’ initial general reactions to prescription drug print ads featuring products whose efficacy is based on composite scores;

2. participants’ initial specific reactions to efficacy claims based on composite scores; and

3. how understanding of composite scores affected participants’ reactions to efficacy claims.

The second study examined many of the same questions in a nationally representative survey to extend the findings of the focus group study. The purpose of the study was to explore:

1. to what extent consumers understand FDA authority and the approval process for prescription drugs;

2. how well consumers comprehend the concept of composite scores; and

3. whether exposure to DTC ads with composite scores, paired with the educational intervention about composite scores developed in Study 1, influences perceptions of a drug’s efficacy and risk.

We predicted that older adults and people with less education would have more difficulty understanding composite scores based on findings from the literature (Cole & Balasubramanian, 1993; Shiffman, Gerlach, Sembower, & Rohay, 2011). We also expected that composite score comprehension and recall of drug benefits would be positively associated with the perceived clarity of the ad.

Method

Study 1

Participants and Setting.

A total of 38 adult participants from St. Louis, Missouri, and Greenbelt, Maryland, joined one of four in-person focus groups in September of 2011. Exclusion criteria included self or family employment in the market research field and participation in focus groups within the past 3 months. Twenty-one participants (55%) were female; 23 participants (61%) were Caucasian; 9 participants (24%) were African-American; 2 participants (5%) were Asian-American; 1 participant (3%) was Native American; and 3 (8%) reported Hispanic identification. Groups were divided by higher and lower education, defined as the completion of some college or not.

Procedure.

This study was granted approval from the relevant Institutional Review Boards. A trained moderator led each focus group using a semi-structured guide. Participants saw three actual DTC print ads, selected for their inclusion or exclusion of composite scores. The first ad, for Lipitor, indicated to treat high cholesterol, was used as a control because Lipitor had demonstrated efficacy based on a single numerical endpoint. This particular ad included percentage improvement in cholesterol levels. The other two products’ efficacy was based on composite scores. The ad for Actonel, indicated to prevent osteoporosis, featured a middle-aged woman and highlighted a number of bones that might be vulnerable to fracture (e.g., wrist, hip). The claim in this case was based on a composite score of efficacy data for all non-vertebral locations studied. The other ad, for Nasonex, indicated to treat nasal allergy symptoms, featured an animated bee and mentioned relief from nasal congestion. In this ad, effectiveness was measured by a composite of efficacy regarding multiple symptoms.

After participants discussed their reactions to the ads in general and the efficacy claims specifically, the moderator defined composite scores, provided an example, and asked how this changed opinions, if at all. The example was that of a decathlon, where competitors participate in 10 events. Competitors are not required to win all events to come in first; rather, the person with the highest overall score places first. Thus, if a competitor wins some events by a large margin but places second or worse in other events, that person may still be the winner.

Discussions lasted approximately 90 minutes, and participants received a $75 honorarium. Sessions were audiotaped and transcribed. A qualitative transcript-based analysis of the data was completed. The analysis was organized around the sections of the moderator guide and focused on examining participants’ initial reactions to prescription drug advertisements, participants’ initial reactions to effectiveness claims based on composite scores; and how an understanding of composite scores impacted participants’ reactions to these effectiveness claims. For each section, we compared the similarities and differences in responses between the higher education and lower education groups.

Study 2

Participants and Setting.

In 2014, GfK Custom Research invited 2,957 general population panelists to participate in an online study. GfK’s KnowledgePanel is a probability-based consumer panel that is nationally representative of adults in the U.S. (Dennis, 2010). Nevertheless, participants without broadband access to the Internet and those who otherwise could not view the print ad were excluded. A total of 1,891 panelists responded to the invitation and 1,629 participants completed the survey. Compared with nonresponders, individuals who completed the survey were more likely to be male, older, White, and non-Hispanic and have more income and a bachelor’s degree or higher (all ps < .001, except sex, where p < .01). However, the data were weighted back to the U.S. population on gender, age, income, and race/ethnicity.

Participants ranged in age from 18 to 92 years old (M = 49.90, SD = 16.74) and the median age was 51 years old. They were evenly split on gender (50.3% male). The majority reported being White (82.3%) and having some college or a college degree (66.2%). Thirty-eight percent (37.7%) reported a household income of less than $50,000 a year; 30% reported $100,000 a year or more; 16.9% reported $50,000 to $74,999; and 14.8% reported $75,000 to $99,999.

Procedure.

This study was granted approval from the relevant Institutional Review Boards. Initial survey items measured participants’ understanding of FDA’s “safe and effective paradigm,” drug approval requirements and authority, and participant attitudes about drug side effects and side effects information in general. Next, all participants saw a print ad for a fictitious osteoporosis drug, Foramel, featuring composite score information (Figure 1). The fictitious ad mimicked existing ads with composite scores while ensuring participants had no previous product knowledge or experience. Participants completed questions including perceived drug efficacy and risks, perceived clarity of the ad’s benefit information, trust in the ad’s benefit information, recall of the drug’s benefits, and open-ended composite score comprehension. Participants then read a brief educational intervention with information on how composite scores are derived (Figure 2) and viewed the same ad a second time. Following the second ad exposure, the remaining survey questions evaluated participants’ comprehension of the composite score information within the ad and reassessed perceived drug efficacy and risks, perceived ad clarity, and trust in information. We debriefed participants by explaining that Foramel is fictitious.

Figure 1.

Ad for fictitious osteoporosis drug Foramel, shown to participants in Study 2.

Figure 2.

Educational intervention provided to participants after first viewing the Foramel ad in Study 2.

Measures.

FDA knowledge.

To measure understanding of FDA’s “safe and effective” paradigm, we asked participants to choose the correct statement among four statements about when FDA approves a prescription drug. We asked four true/false questions from Schwartz and Woloshin (2011) to measure understanding of FDA approval requirements. Then participants were asked to indicate which of four products (prescription drugs, over-the-counter drugs, herbal remedies, vitamins, or none of the above) required FDA approval before marketing; what percentage of new prescription drugs must be approved by FDA before they are sold to consumers (all [100%], most [50%−99%], some [1-49%], none [0%], or don’t know); and whether FDA must approve prescription drug ads before they appear in magazines or on television (true, false, or don’t know). See Table 1 for question wording and correct responses.

Table 1.

Weighted percentage for Study 2 FDA knowledge questions.

Question Wording	Correct Response	% of correct responses	% of incorrect responses
When the FDA approves a prescription drug it means that…	The drug offers benefits that are greater than the risks	68.7	31.3
The FDA only approves prescription drugs that have been found to be extremely effective	False	55.5	44.5
The FDA approves prescription drugs only if they do not have serious side effects	False	77.6	22.4
Only prescription drugs that have been found to be extremely effective can be advertised to consumers	False	74.2	25.8
Prescription drugs that have serious side effects cannot be advertised to consumers	False	83.1	16.9
Which of the following products does FDA approve before they are sold to consumers?a	Prescription drugs	92.4	7.6
What percentage of new prescription drugs must be approved by the FDA before they are sold to consumers?	All of them (100%)	65.2	33.8
FDA must approve prescription drug advertisements before they appear in magazines or on TV.	False	31.2	68.8

Note.

^aParticipants could select more than one answer to this question.

Composite score comprehension.

Before the educational intervention, we measured composite score comprehension with one open-ended question: “What does the term “composite score” mean as related to prescription drugs?” Two independent coders examined responses to this item.

After the educational intervention and the second ad exposure, participants responded to three true/false items related to composite score comprehension, for a closed-ended composite comprehension measure: “Foramel prevents all bone fractures,” “Foramel might not prevent all bone fractures but it is more effective than no treatment,” and “Foramel might not prevent all bone fractures equally.” Correct answers (false for the first item, true for the second and third items) were then summed to create a range of 0 (no correct responses) to 3 (all correct responses).

The following four items were asked pre- and post-educational intervention.

Perceived efficacy was measured with two items. One assessed likelihood of benefit (“In your opinion, if 100 people take Foramel, for how many will the drug work?”) on a 6-point scale (0 people, 20 people, 40 people, 60 people, 80 people, or 100 people). The second item assessed the magnitude of benefit (“In your opinion, if Foramel did help reduce bone fractures, how effective would Foramel be?”) on a 6-point scale (1 = not at all effective, 6 = extremely effective).

Perceived risk was measured with two items. One assessed likelihood of risk (“In your opinion, if 100 people take Foramel, what percentage of them will have side effects?”). Based on pretesting, this item was measured using a sliding scale from 0-100%. The second item assessed the magnitude of risk (“In your opinion, if Foramel did cause side effects, how serious would they be?”) on a 6-point scale (1 = not at all serious, 6 = very serious).

Clarity of the ad was measured with one item (“The ad clearly presented the benefits of Foramel”) on a 5-point scale, which we reverse-coded so that higher agreement received a higher score (1 = strongly disagree, 5 = strongly agree).

Trust in information was measured with one item (“How likely is it that the benefits of the drug presented in this ad are true?”) on a 7-point scale (1 = not at all likely, 7 = extremely likely).

Benefit recall was asked pre-educational intervention and measured by asking participants the open-ended question “According to the ad, what are the benefits of Foramel?” Responses were coded as either “correct” (responses related to osteoporosis and/or strengthening bones/preventing fractures) or “incorrect.”

Analyses

The data were weighted back to the U.S. population on gender, age, income, and race/ethnicity. We report descriptive statistics for the FDA knowledge items. We used multiple linear regression to test whether the post-educational intervention measure of composite score comprehension was associated with post-educational intervention measures of perceived efficacy, perceived risk, clarity of the ad, and trust in information. These models controlled for age, gender, income, education, and current use of prescription medications (yes/no). We used multinomial logistic regression to test predictions about education and age, with education and age included in the same model. We used multiple linear regression to test predictions about the association between benefit recall and post-education perceived clarity, controlling for age, gender, income, education, and current medication use. We also tested our regression models without covariates; because the patterns of results did not change we do not report the bivariate model results. We used paired samples t-tests to assess whether measures changed after the educational intervention.

Results

Study 1

Initially, most participants thought all three drugs they saw would be effective. Several participants mentioned that the presence of a statement about FDA approval was comforting. The Actonel ad did not contain this information, and some participants thought this meant that the drug was less effective than the others. The presence of statistical information in the Lipitor ad also provided reassurance, although some participants misinterpreted what the numbers meant. Specifically, the ad stated that Lipitor reduced cholesterol levels by 39-60%, and some participants thought this meant that Lipitor worked in 39-60% of the people who took it.

When asked to think about the efficacy of the two products that were based on composite scores, participants had different impressions. Most participants took the Nasonex ad, with its claims of treating congestion, at face value. They thought that the term congestion was intuitive and believed that the product would effectively treat any symptoms related to allergies. A few participants complained that the term congestion was not specific enough.

The Actonel ad called out six particular bones and most participants believed this meant that the drug was equally effective for all six bones. A statement in the ad read, “In Actonel studies, fractures beyond the spine were measured as a group, not separately.” A few participants in the higher and lower education groups understood this to be an indication of a composite score, but many participants thought this meant that participants were tested in a group setting, rather than being interviewed individually.

One to two participants in each group said they had heard of the term “composite score,” generally in the form of the ACT college entry test, grades in school, or sports. Few participants recognized the term in relation to prescription drugs. After the moderator described what a composite score was and how it applied to prescription drugs, he provided a concrete example, and participants viewed the ads again. Some participants felt the ads were less convincing, and a few mentioned they thought the ads were deceptive. The groups were mixed on whether there should be additional information about each individual symptom or whether there was already sufficient information in the ad. Overall, there was no evidence of an association between participant education and reactions to the ads or responses to the questions.

Study 2

FDA knowledge.

Responses to questions about FDA approvals and authorities are shown in Table 1. Out of four true/false questions that assessed understanding of FDA approval requirements, nearly half of the participants answered all four questions correctly. Many participants incorrectly thought that FDA only approves prescription drugs that have been found to be extremely effective, indicating that a substantial number of participants lacked knowledge about how effective drugs must be to receive FDA approval. Nearly all participants knew that FDA must approve prescription drugs before they are sold, and most knew that FDA approval applied to all new prescription drugs and meant that the drug offers benefits that are greater than the risks. However, few participants understood FDA authority related to ads.

Pre- and post-educational intervention measures.

In response to the open-ended pre-educational intervention question, “What does the term “composite score” mean as related to prescription drugs?” only 13.6% of participants responded with an answer indicating a combination or average of scores, and another 5.7% responded that it involved the effectiveness of the drug. Almost half (45.3%) of participants responded that they did not know, and another 22% of participants responded in some unrelated way that was categorized “other.” Given that over two-thirds of participants did not provide a relevant answer, we did not further analyze this item.

In terms of the closed-ended post-educational intervention composite score comprehension question, overall, 77.2% of participants answered all three questions correctly, resulting in a mean of 2.68 (SE = 0.02). After the educational intervention, greater composite score comprehension was associated with greater trust in information (β = .30, p < .001), greater perceived efficacy (likelihood of benefit β = .23, p < .001; magnitude of benefit β = .23, p < .001), higher perceived ad clarity (β = .15, p < .001), and lower perceived risk (likelihood of risk β = −2.73, p = .031).

We predicted that people with lower levels of education would have more difficulty comprehending composite scores. This hypothesis received strong empirical support. Multinomial logistic regression results indicated that for each additional year of education: the odds of having the lowest comprehension score (0) versus the highest score (3) decreased by 16% (OR = 0.84, 95% Confidence Interval [CI] = 0.77, 0.91, p < .001); the odds of having a comprehension score of 1 versus 3 decreased by 19% (OR = 0.81, CI = 0.73, 0.89, p < .001); and the odds of having a comprehension score of 2 versus 3 decreased by 8% (OR = 0.91, CI = 0.85, 0.96, p = .001). Although we predicted that older adults would have more difficulty understanding composite scores, this was not supported. Age was not related to comprehension of composite scores in the multinomial logistic regression model (OR = 1.00, 95% Confidence Interval [CI] = 0.99, 1.01, p > .05).

Most participants correctly recalled the drug’s benefit (80.9%). We expected that recall of benefit information would be positively associated with how clearly participants thought the ad presented the benefits of the drug. We found that participants who recalled the drug’s benefits agreed more strongly that the ad clearly presented the drug benefits after the educational intervention than participants who did not recall the drug’s benefits (β = .23, p < .001).

In examining whether participants’ responses differed before and after the educational intervention, we found that only perceived clarity differed, t(1607) =−9.44, p < .001 (Table 2). Participants reported lower perceived ad clarity after the educational intervention. Perceived efficacy, perceived risk, and trust in information did not differ before and after the educational intervention, ps > .05.

Table 2.

Weighted means and standard errors for Study 2 pre- and post-education intervention outcomes.

Outcome	Possible range	Pre-educational intervention	Post-educational intervention
Perceived efficacy likelihood	1-6	4.11 (0.03)	4.09 (0.03)
Perceived efficacy magnitude	1-6	4.47 (0.03)	4.46 (0.03)
Perceived risk likelihood	0-100	40.76 (0.70)	40.62 (0.70)
Perceived risk magnitude	1-6	3.82 (0.03)	3.85 (0.04)
Perceived clarity	1-5	4.01 (0.02)	3.84 (0.02)*
Trust in information	1-7	4.81 (0.04)	(4.80 (0.04)

^*Note. Significantly different from pre-educational intervention, p < .001.

Discussion

These two studies demonstrate that despite the prevalence of prescription drugs approved on the basis of composite scores, consumers have little knowledge of this concept or how it may relate to the information they receive about these drugs. The first study provided prompted- and unprompted-reactions indicating lower levels of trust in ads after learning how composite scores relate to the claims in the ads. The second study, using a nationally representative sample, found that after learning about composite scores, most people understood the composite score and perceived the ad as less clear. Although the first study did not suggest strong differences by education level, the second study demonstrated that education level increased composite score comprehension. This research represents some of the first exploration of the transmission of composite score information in DTC advertising and supports a recent experimental study showing that additional information about composite scores may be required to clarify the concept for consumers (Williams et al, 2015). This research suggests that pharmaceutical companies, regulators, and public health advocates should consider the communication of composite scores when discussing how to evaluate effective DTC advertising.

As shown in Study 1, participants generally viewed the Actonel ad as more deceptive than the Nasonex ad. There are a number of reasons this may be the case: perhaps because (1) Actonel listed specific symptoms and therefore was thought to be more deceptive; (2) Actonel treated a more serious condition and had a more serious risk profile; or (3) patients thought they would be able to tell for themselves whether Nasonex worked, whereas they would not be able to do so with Actonel. This last thought appeared to have more traction in the groups, with several participants voicing it. These potential explanations highlight the many factors that influence perceptions about prescription drugs, even when discussing specific aspects of drug efficacy such as composite scores.

These groups provided support for our expectation that laypersons have little knowledge about composite scores and provided impetus for conducting a more representative survey of individuals. Specifically, many participants were reassured by an FDA-approved statement, causing us to wonder how much participants know about the prescription drug approval process and the authority of the FDA. Second, the groups demonstrated that some people do not know what a composite score is and that disclaimers about composite scores may be ineffective without this knowledge. Finally, the groups showed that a concrete example such as a decathlon may be successful in helping consumers grasp the concept of a composite score.

As shown in Study 2, knowledge of FDA workings was mixed. Results show that a number of participants were unclear about FDA’s authority and approval requirements; in fact, our findings were nearly identical to the studies of Schwartz and Woloshin (2011) and Bell et al. (1999), which also used nationally representative samples. Together, these findings suggest that individuals’ perceptions about FDA and its authorities regarding prescription drugs have been stable over time.

Most participants did not understand FDA’s authority over prescription drug ads. In actuality, FDA provides comments prior to initial ad release if requested by the pharmaceutical company and may issue an untitled letter or Warning letter against violative ads (U.S. FDA Prescription Drug Advertising Regulations, 2014). FDA does not approve drug ads. In addition, pharmaceutical companies are free to disseminate ads without FDA comment. All prescription drug ads must be submitted to FDA at the time of initial dissemination. Few participants correctly understood that FDA does not pre-approve prescription drug ads. This misunderstanding could have implications for how people process information in the ads. When individuals process information more peripherally (as they might when watching television or viewing a magazine ad) as opposed to centrally, perceptions of ad credibility can have a greater influence on attitudes and behaviors (Petty & Cacioppo, 1986). If participants are not paying close attention to the ad (processing peripherally) and believe FDA has approved such ads, they may not thoughtfully process the information and may, thereby, have greater trust in the information presented. This makes ads that violate the regulations by overstating the efficacy of the drug or minimizing the risk of the drug potentially more problematic.

As in Study 1, most participants could not define composite scores initially. After the educational intervention, most participants understood the composite score used in the ad. Participants who had less education exhibited poorer post-educational intervention composite score comprehension than participants with more education, supporting our hypothesis. This is consistent with prior research that has found education to be related to literacy and numeracy (Institute of Medicine [IOM], 2014; Williams et al., 1995), both of which have been associated with increased comprehension of health information for making informed health decisions (Davis, Federman, & Wolf, 2009; IOM, 2014). Unlike education, age was not related to comprehension of composite scores, as we had hypothesized. It is possible that processing decrements that occur with age are offset by increased experience with medications and other life experiences, leading to possible exposure to composite score information.

We found that, after the educational intervention, participants who had greater comprehension of composite scores had greater trust that the benefits of the drug presented in the ad were true, they perceived that the drug would be more likely to work and have a greater magnitude of efficacy if it did work, and they perceived greater clarity of the benefit information in the ad . An understanding of composite scores involves recognition that the drug addresses only some and not all components of the score. It is possible that participants who had this understanding interpreted the ad with a lower bar of expectation, knowing that efficacy did not require improvement in each and every symptom. It is also possible that people familiar with composite scores were more comfortable with the ad as a whole, leading to overall positive reactions. This interpretation is supported by the finding that those who recalled more benefits also reported higher perceived clarity.

After the educational intervention, participants who understood composite scores also perceived a lower likelihood of risk from side effects. The association between composite score comprehension and risk perceptions is consistent with other findings that as perceptions of drug benefits increase, perceptions of risks decrease (O’Donoghue et al., 2014). The affect heuristic, which posits that a positive or negative feeling can be associated with the stimulus, affecting later judgments and decisions (Finucane, Alhakami, Slovic, & Johnson, 2000; Slovic & Peters, 2006; Slovic, Finucane, Peters, & MacGregor, 2007), may explain this finding.

Finally, we examined whether participants’ perceptions changed after viewing a definition and an example of a composite score in the context of the ad that we had presented. We found that perceptions about the clarity of the ad decreased, contrary to expectation, but perceptions of how effective the drug would be, how risky it was, and how much participants trusted the information did not change. The educational intervention may have caused participants to become aware of the complicated nature of the ad in a way they had not considered before. Given the complexity of composite scores, it is conceivable that this alone reduced perceived clarity. To borrow from the field of self-assessment (Kruger & Dunning, 1999), if people do not know what they do not know, then they may respond more positively on a measure of perceived clarity. As with all research, our studies had limitations. Study 1 included a small and nonrandom sample of individuals whose opinions may have been shaped by group dynamics. Study 2 employed a pre-post design and had a low response rate. We weighted our sample back to the US population, but this response rate should be considered when drawing conclusions from our research. Our second study examined responses to only one ad execution for one product. Future research should examine the understanding of composite scores in other conditions and executions. Finally, both studies rely on self-report rather than actual behavior. Future research that taps actual information-seeking behavior would be valuable.

The results of these investigations reveal that there are gaps in general knowledge both about FDA procedures in general and composite scores specifically. Knowing what authorities FDA has and whether ads are reviewed before dissemination may influence individuals’ opinion of the benefits and risks of advertised drugs and whether they should consult their health care professional about the product. This study now establishes three independent examinations (including Bell et al., 1999; Schwartz & Woloshin, 2011) that demonstrate there are widely held misconceptions about the FDA approval process and its authorities. If resources allow, it appears that an outreach program to spread accurate information about FDA’s processes would be valuable to help consumers make informed decisions about the drugs they see advertised. Moreover, understanding composite scores would assist consumers in accurately interpreting the claims of many DTC ads. Future research is needed to further explore ways to convey this common component of drug efficacy because it plays a large role in the advertising of the products, unbeknownst to many consumers.