Abstract
Background.
Physician-targeted prescription drug advertisements sometimes include price comparisons between products that may misleadingly imply equivalence of efficacy and safety or misrepresent true savings, suggesting the potential utility of a context statement to explain what the claims do and do not mean.
Methods.
We manipulated the presence of a price claim and a context statement in a 1 × 3 (control condition, price-comparison-only, price-comparison-plus-context) between-subjects design. Physicians (N = 1,438), randomly assigned to condition, viewed the prescription drug ad and answered a brief survey. Primary outcome measures included recognition, perceived importance, and impact of the price-comparison claim, and recognition, understanding, and effectiveness of the context statement.
Results.
The majority of physicians accurately recognized the price claim (76.0%) but far fewer accurately recognized the associated context statement (44.9%). The context statement did not affect evaluations of the price-comparison claim importance or accuracy and did not have the intended effects on perceptions of uncertainty about drug interchangeability. Physicians may be affected by price-comparison claims in thinking that the drug has risks that are relatively less severe. Price-comparison claims also affected intentions to look for information about the drug.
Conclusions.
Adding a realistic context statement to a physician-targeted prescription drug ad did not generate sufficient awareness of claim caveats to differentiate price-comparison response of those exposed to the context statement from those who were not.
Keywords: risk communication, health communication, pharmaceutical decision making, disclosure, prescription drug, advertising as topic, price comparisons
Introduction
Pharmaceutical companies sometimes opt to include sales promotion offers (e.g., coupons) or other financial incentives (e.g., rebates, free refills, free trial offers) or make price-comparison claims in advertising of prescription medications [1–4]. Johar [5] also has noted the predominance of explicit talk about price in interactions between prescribers and pharmaceutical representatives. How do physicians engage with price-comparison information appearing in advertisements directed toward them? Consumers appear to be sensitive to available information regarding prescription drug pricing [6], but it is unclear whether physicians are equally affected. A physician’s prescribing decisions should be based on a balance of considerations such as medical need, safety, efficacy, price, and reimbursements [7]. As such, price-comparison advertising may affect some physician perceptions relevant to decision making.
In addition, many physicians feel some pressure to comply with patient requests when prescribing medication [8]. For instance, Friedman and Gould [7] found that 43% of surveyed physicians agreed strongly or somewhat strongly that direct-to-consumer (DTC) advertising created tension between themselves and their patients and 73% felt that DTC advertising led patients to inappropriately pressure their physicians to prescribe drugs. Perhaps this explains why more than 45% of doctors surveyed will prescribe a patient-requested brand more than half of the time despite the fact that 66% of surveyed doctors felt that DTC ads created an incorrect preference for brands in cases where a generic would be sufficient [7].
Importantly, while these studies offer indirect support for an effect of advertising on physician prescribing behavior, we currently do not have direct empirical evidence of the impact of price-comparison information or disclaimers specifically in physician-targeted prescription drug ads on prescriber perceptions or behavior. To fill this gap, our study examined whether physicians engage with price-comparison information in prescription drug advertising and whether exposure to price-comparison information affects perceptions of drug risk and benefits and behavioral intentions.
Some studies have noted that physicians tend to underestimate the price of branded drugs and overestimate the price of generic drugs and that they also tend to prescribe advertised brands despite the availability of generic equivalents that might cost less [9–11]. These studies may suggest that price considerations can be overshadowed, or perhaps are not the most important consideration, when prescribing medication. However, other studies have argued that physician responsiveness to patient requests for medication is actually driven at least in part by prescriber sensitivity to patient concerns about price [12]. In addition, Cogdill and Nappi [13], who studied prescribing behavior, found that 86% of physicians checked a discounted medication list at least sometimes and 79% assessed a patient’s prescription insurance coverage at least once before writing a prescription.
Another reasonable concern is that price comparisons might imply equivalency between prescription drug products in terms of their efficacy and safety. Likewise, price comparisons might imply a fixed cost in terms of what a consumer or third-party will ultimately pay. In cases where equivalency in these domains has not been determined, a context statement inserted into an advertisement may offer clarification. Current literature does not offer robust empirical evidence of context statement effects, however. A number of researchers (e.g., Gilhooley [14]) have noted the potential importance of disclosure information for DTC advertising viewers, although none of those authors have explicitly focused on price-comparison disclosure information. The available literature also lacks robust discussion of the effects of disclosure in price-comparison advertising. Moreover, we have virtually no evidence of physician response to context statements in physician-targeted advertisements.
Assuming price and pricing information are at least somewhat salient to prescribers and hold potential to affect prescribing behavior, especially as prescribers consider monetary barriers to their patients’ medication acquisition and adherence, we need research to examine the exact relationship between the exposure to prescription drug price-comparison information and context statements intended to clarify claims and the prescriber’s responses to the advertised drug. In this article, our aim is to examine the extent to which physicians engage with price-comparison information presented in a prescription drug advertisement to treat diabetic neuropathy, the noticeability and understandability of a prominent context statement related to the price-comparison claim, and the impact of both the price-comparison claim and the context statement on physicians’ understanding of risks and benefits and their intentions to seek information and talk to their patients. Based on this aim, we proposed the following research questions:
RQ 1. Do physicians notice price comparison information in prescription drug ads?
RQ 2a. Do physicians notice the context statement in prescription drug ads?
RQ 2b. Do physicians comprehend the context statement in prescription drug ads?
RQ 3a. Does contextual information affect physicians’ perceptions of the price comparison claim?
RQ 3b. Is adding contextual information about efficacy and safety sufficient to correct physicians’ impressions that drugs are interchangeable?
RQ 4. What is the impact of price comparison information and context statement on physicians’ intentions and drug-related perceptions?
Methods
Stimuli
We conducted a randomized, controlled study exposing participants to one of three experimental conditions, each containing a physician-targeted two-page print ad for a fictitious prescription drug, called Veridan, designed to treat diabetic neuropathy. (The team developing the fictitious drug and a related drug profile included an advisor holding an MD degree and direct consultation with U.S. Food and Drug Administration staff holding PharmD degrees.) The study included a control condition featuring an ad that mentioned saving money in general: “Prescribing Veridan to treat diabetic peripheral neuropathy could save your patients money”; a price-comparison-only condition, where Veridan was compared with Lyrica, and the audience was told that switching to Veridan could save them money: “If you currently prescribe Lyrica to treat diabetic peripheral neuropathy, switching your patients to Veridan could save them up to $560.00 a year”;* and finally, a price-comparison-plus-context condition, where contextual information was included beneath the subheading: “The price savings presented may not reflect the actual savings by consumers or third-party payers. The products in this price comparison may or may not be equally effective or safe.” This latter condition is depicted in Figure 1.† The context statement was revised based on responses during cognitive testing (n = 9) and checked for comprehensibility during pretesting (n = 280). The additional contextual information was designed to be a realistic example of how disclosures might appear in real ads, but some steps were taken to highlight that information. For example, the size of the disclosure type face was the same as the surrounding text, the text was bolded, and an asterisk was used to link the context statement with the price comparison in the subheading. The manipulations were located on the top half of the first page of the ad directly underneath the headline. The second page of the ad, which included the brief summary page, was identical across the three conditions.
Figure 1.
Stimuli used in price-comparison-plus-context condition.
Participants
We identified potential participants through a research partner’s existing online health care panel through Research Now®. Physicians who indicated on their panel profile that they spoke English and their primary medical specialty was family practice, general practice, internal medicine, or endocrinology were targeted for the study. Research Now verifies a physician’s practicing status against association and governmental databases such as the Drug Enforcement Agency number (DEA#) and the American Medical Association Medical Education Number (ME#). Sixty-seven percent of the physician sample came from the health care panel. About 33% of the sample came from Research Now’s general panel. A screener item confirmed whether physicians specialized in primary care, internal medicine, or endocrinology in order to be eligible. Once a sample was selected, e-mail invitations were sent encouraging panelists to participate in the study. Members who were interested clicked on a secure, nonidentifiable hyperlink within the email, read an initial informed consent screen, answered screener items, and completed an informed consent form. Respondents were then randomly assigned to one of the three experimental arms, viewed the stimuli, and completed the web-based survey. Endocrinologists were offered an honorarium of $45, and primary care physicians were offered an honorarium of $35 for completing the survey.
Our sample included 1,438 physicians. Physicians were primarily male (66.2%, 952/1,437), White (68.2%, 909/1,332), non-Hispanic (95.4%, 1,271/1,332), between the ages of 45 and 59 (43.2%, 618/1,432) and had been a health care provider for 18 years on average. Eighty-four percent were primary care physicians and the rest were endocrinologists. About 37% of the patients physicians treated had diabetes. A detailed summary of physicians’ demographic characteristics appears in Table 1. Participant flow through the study is shown in Figure 2.
Table 1.
Provider Demographic Characteristics
| Characteristics | Control n (%) | Price-comparison-only n (%) | Price-comparison-plus-context n (%) | Total n (%) |
|---|---|---|---|---|
|
| ||||
| Gender | n=505 | n=478 | n=454 | n=1,438 |
| Male | 346 (68.5%) | 306 (64.0%) | 300 (66.1%) | 952 (66.2%) |
| Female | 159 (31.5%) | 172 (36.0%) | 154 (33.9%) | 485 (33.8%) |
| Age | n=503 | n=477 | n=452 | n=1,432 |
| 18–29 | 6 (1.2%) | 7 (1.5%) | 5 (1.1%) | 18 (1.3%) |
| 30–44 | 156 (31.0%) | 162 (34.0%) | 176 (38.9%) | 494 (34.5%) |
| 45–59 | 229 (45.5%) | 214 (44.9%) | 175 (38.7%) | 618 (43.2%) |
| 60+ | 112 (22.3%) | 94 (19.7%) | 96 (21.2%) | 302 (21.1%) |
| Race/Ethnicity | n=471 | n=440 | n=421 | n=1,332 |
| White, non-Hispanic | 329 (69.9%) | 300 (68.2%) | 280 (66.5%) | 909 (68.2%) |
| Black, non-Hispanic | 7 (1.5%) | 11 (2.5%) | 14 (3.3%) | 32 (2.4%) |
| Hispanic | 25 (5.3%) | 19 (4.3%) | 17 (4.0%) | 61 (4.6%) |
| Other, non-Hispanic | 106 (22.5%) | 105 (23.9%) | 105 (24.9%) | 316 (23.7%) |
| Multiracial, non-Hispanic | 4 (0.9%) | 5 (1.1%) | 5 (1.2%) | 14 (1.1%) |
| Physician Type | n=506 | n=478 | n=454 | n=1,438 |
| Primary Care Physician | 431 (85.2%) | 404 (84.5%) | 371 (81.7%) | 1,206 (83.9%) |
| Endocrinologist | 75 (14.8%) | 74 (15.5%) | 83 (18.3%) | 232 (16.1%) |
|
| ||||
| Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | |
|
| ||||
| Years practicing | 18.3 (9.75) | 17.4 (9.15) | 17.2 (9.42) | 17.7 (9.46) |
| Proportion of patients with diabetes | 0.36 (0.21) | 0.37 (0.22) | 0.38 (0.21) | 0.37 (0.21) |
Figure 2.
Consort diagram.
Measures
Recognition of price-comparison and context statement
Participants were provided with eight statements, including the price-comparison claim and the context statement. We asked them to select from the list all of the statements that were in the prescription drug advertisement they saw. Thus, two separate items captured correct recognition of the price-comparison claim (yes/no) and correct recognition of the context statement (yes/no).
Perceived importance of the price-comparison claim
Participants were shown the price-comparison claim via the questionnaire and were asked to rate how important that claim would be if they were deciding whether Veridan is a good option for their patients. Responses ranged from 1 = not at all important to 6 = very important.
Context statement comprehension
As a measure of physicians’ understanding of the context statement, they were asked to explain the statement in their own words. We coded the open-ended responses to create a measure of correct interpretation (correct/incorrect). To examine whether certain parts of the context were more salient, for those with correct interpretation, we also captured whether physicians mentioned both price and efficacy/risk elements in their explanation, whether they only mentioned price savings, or whether they only mentioned efficacy/risk.
Effectiveness of context statement
Multiple single-item measures were used to examine if the context statement had the intended effects. First, we examined if the context statement was able to temper perceptions of the price-comparison claim by asking physicians in the price-comparison-only and price-comparison-plus-context conditions how accurate they believed the price-comparison claim was, ranging from 1 = not at all accurate to 6 = very accurate. As a direct measure of whether physicians understood the portion of the context statement related to efficacy/risk, we asked participants to indicate their level of agreement with the statement “the ad does not provide enough information for me to know if Veridan and Lyrica are interchangeable.” The responses ranged from 1 = strongly disagree to 6 = strongly agree. We also included two other proxy measures for context statement effectiveness: a question measuring perceived drug efficacy and another question measuring perceived drug safety. These items were analyzed separately and asked participants (1) whether Veridan was better or worse at relieving diabetic nerve pain than Lyrica and (2) how risky or safe Veridan was compared with Lyrica. Each item had a 7-point response option scale that ranged from 1 = Veridan is much worse to 7 = Veridan is much better or 1 = Veridan is much safer to 7 = Veridan is much riskier. For each item, the middle point was labeled “They are the same.” Importantly, participants were also shown a box underneath the response options that said “I don’t know based on the ad.” This was the key focus of the questions and served as a proxy for context statement effectiveness. As evidence of the context statement working as intended, we expected that a larger percentage of physicians who saw the context statement would choose that response. When analyzing these items, we only focused on the percentage of physicians who chose that response in the price-comparison-only and price-comparison-plus-context conditions.
We also examined unintended effects of price-comparison claim and the correction for those perceptions in the presence of a context statement by asking physicians about their perceived efficacy and risks associated with Veridan. Perceived efficacy was measured by asking participants to indicate “How much relief does Veridan give people from their diabetic nerve pain?” Responses could range from 1 = very little relief to 6 = complete relief. We also assessed perceived risk by asking participants to report on a scale ranging from 1 = not at all serious to 6 = very serious: “Overall, Veridan’s side effects are …” Finally, we asked participants to indicate the likelihood that they would (1) look for more information about Veridan and (2) talk with patients about Veridan after seeing the Veridan ad. Response options for each item ranged from 1 = not at all likely to 6 = very likely.
Analyses
Analyses were conducted using SAS 9.4 and SPSS 23.0. We used an iterative multiple-coder approach to establish acceptable intercoder reliability for our open-ended item (based on Krippendorff’s alpha ≥.70, as per Hayes and Krippendorff [15]).
We conducted one-way analysis of variance tests (ANOVAs) to examine the effectiveness of the context statement on accuracy and importance perceptions of the price-comparison claim, drug interchangeability perceptions, general drug perceptions, and physician intentions. If the omnibus ANOVA test was significant, we conducted pairwise comparisons to identify significant differences between specific experimental arms focusing on those in the price-comparison-only condition versus those in the price-comparison-plus-context condition. We used a Bonferroni-adjusted significance threshold of 0.0167 based on the three comparisons between the three ad conditions.
Additionally, to assess the effectiveness of the context statement on participants’ perceived comparative efficacy and safety, we used Z-tests to examine the percentage of participants in the price-comparison-only condition versus those in the price-comparison-plus-context condition who indicated they “did not know based on the ad” whether Veridan had higher or lower efficacy/risk compared with Lyrica.
Finally, we conducted ordinary least squares regressions to test the predictive power of price importance perceptions on drug-related intentions and whether the effect of price importance varied by experimental condition. In the first step of the analysis, we entered the main effect of price importance and experimental condition, followed in the second step by the interaction between price importance and experimental condition.
Results
Recognition of price-comparison claim (RQ 1)
The majority of physicians recognized the price-comparison claim; about 78.0% (373/478) of physicians in the price-comparison-only condition and 73.8% (335/454) of physicians in the price-comparison-plus-context condition recognized the price-comparison claim.
Recognition and comprehension of context statement (RQ 2)
Less than half of physicians (44.9%, 204/454) in the price-comparison-plus-context condition correctly recognized the context statement. Among all conditions, when asked to imagine they were explaining the context statement to a patient (after being shown the statement via the survey), 78.6% (1,101/1,401) correctly mentioned at least one of the elements of the context statement. Of those, 40.0% (440/1,101) of physicians mentioned both price and efficacy/risk elements, a little more than half (57.6%, 634/1,101) mentioned only the portion about the price savings, and very few mentioned only efficacy/risk (2.5%, 27/1,101).
Effectiveness of context statement: perceived claim accuracy and importance and perceived comparative efficacy and safety (RQ 3)
We found no evidence to suggest that seeing the context statement affected how accurate (F(1, 930) = 0.30, p = .584) or important (F(2, 1,435) = 1.51, p = .221) physicians believed the price-comparison claim to be (Table 2). Next, we assessed the effectiveness of the context statement by examining physicians’ perceptions of whether the ad contained insufficient information to know or determine whether Veridan and Lyrica are interchangeable. If the context statement was effective at conveying uncertainty about the interchangeability of the two drugs, then participants shown the context statement should have greater agreement that the ad contained insufficient information to know if Veridan and Lyrica are interchangeable than participants who saw the price-comparison claim without the context statement. Thus, the focus of this analysis is on the price-comparison-only condition versus price comparison-plus-context condition. We did not find support for this proposition, F(2, 1,434) = 0.81, p = .444; physicians in the price-comparison-only and price-comparison-plus-context conditions reported similar levels of agreement that the ad contained insufficient information to know if Veridan and Lyrica are interchangeable (Table 2).
Table 2.
Means (SE) of Continuous Dependent Variables by Experimental Condition
| Experimental Condition Mean (SE) |
|||
|---|---|---|---|
| Dependent Variable | Control | Price-comparison-only | Price-comparison-plus-context |
|
| |||
| Perceived accuracy of price-comparison claim | — | 4.25 (0.05) | 4.21 (0.05) |
| Perceived importance of price-comparison claim | 4.79 (0.04) | 4.73 (0.05) | 4.67 (0.05) |
| Belief in insufficient information to judge interchangeability | 4.24 (0.06) | 4.29 (0.06) | 4.19 (0.06) |
| Perceived drug risk severity | 3.55 (0.05)ab | 3.41 (0.04) a | 3.58 (0.05) b |
| Perceived drug efficacy | 4.13 (0.04) | 4.06 (0.04) | 4.14 (0.04) |
| Intentions to look for more information | 4.24 (0.06) a | 4.47 (0.06) b | 4.43 (0.06)ab |
| Intentions to talk to patients about Veridan | 3.81 (0.06) | 3.95 (0.06) | 3.87 (0.06) |
NOTES: Perceived accuracy items could range from 1 = not at all accurate to 6 = very accurate. Perceived importance items could range from 1 = not at all important to 6 = very important. Belief in insufficient information could range from 1=strongly disagree to 6 = strongly agree. Perceived risk could range from 1 = not at all serious to 6 = very serious. Perceived efficacy could range from 1 = very little relief to 6 = complete relief. Intentions to look for information and intention to talk to patients could range from 1 = very unlikely to 6 = very likely.
Means in bold font with different superscripts are significantly different.
Additionally, we assessed effectiveness of the context statement by examining differences in perceived comparative efficacy and safety between those in the price-comparison-only condition and those in the price-comparison-plus-context condition. The assumption was that if the context statement was effective, it should convey to participants that they cannot know whether Veridan is better or safer than Lyrica based on the information in the ad. Therefore, we expected that participants in the price-comparison-plus-context condition should be more likely to choose “don’t know based on the ad” than participants in the price-comparison-only condition. We found that the difference between the percentage of physicians in the price-comparison-plus-context and price-comparison-only conditions that chose “I don’t know based on the ad” was borderline significant and counter to our expectations. Specifically, a larger percentage of participants in the price-comparison-only condition (46.4%, 222/478) vs. price-comparison-plus-context condition (40.1%, 182/454) reported not knowing based on the ad if Veridan had higher/lower efficacy compared with Lyrica (p = .050) (Table 3). In terms of the perceived comparative safety, there was a significant difference between participants in the price-comparison-only condition and those in the price-comparison-plus-context condition. Also, counter to expectations, a larger percentage of physicians in the price-comparison-only condition (34.7%, 166/478) indicated that they could not determine based on the ad whether Veridan was safer/riskier compared with physicians in the price-comparison-plus-context condition (28.4%, 129/454), p = .038. We also noted one other pattern among our experimental conditions relative to the control condition: on average, those in the control condition were most likely to indicate that they “don’t know” whether the products are interchangeable. This should not be altogether surprising because the control ad did not include any reference to Lyrica and instead only mentioned a general price savings.
Table 3.
Perceived Comparative Efficacy and Safety (Percentage) by Experimental Condition
| Experimental Condition N (%) |
|||
|---|---|---|---|
| Dependent Variable | Control± | Price-comparison-only | Price-comparison-plus-context |
|
| |||
| Perceived comparative safety* | |||
| Don’t know if Veridan is safer or riskier than Lyrica based on the ad | 183 (36.2%) | 166 (34.7%) a | 129 (28.4%) b |
| Perceived comparative efficacy* | |||
| Don’t know if Veridan is better or worse than Lyrica based on the ad | 254 (50.2%) | 222 (46.4%) | 182 (40.1%) |
NOTES:
Examined (using Z-tests) the percentage of participants in the price-comparison-only condition versus those in the price-comparison-plus-context condition who indicated they “did not know based on the ad” whether Veridan had higher or lower efficacy/risk compared with Lyrica.
The control group was included for descriptive purposes only; no statistical tests were conducted using the control group.
Percentages in bold font with different superscripts are significantly different.
Effectiveness of context statement: Intentions and general drug perceptions (RQ 4)
To examine whether the perceived importance of the price-comparison claim affected physicians’ drug-related intentions, we ran two separate regression models. In these models, we also examined whether experimental condition predicted intentions after controlling for perceived price-claim importance, and whether experimental condition moderated any relationship between price importance and intentions. In the regression models, price-comparison-only was the reference category. Across the two intentions measures, we found that perceptions of price importance affected physicians’ intentions (Table 4). Positive relationships were identified between perceived price-comparison claim importance and intentions to look for information about Veridan (B = 0.45, p < .001) and to talk with patients about Veridan (B = 0.45, p < .001). Compared to participants in the price-comparison-only condition, participants exposed to the control ad had lower intentions on average to look for information (B = −.26, p < .001) or to talk with their patients about Veridan (B = −.17, p = .034). The predictors accounted for 13.3% of the variance in intentions to look for information and 12.5% of the variance in intentions to talk to patients. Counter to predictions, adding the interaction between price importance and experimental condition in the second step of the equation was not associated with a significant increase in the variance explained in intention to look for information (R2Δ = .002, F(2, 1,432) = 1.34, p = .262) and intentions to talk to patients (R2Δ = .002, F(2, 1,432) = 1.43, p = .239).
Table 4.
OLS Regression Summary Results for Intentions to Look for Information and Talk to Patients about Veridan
| Look for information about Veridan | Talk with patients about Veridan | |||||||
|---|---|---|---|---|---|---|---|---|
|
|
||||||||
| B | SE | t | p | B | SE | t | p | |
|
| ||||||||
| Step 1 | ||||||||
| Price importance | .446 | .031 | 14.49 | <.001 | .447 | .031 | 14.21 | <.001 |
| Condition | ||||||||
| Control | −.255 | .077 | −3.31 | .001 | −.167 | .079 | −2.12 | .034 |
| Price-comparison-plus-context | −.017 | .079 | −.21 | .834 | −.055 | .081 | −.68 | .498 |
| Step 2 | ||||||||
| Price importance | .515 | .053 | 9.78 | <.001 | .518 | .054 | 9.61 | <.001 |
| Condition | ||||||||
| Control | −.253 | .077 | −3.29 | .001 | −.166 | .079 | −2.11 | .035 |
| Price-comparison-plus-context | −.019 | .079 | −.25 | .806 | −.058 | .081 | −.72 | .471 |
| Interaction Effect | ||||||||
| Price importance × Control | −.097 | .075 | −1.29 | .199 | −.088 | .077 | −1.15 | .250 |
| Price importance × Price-comparison-plus-context | −.114 | .075 | −1.51 | .130 | −.127 | .077 | −1.65 | .100 |
| Price importance × Price-comparison-only (reference) | — | — | — | — | — | — | — | — |
NOTES: Effective sample size for any individual analysis may differ based on the number of participants who answered the specific items.
Bolded p value indicates significant result.
Next, we examined whether exposure to price-comparison information affected intentions to look for more information about the drug or talk to patients about Veridan. Experimental condition had a significant effect on physicians’ intentions to look for more information about Veridan (F(2, 1,435) = 4.39, p = .013, f = .10) such that those who viewed the price-comparison claim without a context statement (M = 4.47) reported greater intentions to seek information about Veridan than those who saw the general price savings statement (M = 4.24, p = .005, d = 0.18). Physicians in the general savings information control condition reported lower intentions than those who saw the price-comparison claim with the context statement (M = 4.43) as well, but this difference was borderline nonsignificant based on a Bonferroni adjusted p value to account for multiple comparisons (p = .026). In addition, physicians who viewed a general price savings statement, those who viewed a price-comparison claim, and those who viewed a price-comparison claim and context statement reported similar intentions to talk with their patients about Veridan, (F(2, 1,435) = 1.43, p = .241) (Table 2).
Finally, we examined the two items related to perceived risk and perceived efficacy to determine whether seeing a specific price-comparison claim resulted in inflated efficacy or reduced risk perceptions compared with seeing just a general price savings statement or seeing a price-comparison claim with a context statement. There was a significant effect of experimental condition on risk severity perceptions, F(2, 1,435) = 3.81, p = .023, f = .07), such that physicians in the price-comparison-only condition (M = 3.41) compared with those in the price-comparison-plus-context condition (M = 3.58) reported lower overall severity of side effects (p = .012) (Table 2). Additionally, physicians seeing the price-comparison claim reported lower overall severity of side effects than those who viewed a general price savings statement, but this difference was borderline nonsignificant at our Bonferroni adjusted p-value (p = .026). We found no significant effect of experimental condition on efficacy perceptions, F(2, 1,435) = 1.17, p = .312).
Discussion
We examined the extent to which physicians engaged with price-comparison information presented in a prescription drug advertisement to treat diabetic neuropathy, the noticeability and understandability of a prominent context statement related to the price-comparison claim, and the impact of both the price-comparison claim and the context statement on physicians’ understanding of risks and benefits and their intentions to seek information and talk to their patients.
While a sizeable majority of physicians (76.0%) accurately recognized the price claim, less than half of physicians assigned to view the ad with a context statement (44.9%) recognized the presence of the context statement. The latter finding emerged despite the context statement being presented prominently in the top half of the ad in conjunction with an asterisk, reflecting what might be considered a best-case scenario. When asked to assess the context statement, the majority (78.6%) appeared to accurately understand what the context statement meant; however, 57.6% of those with the correct interpretation focused only on explaining the portion of the statement focused on price (meaning they tended to only note the price-related caveat and not the efficacy/risk stipulation). This finding may have resulted because the portion of the statement related to efficacy/risk appeared after the sentence about price. The overall lesson here may be that physicians are unlikely to attend to supplementary contextual information, and in cases when they do, their attention may be limited to the first concept communicated. Health care providers often maintain exceptionally busy schedules (e.g., Baron [16]); so limiting attention to contextual details in ads may be an adaptive response.
This pattern of context statement engagement (or lack thereof) helps explain a general lack of experimental-condition differences on other outcomes. Physicians perceived the price-comparison claim to be accurate and important, and the addition of the context statement did not temper these evaluations: we found no statistically significant differences between ad conditions. Additionally, seeing a context statement did not moderate the effect of price-comparison claim importance on intentions to look for information and talk to patients about the drug.
Overall, adding the context statement did not encourage more uncertainty about the interchangeability of the two drugs despite such uncertainty being one of the intended effects of the context statement. Physicians in the price-comparison-only and the price-comparison-plus-context conditions reported similar levels of agreement that the ad lacked the necessary information to assess whether Veridan and Lyrica were interchangeable. However, when we examined interchangeability related to safety and efficacy specifically, we also found some unexpected results. The difference in percentage of physicians in price-comparison-plus-context and price-comparison-only conditions who reported uncertainty regarding Veridan’s comparative efficacy was borderline significant; unexpectedly, a higher percentage of physicians in the price-comparison-only condition (46.4%) than physicians in the price-comparison-plus-context condition (40.1%) reported uncertainty related to efficacy. We also found that a higher percentage of physicians in the price-comparison-only condition (34.7%) than those in the price-comparison-plus-context condition (28.4%) reported not knowing based on the ad whether Veridan was safer/riskier. In sum, the addition of the context statement to the price-comparison claim yielded a mixed effect on perceptions of product interchangeability related to safety or efficacy.
With respect to the general risk and efficacy perceptions, we did not find evidence to suggest that a price-comparison claim, with or without the context statement, inflates efficacy perceptions compared with ads with a general price savings statement. However, adding the context statement may have led physicians to be more cautious regarding Veridan’s serious side effects in that participants seeing price-comparison information without the additional context actually reported lower perceived seriousness of Veridan’s side effects than people seeing the context statement. Without the context statement, physicians may have been swayed by price-comparison information to think that the drug risks are relatively less severe than they would otherwise believe.
Finally, although adding the context statement did not specifically affect physicians’ behavioral intentions, including price-comparison information overall did in comparison to showing physicians a more general savings claim. Participants who saw the price-comparison claim (with or without the context statement) compared with those who saw a general price savings statement were more likely to intend to look for information about the drug, although they did not intend to go so far as to talk to their patients about the drug. This finding could reflect some caution by physicians who, when seeing price-comparison information, would want to gather more information about the drug before potentially discussing it with their patients.
Our experimental design, realistic stimuli, and adequate sample size allowed us to answer our general research questions in a useful way. Nonetheless, the study has a number of limitations that we should note. First, we asked participants to respond to a series of questions to gauge their perceptions and reactions immediately after viewing the advertising in question. It is conceivable that ad effects would fade or change over time after exposure, suggesting that our indicator of short-term ad effect may not generalize to long-range effects. Our outcome measures also are self-reported and do not include evidence of actual prescribing behavior subsequent to the study (a possibility constrained by the fact that the advertised drug was fictitious). Additionally, dependent measures in this study comprised single items rather than composites, which is an important limitation of our measurement approach. Future research should investigate the potential for multiple-item measures of our constructs to deliver improved reliability and validity. As another limitation, our sample was limited in that it included physicians but not nurse practitioners, physician’s assistants, or other health care professionals. Finally, we administered this study online, which means we lacked some control over the exact circumstances in which physicians participated in the study. For example, when individuals complete a study in person, we can verify that they completed the study in one sitting or verify the nature of any breaks taken. Respondents in our study were able to walk away, attend to patients, or become distracted, which could potentially affect responses. However, we examined the impact of study time and presented evidence suggesting that the amount of time taken to complete the survey was unlikely to produce biased results.
Although price-comparison claims can be found in prescription drug advertising, little is known about the true prevalence of such claims or how price information should be conveyed. Including comparative price information could facilitate physicians’ ability to provide cost-effective, high-value care [17]. However, any potential benefits of price-comparison claims would be negated if such claims led to incorrect inferences about risk and efficacy or if such claims omit information to support or clarify claims. When noticed, the context statement had some of the intended effects and was able to temper potential misleading impressions. However, given the mixed results in this study, including a high percentage of physicians who either did not remember seeing or could not accurately report the entire content of the context, an evaluation that such disclosures are operating as intended may be necessary to ensure they are having their intended effect.
Conclusions
In light of price-comparison claims as an advertising strategy, context statements may offer relevant caveats regarding the claims. Current literature lacks empirical evidence on viewer reactions to such context statements generally, and specifically with regard to physician populations. We developed realistic prescription drug print advertising promoting a fictitious diabetic neuropathy drug, and we experimentally tested the effects of showing physicians a context statement intended to clarify the implications of a price-comparison claim. Although inclusion of specific price-comparison information appears to affect physician response to the advertised drug to some extent, we did not find evidence that inclusion of a context statement affected physician understanding or uncertainty about product efficacy or actual cost to patients in ways consistent with the intended effects of the context statement. A majority of physicians did not notice or recognize the context statement, which likely undermined its general effect.
ACKNOWLEDGMENTS
We would like to thank Kayla Gray and Maria Ashbaugh of RTI International for creating the fictitious drug advertisements.
FUNDING DETAILS
Funding was provided by the Office of Prescription Drug Promotion, Center for Drug Evaluation and Research, U.S. Food and Drug Administration.
Biographies
Kevin R. Betts is a Social Science Analyst in the Food and Drug Administration’s Office of Prescription Drug Promotion, where he provides research and consulting services pertaining to prescription drug communications.
Kathryn J. Aikin is a Senior Social Science Analyst and the Research Team Lead in the Food and Drug Administration’s Office of Prescription Drug Promotion. Dr. Aikin’s research has focused on topics related to promotion of prescription drugs.
Vanessa Boudewyns is a Social Scientist in RTI International’s Center for Communication Science. Her work takes an interdisciplinary approach to the study of message design and health behavior prediction and change.
Mihaela Johnson is a Social Scientist in RTI International’s Center for Communication Science whose research focuses on evaluating the effectiveness of communication strategies and health messages in the media environment.
Alex Stine is a doctoral student with Duke University’s Department of Psychology and Neuroscience. Previously, he was an analyst in RTI International’s Center for Communication Science.
Brian G. Southwell directs the Science in the Public Sphere program in the Center for Communication Science at RTI International. He also is a faculty member at Duke University and the University of North Carolina at Chapel Hill.
Footnotes
Lyrica was chosen because it is also indicated for treatment of diabetic neuropathy.
Full size high-definition copies of stimuli from all experimental conditions are available by request directed to the corresponding author.
DISCLOSURE STATEMENT
The authors report no conflicts of interest.
ETHICS APPROVAL
The study presented in this article was provided an exemption from FDA’s Research Involving Human Subjects Committee under 45 CFR 46.101(b)(2). All participants provided their informed consent to participate in the research.
Contributor Information
Kevin R. Betts, U.S. Food and Drug Administration, 10903 New Hampshire Ave, Silver Spring, MD 20993.
Kathryn J. Aikin, U.S. Food and Drug Administration, 10903 New Hampshire Ave, Silver Spring, MD 20993.
Vanessa Boudewyns, RTI International, 3040 E. Cornwallis Road, P.O. Box 12194, Research Triangle Park, NC 27709.
Mihaela Johnson, RTI International, 3040 E. Cornwallis Road, P.O. Box 12194, Research Triangle Park, NC 27709.
Alex Stine, RTI International, 3040 E. Cornwallis Road, P.O. Box 12194, Research Triangle Park, NC 27709.
Brian G. Southwell, RTI International, 3040 E. Cornwallis Road, P.O. Box 12194, Research Triangle Park, NC 27709.
References
- [1].Bhutada NS, Cook CL, Perri M 3rd. Consumers responses to coupons in direct-to-consumer advertising of prescription drugs. Health Mark Q. 2009;26:333–346. [DOI] [PubMed] [Google Scholar]
- [2].Palmer JA, Timm AR, Neumann PJ. Drug company advertising in medical journals about the health-economic advantages of their products for 2000–2006 versus 1990–1999. J Manag Care Pharm. 2008;14:749–755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Woloshin S, Schwartz LM, Tremmel J, et al. Direct-to-consumer advertisements for prescription drugs: what are Americans being sold? Lancet. 2001;358:1141–1146. [DOI] [PubMed] [Google Scholar]
- [4].Weppner WG, Hollon MF, Chew LD, et al. Direct-to-consumer offers for free and discounted medications on the internet: a content analysis of “e-samples”. Arch Intern Med. 2009;169:2024–2030. [DOI] [PubMed] [Google Scholar]
- [5].Johar K An insider’s perspective: defense of the pharmaceutical industry’s marketing practices. Albany Law Rev. 2012;76:299–334. [PubMed] [Google Scholar]
- [6].Glasgow C, Schommer JC, Gupta K, et al. Promotion of prescription drugs to consumers: case study results. J Manag Care Pharm. 2002;8:512–518. [DOI] [PubMed] [Google Scholar]
- [7].Friedman M, Gould J. Physicians’ attitudes toward direct-to-consumer prescription drug marketing. J Med Market. 2007;7:33–44. [Google Scholar]
- [8].Becker SJ, Midoun MM. Effects of direct-to-consumer advertising on patient prescription requests and physician prescribing: a systematic review of psychiatry-relevant studies. J Clin Psychiatry. 2016;77:e1293–e1300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Allan GM, Lexchin J, Wiebe N. Physician awareness of drug cost: a systematic review. PLoS Med. 2007;4:e283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Korn LM, Reichert S, Simon T, et al. Improving physicians’ knowledge of the costs of common medications and willingness to consider costs when prescribing. J Gen Intern Med. 2003;18:31–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Reichert S, Simon T, Halm EA. Physicians’ attitudes about prescribing and knowledge of the costs of common medications. Arch Intern Med. 2000;160:2799–2803. [DOI] [PubMed] [Google Scholar]
- [12].Czechowski JL, Tjia J, Triller DM. Deeply discounted medications: implications of generic prescription drug wars. J Am Pharm Assoc. 2010;50:752–757. [DOI] [PubMed] [Google Scholar]
- [13].Cogdill B, Nappi JM. Assessment of prescribers’ knowledge of the cost of medications. Ann Pharmacother. 2012;46:200–207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Gilhooley M Heal the damage: prescription drug consumer advertisements and relative choices. J Health Law. 2005;38:1–44. [PubMed] [Google Scholar]
- [15].Hayes AF, Krippendorff K. Answering the call for a standard reliability measure for coding data. Commun Methods Meas. 2007;1:77–89. [Google Scholar]
- [16].Baron RJ. What’s keeping us so busy in primary care? A snapshot from one practice. N Engl J Med. 2010;362:1632–1636. [DOI] [PubMed] [Google Scholar]
- [17].Riggs KR, DeCamp M. Providing price displays for physicians: which price is right? JAMA. 2014;312:1631–1632. [DOI] [PMC free article] [PubMed] [Google Scholar]