Abstract
Purpose: Exhibit hall booths at medical conferences are a key avenue for promoting prescription drugs to health care providers (HCPs). Because HCPs spend considerable time interacting with industry representatives at conferences, we explored how representatives’ credentials might influence HCP perceptions and prescribing intentions of promoted drugs. We also examined how disclosures of clinical trial data limitations about these drugs during conference interactions might influence HCP perceptions and intentions.
Methods: We conducted a 2 × 2 factorial experimental study with HCPs (n = 430) during or immediately after their attendance at 1 of 12 US medical conferences. Participants viewed video stimuli depicting an exhibit hall interaction between an industry representative and an HCP discussing a fictitious drug for preventing nausea and vomiting. Participants were randomly assigned to 1 of 4 experimental conditions that varied (1) the representative’s educational credentials (MBA vs MD) and (2) a disclosure noting clinical trial data limitations (present vs absent). Participants then completed an online questionnaire with questions about the fictitious drug, such as perceived efficacy and perceived risk.
Results: Industry representative credentials had no influence on HCP perceptions and intentions to prescribe the drug, though representatives with medical degrees were rated as having slightly more medical knowledge. Conversely, the disclosure significantly reduced drug efficacy perceptions and led to less positive drug attitudes, although it did not influence prescribing intentions.
Conclusions: The findings suggest that HCP perceptions and intentions are not swayed by the industry representative credentials but that data limitation disclosures can temper HCP perceptions of drugs promoted at medical conferences.
- Conferences
- Credentialing
- Disclosure
- Exhibition
- Health Communication
- Intention
- Pharmaceutical Economics
- Pharmaceutical Industry
- Pharmaceuticals
Introduction
Pharmaceutical companies spent $20.3 billion promoting their products to health care providers (HCPs) in 2016, accounting for 88% of their marketing budgets.1 A sizeable portion of their expenditures on HCP-directed promotion went toward participation at national medical conferences where companies use sponsorship for product exposure, including sharing information about their products in presentations and exhibit halls. Almost 87% of physicians attend at least one medical conference annually, with 77% of attendees visiting exhibit hall booths sponsored by pharmaceutical or biotech companies.2,3 HCPs typically spend 1 hour each day in exhibit halls and up to 21 minutes at each booth,3,4 comparatively longer than detailing visits in HCP offices.5,6 Although many factors have changed recently because of the COVID-19 pandemic, exhibit booths continue to help pharmaceutical companies market products to many HCPs and engage in relatively lengthy discussions. The current study uses an experimental design to examine how the characteristics of these exhibit hall booths—such as the educational credentials of the industry representatives and the disclosure of data limitations—influence HCP perceptions of drugs promoted at the booths.
Evidence suggests that source credibility and disclosures may affect HCP prescription drug perceptions in the context of conference exhibit halls. Decades of research have demonstrated that higher levels of perceived source credibility—typically operationalized as both trustworthiness and expertise—are associated with greater persuasiveness and attitude changes in information recipients.7–12 Source qualifications and credentials are one element that recipients use to assess a source’s credibility,13 and perceived similarity between the information source and recipient can lead to higher credibility perceptions.14–16 Other studies have examined the effects of endorsers with professional expertise versus those with product experience on attitudes toward the brand and promotion,17,18 finding that individuals distinguish between expertise and experience but that only expertise affects product perceptions. Based on previous research,18 we examined whether the professional and product expertise of a representative, as operationalized by profession, affected attitudes and perceptions of the promoted product.
Past research also has demonstrated that disclosures about prescription drug limitations can influence HCP perceptions and intentions. For example, researchers19 found that HCPs, as expected, reported weaker intentions to prescribe a drug after viewing a disclosure about unsupportive data of off-label use. Further research20 determined that disclosures can inform HCPs about the limitations of sales aid data displays, but these changes in perceptions did not influence prescribing intentions. Given the situation-specific influence of disclosures, it is not clear how disclosures might affect HCP perceptions of promoted drugs in an exhibit hall context.
We conducted an experimental study to answer 2 research questions:
Research Question 1: How do the credentials of an industry representative (eg, MD vs MBA) at an exhibit hall booth influence HCP perceptions of promoted drugs?
We hypothesized that HCPs exposed to a representative with a medical degree will (1) rate the representative as having more positive personal characteristics, (2) perceive the promoted drug as having higher efficacy and lower risk, and (3) have stronger intentions to prescribe the promoted drug.
Research Question 2: How does the presence or absence of a clinical trial data limitations disclosure influence HCP perceptions of promoted drugs?
We hypothesized that HCPs who view a disclosure will (1) perceive the promoted drug as having lower efficacy, (2) have less positive attitudes toward the drug, and (3) have weaker intentions to prescribe the drug.
Methods
Study Design
As part of a larger study regarding medical conference attendance in general, we conducted a 2 × 2 factorial experimental study that enrolled HCPs who attended 1 of 12 selected medical conferences and agreed to complete an online survey within 7 days of their conference attendance. This experiment, independent of subsequent survey questions, was the first task participants completed. We randomly assigned eligible HCPs to 1 of 4 experimental conditions, each depicting a different version of a video simulating a conference exhibit hall interaction between an industry representative and an HCP discussing a fictitious prescription drug for preventing nausea and vomiting. The experimental conditions varied (1) the educational credentials of the industry representative (MBA vs MD) and (2) the disclosure of clinical trial data limitations (present vs absent). These experimental manipulations were delivered both verbally and visually in the videos. After viewing the stimuli, study participants answered questions about the fictitious drug, such as perceived efficacy, perceived risk, and intention to prescribe.
Stimuli
We developed a profile for a fictitious prescription drug indicated to treat nausea and vomiting intended to be relevant across multiple medical specialties. We developed trade and generic names for the drug (Nomestra; pravasetron); created professional‐quality branding; specified the drug’s indications, risks, and benefits based on real medications indicated to treat the same conditions; and created charts depicting fictitious clinical study efficacy data for the drug. The fictitious drug’s profile was reviewed by our medical advisor for accuracy and realism.
We next created 4 versions of a video simulating a conference exhibit hall interaction between an industry representative and an HCP, with video durations ranging from 2 minutes and 37 seconds to 3 minutes depending on the experimental condition. The videos were identical in every way (eg, actors, script, setting, background attendee movements, scene changes) except for the verbal and visual manipulations of (1) the industry representative’s credentials and (2) the disclosure of data limitations. The representative’s credentials were presented verbally during a conversation with the HCP about attending the same college, and the credentials were presented visually during a close-up view of the representative’s business card (Figure 1). In conditions where the data disclosure was present, the representative verbally reads and explains the disclosure when presenting a poster with efficacy data, which includes the disclosure as a footnote (Figure 2). Study participants were also shown a close-up of the disclosure on a flyer. The disclosure emphasizes the retrospective nature of reporting drug efficacy and read: “In clinical studies, patients retrospectively reported nausea and vomiting severity up to 2 weeks following initiation of treatment.” The efficacy data shown on the poster and flyer were the same across all experimental conditions.
Visual Manipulation of Industry Representative Credentials.
Disclosure Manipulation (Poster with Disclosure Present).
Questionnaire and Measurements
We developed a questionnaire with closed- and open-ended items that mapped to our hypotheses. Before the study, we conducted cognitive interviews with HCPs (n = 9) to assess the clarity and flow of the video and the questionnaire. We also conducted a pretest with HCPs (n = 25) to assess data collection procedures, random assignment functionality, and programming functionality, as the survey was programmed for online administration, with stimuli embedded as the first task. Participants could not progress in the survey until the video ran completely. Participants confirmed they viewed the video by self-report, and then completed the questionnaire.
We assessed exposure to experimental manipulations by measuring participant recall of both the representative’s credentials and the disclosure. We measured perceptions of the representative by asking participants to rate on a 6-point semantic differential scale the extent to which the representative was sincere (insincere), trustworthy (untrustworthy), expert (not an expert), experienced (inexperienced), friendly (unfriendly), pushy (not pushy), and a good (or bad) communicator. We averaged ratings to create a composite score of perceptions of the representative. We separately asked participants to rate the representative’s product expertise and medical expertise, respectively, on scales from 1 (very little expertise) to 5 (great deal of expertise). We presented the disclosure to all participants in the questionnaire and asked them to rate the statement’s importance when deciding whether the drug was a good option for patients on a scale of 1 (not at all important) to 6 (extremely important) the statement.
We asked participants to rate the fictitious drug on whether they perceived its risks as (1 (not at all) to 6 (extremely) serious and whether they 1 (strongly disagree) to 6 (strongly agree)) that the advertised drug is safer than similar medications. We also asked participants to rate the drug’s perceived efficacy (absolute; compared with similar medications). We measured drug attitudes by asking participants to rate on a semantic differential scale from 1 to 6 the extent to which they perceived the drug as useful (not useful), helpful (not helpful), and a good (bad) option, and we averaged items to create a composite score. Finally, we asked participants to rate (scale 1 to 6) their intention (1 (not at all likely) to 6 (extremely likely)) to prescribe the drug based only on the information in the video.
Sampling Frame and Eligibility
We used a stepwise approach to select medical specialties, medical conferences, and participants for the study. First, we selected 12 medical specialties to increase the generalizability of the findings (Table 1). We prioritized specialties that had a high number of branded medications, high volume of prescriptions, and robust drug development and promotional spending.
Medical Conferences Through Which Participants Were Recruited and Data Collected
Second, we identified medical conferences within the selected specialties, resulting in a list of 88 potential conferences. We narrowed the list by applying eligibility criteria: (1) historic attendance of 5000+ individuals; (2) primary audience of prescribers and clinicians; (3) scheduled during data collection time frame; and (4) located within the United States. Applying these criteria yielded 26 eligible conferences, and we selected 12 conferences (1 per specialty) for the study.
Finally, participants were eligible if they attended (or were currently attending) one of the selected conferences within the past 7 days (in-person or virtually), were authorized to prescribe medications (excluding pharmacists), and typically spent 20% or more time on direct patient care. We excluded individuals who could not read and speak in English; who reported having vision or hearing problems that would interfere with stimuli exposure; who were currently employed by the US government or a pharmaceutical/biotech company; or who were enrolled in another study wave. To recruit participants, we sent e-mail invitations to a random sample of HCPs registered with Medscape/WebMD’s online HCP survey panel, targeting panelists in the selected medical specialties. Participants were offered a $50 electronic Amazon.com gift card for completing the study.
Data Collection
We collected data at discrete timepoints during February – November 2022 (Table 1). Data collection periods were timed to coincide with the selected conferences, up to 7 days after each conference ended. Within each data collection wave, panelists who met the eligibility criteria and consented were randomly assigned to 1 of the 4 experimental conditions using permuted block randomization. This approach allowed for random assignment to conditions while maintaining balance across conditions.21
Data Analysis
For scale-based composite variables, we calculated internal reliability using Cronbach’s α (threshold 0.75). We also checked bivariate associations between potential covariates (ie, age, sex, race/ethnicity, years in practice, prescription volume) and outcome variables using Pearson’s correlation and Eta (threshold r ≥ 0.30 and η ≥ 0.24, respectively).22 No variables met the threshold for multivariate analysis.
To test hypotheses, we conducted two-way ANOVAs and logistic regressions to examine the main effects of the 2 experimental factors and their interaction on continuous and categorical outcome variables, respectively. If an interaction was significant, we conducted planned comparisons to examine differences between experimental groups using a Bonferroni-adjusted P-value of 0.0167. If an interaction was not significant, we examined only the main effects using a P-value of 0.05. In cases where homogeneity of variance was not satisfied, we used a restricted maximum likelihood function, which takes into consideration unequal variances across groups. In these instances, we report the effect size, ƞ2, which should be viewed as an approximation.
Results
Participant Characteristics
A total of 430 participants completed the study, evenly spread across the 4 experimental conditions as well as the 12 medical specialties (n = 107 to 108 per experimental condition; n = 34 to 36 per specialty). Participants were reasonably varied on demographic and professional characteristics (Tables 2 and 3). Most participants were physicians (93%), with some physician assistants (6%) and nurse practitioners (1%). Participants most commonly worked in private/group practices (44%), academic hospitals (34%), and/or community hospitals (30%), and three-quarters of participants (75%) reported having a patient load of 26 to 75 individuals per week. Experience in clinical practice ranged from 1 to 50 years.
Participant Demographic Characteristics
Participant Professional Characteristics
To characterize our sample relative to practicing US clinicians, we compared participants’ demographic characteristics against the demographic benchmarks reported by the Association of American Medical Colleges (AAMC) (2021).23 The sex breakdown of participants was similar to AAMC benchmarks for all medical specialties except endocrinology (participants = 69% male vs benchmark = 47% male) and rheumatology (participants = 69% male vs benchmark = 53% male). Our sample had a smaller percentage of Black/African-American HCPs in 10 out of 12 specialties; had a smaller percentage of Asian HCPs in 7 out of 12 specialties; and had a smaller percentage of Hispanic HCPs in 3 out of 12 specialties. On average, participants were slightly younger than practicing clinicians.
Recall of Experimental Manipulations
Recall of the representative’s credentials was high, with 91% in the MD conditions and 84% in the MBA conditions correctly recalling the credentials (P = .042). Disclosure recall among those exposed to it was generally high (79%), although nearly a third of participants in the nondisclosure conditions incorrectly reported seeing it (29%) (P < .001).
Research Question 1: Industry Representative Credentials
Credentials and Perceived Attributes of Industry Representative
Contrary to our hypothesis, we found that participants rated medical and business representatives similarly on positive characteristics (credible, knowledgeable, likable, familiar, articulate). The average rating for medical representatives was 5.06 versus an average rating of 4.95 for business representatives (P = .222). When asked specifically about the representative’s product expertise, participants rated medical and business representatives similarly (4.27 vs 4.33, P = .321). However, participants did rate medical representatives as having significantly higher medical expertise (4.20 vs 3.82, P < .001, η2 = 0.04), though the effect size was small.
Credentials and Perceived Drug Efficacy
Our second hypothesis was that HCPs exposed to a medical representative would perceive the promoted drug as having higher efficacy than those exposed to a business representative. However, we found that participant ratings of absolute drug efficacy and comparative drug efficacy did not differ based on credentials (4.68 vs 4.68, P = .989 and 4.48 vs 4.40, P = .355, respectively) (Figure 3).
Effect of Industry Representative Credentials on Participant Perceptions of Promoted Drug.
Credentials and Perceived Drug Risk
When we examined the relationship between representative credentials and perceived drug risk, we again found that participants perceived the absolute risk of the drug similarly regardless of whether it was promoted by a medical or business representative (2.65 vs 2.78, P = .212). Likewise, participant ratings for comparative drug risk were similar regardless of representative credentials (3.86 vs 3.85, P = .920).
Credentials and Intention to Prescribe Drug
We hypothesized that HCPs exposed to a medical representative would have stronger intentions to prescribe the promoted drug than those exposed to a business representative. However, participants had similar prescribing intentions regardless of representative credentials (4.31 vs 4.33, P = .864).
Research Question 2: Data Limitations Disclosure
Disclosure and Perceived Drug Efficacy
As hypothesized, participants who viewed the disclosure rated the drug’s absolute efficacy lower than those who did not view the disclosure (4.56 vs 4.80, P < .004, η2 = 0.02), though the effect size was small (Figure 4). We likewise found that participants who viewed the disclosure were less likely to agree that the drug was more effective than other prescription drugs indicated for nausea and vomiting (4.26 vs 4.63, P < .001, η2 = 0.04). As with absolute efficacy, the effect size was small.
Effect of Data Limitations Disclosure on Participant Perceptions of Promoted Drug.
Disclosure and Attitudes Toward the Drug
As hypothesized, we found that participants who viewed the disclosure had significantly less positive attitudes toward the promoted drug than participants who did not view the disclosure (4.73 vs 4.95, P < .010, η2 = 0.02), with a relatively small effect size.
Disclosure and Intention to Prescribe Drug
Contrary to our hypothesis, we found that participants who saw the disclosure rated the likelihood that they would prescribe the drug very similarly to participants who did not see the disclosure (4.33 vs 4.31, P = .864).
Discussion
This study examined the potential for industry representatives’ educational credentials and disclosures of clinical trial data limitations at medical conferences to influence HCP’s prescription drug perceptions and prescribing intentions. We found that industry representative credentials did not influence prescribing intentions and had limited influence on perceptions of the representative, with HCPs rating medical and business representatives similarly on key characteristics (eg, credibility, knowledge, likability, familiarity, articulateness). Although HCPs rated representatives’ product expertise similarly regardless of credentials, they did rate medical representatives as having higher medical expertise, as expected. Finally, HCPs rated drug efficacy and risk similarly and had similar prescribing intentions regardless of the representative’s credentials.
Why did not industry representative credentials have a more prominent effect? As discussed in the introduction, we measured the perceived general professional and product expertise of the representative, as is typical in endorsement research.18 We operationalized this by profession because a medical doctor may have great familiarity with the actual use of the product among his or her patients, whereas a business focused representative may have familiarity with product specifications, mechanism of action, and indications, but less, if any, experience with real patients. One possible explanation is that HCPs did perceive the industry representatives as having different levels of credibility (demonstrated by the significant difference in perceived medical expertise), but this difference resulted in individual assessments that cancelled each other out. For example, some participants may have found medical credentials more persuasive, while others may have responded negatively to medical credentials, wondering why a medical doctor was promoting prescription drugs at conferences.
On the other hand, perhaps professional credentials simply did not matter much to participants. Any effect of the manipulation may have been overcome by the information in the rest of the video, which was identical except for the disclosure manipulation. This is exactly what we would hope and expect of clinicians. This also is consistent with research demonstrating that source credibility tends to have more influence when individuals are less knowledgeable about a topic.24 Thus, as highly trained and knowledgeable professionals, HCPs might be less swayed by their perceptions of an industry representative’s credibility. Future research manipulating presentation quality and more nuanced aspects of source characteristics beyond credentials, such as time in current position or even a characteristic such as confidence, might enable researchers to tease apart the line at which participants stop focusing on the message and start using heuristics to assess information.
We found that, unlike credentials, a disclosure did influence perceived efficacy and attitudes toward the drug. HCPs who viewed a disclosure rated the absolute and comparative efficacy of the promoted drug lower and had less positive attitudes toward the product. However, a disclosure did not alter intentions to prescribe the medication, with HCPs who saw a disclosure reporting similar prescribing intentions as HCPs who did not see one. This finding is consistent with Boudewyns et al. (2021),20 who also found differences in perceptions but not intentions. One possibility is that HCPs are not metacognitively aware of how disclosures affect their behavior or they are reluctant to admit the effects if they are aware. However, it is also possible that HCPs would still consider the drug to be an option for some of their patients despite its lower perceived efficacy. Again, this is what we would hope and expect from an HCP—that they recognize the risks and benefits of a drug but are still willing to prescribe it if they believe the drug can help a particular patient.
This study had several strengths. First, we recruited a large sample that was representative of numerous medical specialties. Second, we used a realistic video of an exhibit hall interaction featuring a realistic drug relevant across multiple specialties. Third, we sampled participants from 12 different medical conferences that occurred throughout 2022, reducing the likelihood that results were influenced by external events or conference-specific experiences. Fourth, we intentionally included both in-person and virtual conference attendees so that our findings would be relevant regardless of how conferences and exhibit halls adapted to COVID-19. Most importantly, we captured data in real time, within 7 days of participants’ conference attendance. Thus, participants were still in the “conference mindset” when viewing the stimuli, which likely increased its relevance and participants’ engagement with it.
The study had some limitations as well. Despite the realism of the stimuli, it was a hypothetical scenario and may not represent all HCP-industry interactions at conferences. Moreover, although we had generalizable samples of HCPs, we only used one type of disclosure. It is unclear whether and how our research would generalize to other types of disclosures. Finally, we assessed recall, perceptions, and intentions but not actual behaviors. Tracking actual prescribing behaviors would be a logical next step for future research but would require stimuli that references a real-world medication that HCPs could actually prescribe.
In conclusion, this study contributes to existing research about industry representatives’ communication of prescription drug information to HCPs by examining the influence of representatives’ educational credentials and disclosures of data limitations in the medical conference exhibit hall setting. The study’s findings suggest that HCP perceptions and intentions are not swayed by representatives’ educational credentials but that disclosures can temper HCP perceptions of promoted drugs, which supports their use in multiple situations, including at medical conferences.
Acknowledgments
We would like to thank the following employees of RTI International for their assistance: Dilsey Davis (stimuli development), Shari Lambert (stimuli development), and Leila Kahwati (advice on study design and the development of the study questionnaire and stimuli; thorough review of the manuscript). We also thank Meeta Patel of FDA for help with stimuli development.
Notes
This article was externally peer reviewed.
Funding: This work was supported by the Office of Prescription Drug Promotion, Office of Medical Policy, Center for Drug Evaluation and Research, U.S. Food and Drug Administration (FDA). This article reflects the views of the authors and should not be construed to represent FDA’s views of policies.
Conflict of interest: The authors have no conflicts of interest to report.
- Received for publication October 8, 2024.
- Revision received February 24, 2025.
- Accepted for publication March 5, 2025.
