INSIDE-OUT: Introduction of Speakers at IDWeek Events—Observing for Unconscious Bias Over Time

Abstract

Background

Specialty societies, including the Infectious Diseases Society of America, strive to address gender and racial inequities in professional advancement. Microaggressions remain a persistent and pervasive barrier to these goals. Nonprofessional speaker introductions are a manifestation of race- and gender-based microaggressions, which have not been previously assessed at IDWeek. We assessed disparities in speaker introductions at IDWeek over a 7-year period that included formal gender equity initiatives introduced in 2016.

Methods

We conducted a retrospective observational study of video-recorded IDWeek speaker introductions from 2013 to 2019. Trained coders reviewed presentation video archives to assess a primary outcome of nonprofessional introductions: when a speaker's professional title was not used as the first introduction. We used descriptive statistics, Fisher exact tests, Cochrane-Armitage trend tests, and multivariable logistic regression to characterize relationships between speaker introductions and presentation year, speaker demographics, and speaker-moderator demographic concordance.

Results

Of 1940 videos reviewed, 48.9% of IDWeek speakers received nonprofessional introductions during and before 2016 vs 41.5% of speakers after 2016 (P = .0013). There was an increasing linear trend in the frequency of professional introductions by speaker age group from 47.1% for age <40 years to 65.3% for age >60 years (P < .0001). White moderators more frequently used nonprofessional introductions than moderators from backgrounds underrepresented in medicine (47.7% vs 29.1%, P = .0014). Women-men speaker-moderator pairs had more nonprofessional introductions (54.6%, P < .001).

Conclusions

In the largest assessment of microaggressions in speaker introductions at a national medical specialty conference, we highlighted some progress over time and ample opportunity to further standardize equitable speaker introductions, especially for women and junior speakers

equity, IDA&E, IDWeek, professional introduction, unconscious bias

Gender and racialized disparities in career advancement are ubiquitous in academic medicine, with evidence of discordant pay, underrepresentation in leadership positions, and inequities in invited presentations at national meetings [1–5]. The Infectious Diseases Society of America (IDSA) has committed to addressing gender and racial equity within the profession [6]. Notable achievements include equitable gender representation among IDWeek speakers and program committee members [7], increased proportions of speakers from racial and ethnic backgrounds that are underrepresented in medicine (URiM) [7], and equitable representation of membership across multiple volunteer leadership positions in the organization [8].

Despite meaningful progress toward equitable representation in professional spaces, microaggressions in the workplace compound existing inequities through reinforcing unconscious biases, causing feelings of “otherness.” Presentations at academic medical conference such as IDWeek enhance reputation and career advancement for clinicians, scientists, and public health professionals. In this professional setting, introductions highlight the speaker's authority or expertise in the subject matter; therefore, the expectation should be that speakers are introduced by their professional titles [9]. Nonprofessional speaker introductions at these conferences are one manifestation of gender-based microaggressions, reinforcing unconscious bias when professional titles are not used consistently. Prior literature in institutional and large national conference settings has shown that women are less likely to be introduced by their professional titles than men [9, 10].

Recommendations of the 2016 IDSA Gender Disparity Task Force focused on equitable representation among speakers but did not explicitly address unconscious bias in speaker introductions or prescribe how speakers should be introduced. This study aimed to assess disparities in how speakers are introduced at IDWeek meetings, with particular interest in assessing whether the 2016 IDSA Gender Disparity Task Force was associated with improvements in professional introductions.

METHODS

This retrospective observational analysis of video-recorded speaker introductions from 2013–2019 IDWeek sessions used archives obtained directly from IDSA staff. All speaker and moderator demographics in this cohort were self-identified by participants, and the distribution has been described [7]. We used the Association of American Medical Colleges’ definition of URiM as “racial and ethnic populations that are underrepresented in the medical profession relative to their numbers in the general population” [11]; for the data available in our cohort, this included people self-identifying as Black/African American, American Indian/Alaska Native, Native Hawaiian/Pacific Islander, or Hispanic/Latino. Moderators and speakers were included who had full introductions captured on recordings of plenaries, named lectures, oral abstracts, symposia, meet-the-professor sessions, and panels. If a speaker spoke in more than 1 session, each slot was considered an independent event; however, if a speaker spoke more than once within an individual session, that person was counted only once for that session based on the first introduction.

Ten team members (INSIDE-OUT Study Group) reviewed and coded the videos after receiving standardized training on systematic video review, coding strategies, and code-reporting structure. Two coders reviewed each presentation to document first-time introductions and subsequent use of professional honorifics. R. K. and J. R. M. reviewed and assigned final coding determinations for any discordant team codes. Introductions were coded as “professional” if moderators used titles of advanced degrees or other honorifics (eg, Dr [full name], Dr [last name], Professor [full name]) and “nonprofessional” if moderators referred to speakers without professional honorifics (eg, [First name only] or [First and last name]). The κ statistic for agreement between coders was 0.90, indicating very good interrater reliability.

Data were collected via a web-based form stored on a secure database and summarized with descriptive statistics. Fisher exact test was used to assess the association of professional speaker introduction with speaker demographics (race, ethnicity, gender, and age) and speaker-moderator demographic concordance. The Cochrane-Armitage test of trend was used to assess significant linear trends in professional introduction by year and by age group. Multivariable logistic regression that included speaker gender, age group, and presentation during or before the 2016 IDWeek meeting was used to determine which speaker demographics were independently associated with receiving a professional introduction. All analyses were done with SAS version 9.4 (SAS Institute). P < .05 was considered statistically significant. This work was evaluated and classified as exempt nonhuman subjects research by the University of Nebraska Medical Center Institutional Review Board.

RESULTS

Across 7 IDWeek conferences from 2013 to 2019, we analyzed 1940 speaker slots. Of all speakers, 44.8% were introduced nonprofessionally during their first introduction, and 42% were never introduced professionally during their entire session. Of the 624 speakers who were introduced twice during the session and did not initially receive a professional introduction, 18.7% were introduced professionally on the second instance. During and before IDWeek2016, 48.9% of speakers received nonprofessional introductions as compared with 41.5% after 2016 (P = .0013). The rate of receiving a nonprofessional introduction had a decreasing linear trend by year from 49.5% in 2014% to 35.5% in 2019 (P < .0001; Figure 1). Speaker demographics stratified by receipt of professional or nonprofessional introductions are summarized in Table 1. Men were more likely to receive professional introductions than women, though this difference was not statistically significant (57.2% vs 53.1%, P = .0773). There was an increasing linear trend in professional introductions by age group from 47.1% of speakers aged <40 years to 65.3% of speakers aged >60 years (P < .0001). There were no statistically significant associations in professional introduction by speaker race or ethnicity, so these variables were not included in the multivariable model. In multivariable analysis, presenting during or before the 2016 IDWeek meeting (vs after 2016; odds ratio, 1.34 [95% CI, 1.11–1.62]) and age group (<40 vs >60 years; odds ratio, 2.03 [95% CI, 1.51–2.74]) were independently associated with nonprofessional first introduction. Speaker gender was not independently associated with nonprofessional introduction.

Figure 1.

Line plot of first speaker introduction by year.

Table 1.

Speaker Demographics by Professional and Nonprofessional Forms of First Introduction

	Introductions, No. (Row %; 95% CI)
	Professional	Nonprofessional	Total, No. (Column %)	P Value
Conference year before vs after 2016 Gender Disparity Task Force				.0013a
≤2016	449 (51.1; 47.8–54.4)	429 (48.9; 45.6–52.2)	878 (45.3)
>2016	621 (58.5; 55.5–61.4)	441 (41.5; 38.6–44.5)	1062 (54.7)
Speaker gender				.0773a
Men	556 (57.2; 54.1–60.3)	416 (42.8; 39.7–45.9)	972 (51.9)
Women	478 (53.11; 49.8–56.4)	422 (46.9; 43.6–50.2)	900 (48.1)
Missing	36	32	68
Speaker age, y				<.0001b
<40	271 (47.1; 43.0–51.2)	304 (52.9; 48.8–57.0)	575 (31.6)
40–49	296 (54.2; 50.0–58.4)	250 (45.8; 41.6–50.0)	546 (30.0)
50–59	248 (60.0; 55.3–64.8)	165 (40; 35.2–44.7)	413 (22.7)
≥60	188 (65.3; 59.8–70.8)	100 (34.7; 29.2–40.2)	288 (15.8)
Missing	67	51	118
Speaker race or ethnicity				.5515a
AI/AN or NH/PI	5 (83.3; 53.5–100.0)	1 (16.7; 0–46.5)	6 (0.3)
Asian	160 (59.0; 53.2–64.9)	111 (41.0; 35.1–46.8)	271 (15.0)
Black/African American	30 (54.6; 41.4–67.7)	25 (45.5; 32.3–58.6)	55 (3.1)
Hispanic/Latino	44 (55.0; 44.1–65.9)	36 (45.0; 34.1–55.9)	80 (4.4)
White	753 (54.6; 51.9–57.2)	627 (45.4; 42.8–48.1)	1380 (76.6)
Other	6 (66.7; 35.8–97.5)	3 (33.3; 2.5–64.2)	9 (0.5)
URiMc	79 (56.0; 47.8–64.2)	62 (44.0; 35.8–52.2)	141 (7.9)
Missing	72	67	139
Moderator race or ethnicity				.0160a
AI/AN or NH/PI	2 (66.7; 13.2–100)	1 (33.3; 0–86.8)	3 (0.3)
Asian	83 (54.6; 46.7–62.3)	69 (45.4; 37.5–53.3)	152 (13.5)
Black/African American	31 (68.9; 55.3–82.4)	14 (31.1; 17.6–44.7)	45 (4.0)
Hispanic/Latino	40 (72.7; 60.9–84.5)	15 (27.3; 15.5–39.1)	55 (4.9)
White	452 (52.3; 48.9–55.7)	412 (47.7; 44.3–51.0)	864 (76.5)
Other	6 (54.5; 25.1–84.0)	5 (45.5; 15.9–74.9)	11 (1.0)
URiMc	73 (70.9; 62.1–79.7)	30 (29.1; 20.3–37.9)	103 (9.2)
Missing	456	354	810
Speaker-moderator gender pairs				<.0001a
Speaker: woman–moderator: woman	141 (60; 53.7–66.3)	94 (40; 33.7–46.3)	235 (21.2)
Speaker: man–moderator: man	153 (49.7; 44.1–55.3)	155 (50.3; 44.7–55.9)	308 (27.8)
Speaker: woman–moderator: man	127 (45.4; 39.5–51.2)	153 (54.6; 48.8–60.5)	280 (25.3)
Speaker: man–moderator: woman	184 (64.6; 59.0–70.1)	101 (35.4; 29.9–41.0)	285 (25.7)
Missing	465	367	832

Abbreviations: AI/AN, American Indian/Alaska Native; NH/PI, Native Hawaiian/Pacific Islander; URiM, underrepresented in medicine.

^aFisher exact P value.

^bCochran-Armitage trend test.

^cIncludes Black/African American, AI/AN, NH/PI, and Hispanic/Latino, based on the Association of American Medical Colleges.

Of 1940 speaker-moderator pairs, 832 moderators (42.9%) were missing demographic data (Table 1). Among moderators with demographics, men were more likely to give nonprofessional introductions than women (53.4% vs 37.5%, P < .001). Women-men speaker-moderator pairs had the most nonprofessional introductions (54.6%). There was a statistically significant association between moderator race or ethnicity and professional address. White moderators used nonprofessional address to introduce speakers 47.7%, as compared with 45.4% for Asian moderators and 29.1% for moderators from URiM backgrounds (P = .0014).

DISCUSSION

We report the largest study evaluating the rates of nonprofessional speaker introductions at a medical conference by gender, age, race, and ethnicity, where almost half of the speakers experienced nonprofessional first introductions and significant disparities were noted with younger speakers and with speaker-moderator gender and racial discordance.

Our study has several key findings. First, while over half of speakers were addressed professionally at first introduction, there is still room for improvement as 42% of speakers were never introduced with professional titles at all. This is lower than reported for a variety of other medical/surgical conferences [10, 12, 13]. Second, professional speaker introductions increased over time and especially after 2016, with no statistically significant disparities by speaker gender, race, or ethnicity in professional introductions. The reasons for this may be multifactorial. For example, the IDSA Gender Disparity Task Force was developed in 2016 and focused on equitable gender representation among IDWeek speakers, which could have played a role despite no specific speaker introduction policies. Over time, as data have accumulated highlighting the prevalence of gender disparities and microaggressions in these settings, individual session moderators may have grown increasingly aware of these issues and made intentional efforts to avoid nonprofessional address in introductions. However, there is still room for improvement given higher rates of nonprofessional introductions by moderators who were men or White.

Third, younger speakers more frequently experienced nonprofessional introduction, particularly with discordant moderator-speaker pairs. This suggests that the traditional hierarchy of academic medicine may perpetuate a bias of junior speakers as being less likely to be experts. Irrespective of the moderator's intentions, this may negatively affect engagement or performance of junior IDSA members if they believe that they are not respected or perceived as experts in their field. While we did not analyze our data by academic rank (these data were not consistently available), Stewart et al conducted a similar analysis at the Society of Surgical Oncology and found that trainee speakers were almost 3 times more likely receive a nonprofessional introduction, especially when introduced by more senior faculty [14].

While this study was not designed to understand the intent of moderators who did not use professional introductions, there are multiple published reports of the impact of unconscious bias in health care and academic settings on women and minoritized individuals [15]. This study demonstrates some progress and some opportunities. We previously reported the improvement of gender equity in invited speakers after the IDSA Gender Disparity Task Force, and with this new study in mind, we posit that the prioritization of gender equity may have positively influenced speaker introduction trends. However, consistent with studies in other disciplines [9, 12, 16], women introduced by men in our study received nonprofessional introductions in over half of presentations, whereas women moderators provided professional introductions in 62.5% of all presentations.

Standardization of speaker introductions may be a key opportunity to enact our previous recommendations to address equity with specific and intentional initiatives [7]. As the proportions of IDWeek invited speakers who are women, who are from URiM backgrounds, and who are <50 years old are continuing to increase over time [7], it is important to ensure that these speakers feel a sense of belonging. An action plan to address this should include the development of a formal moderator policy that consists of best practices for moderating a session and professionally introducing speakers. Published guidelines for moderators acknowledge the impactful role that the moderator has in creating a memorable conference session for both the audience and the speakers [17, 18]. Recommendations have included communication with speakers ahead of time to coordinate session timing and disclosure requirements, assessment of need for accommodations, and familiarization with the speaker's topic [18]. We recommend that when the moderator reaches out to the speaker to plan the introduction, objectives of this discussion should be to clarify pronouns and phonetic pronunciation of names, clarify pertinent aspects of speaker biography (eg, professional title), and explicitly describe how the moderator intends to introduce the speaker (Figure 2). An IDWeek moderator orientation toolkit could incorporate these recommendations in addition to highlighting data from our IDWeek analysis and other published studies on microaggressions so that moderators can appreciate the full extent and variation of unconscious biases (ie, factors beyond gender, such as race, ethnicity, age, professional rank, specialty, and geographic location). The toolkit can also reinforce strategies that individual moderators can take to mitigate potential bias, such as familiarizing themselves with the speaker’s expertise and session topic, engaging in deliberative reflection, actively countering stereotypes, and practicing cultural humility [15]. Last, in cases where there is moderator-speaker familiarity or the format of the session lends itself to be more informal (eg, panels or fireside chats), moderators should be instructed to consistently use professional address at first introduction of the speaker followed by a professionalism declaration (eg, “Dr XYZ and I have discussed and mutually agreed to hereafter refer to each other by our first names”).

Figure 2.

Best practices for introducing speakers and avoiding microaggressions in a moderated conference session.

Limitations

This study represents an advancement of prior literature on speaker introductions because of its large size, long duration, and use of speaker-defined—rather than investigator- or software-assigned—demographic data. However, there are several limitations. First, we identified professional honorifics by traditional titles (eg, Dr, Professor). Yet, for some speakers without doctorate degrees (eg, advanced practice providers, community members, infection preventionists), it may be appropriate to address them with alternative titles of respect such as Mr, Mrs, Ms, and Mx. Our previous analysis reported that 4.7% to 4.8% of speakers held nondoctoral degrees [7] where this may have been applicable, though we did not account for this in our analysis of introductions. The study relied on video review; nonrecorded sessions were not included. As with all speaker introduction studies, it is challenging to account for preexisting familiarity of the moderator and speaker, which may influence their use of professional titles; thus, we focused most of our analyses on first introduction. There were high levels of missingness in moderator demographic data, which may affect the strength of our analysis of introductions by moderator groups. As such, improved data collection prior to future repeated analysis may be helpful. We were unable to conduct intersectional analyses (eg, combinations of gender and race), but these analyses should be considered in future work. We acknowledge that racial groupings are nonhomogeneous and time-varying social constructs; unfortunately, our data set did not allow for disaggregation of Asian racial groups or Indigenous tribes. Our sample size limited the ability to analyze American Indian/Alaska Native and Native Hawaiian/Pacific Islander groups separately, so these were considered 1 category. Gender was collected in demographic data as binary male/female; other gender identities were unable to be identified. However, in 2021, IDSA began featuring more inclusive options for self-reported gender identity, such as transgender, nonbinary, and gender nonconforming, and these subgroups merit further analysis.

CONCLUSION

In the largest study of speaker introductions at an academic medical specialty conference, 44.8% of speakers received a nonprofessional first introduction, though rates of professional introduction improved over time and after formation of the IDSA's Gender Disparity Task Force. Younger speakers received more nonprofessional introductions, and moderators who were men gave more nonprofessional introductions, particularly when paired with women speakers. Conference presentations are an important component of developing a national academic portfolio. Speaker introductions contribute to how the speaker is viewed as an authority in the subject matter, and there is a possibility for reinforcement of unconscious biases if professional titles are not used consistently. We advocate that IDSA—and other specialty societies—consider a formal approach to orienting moderators, which includes discussion of nonprofessional introductions as a form of microaggression, in alignment with IDSA's commitment to inclusion, diversity, equity, and access across the organization.

Notes

Acknowledgments. We thank the IDSA staff for providing us with the data from IDWeek. We are grateful to Toni Goeser for her development of Figure 2. Finally, we acknowledge the following members of the INSIDE-OUT Study Group for their contributions, including their affiliations at the time of contribution: Nipunie Rajapakse, Mayo Clinic; Megha Patel, Amy M. Miller, Erin N. Oeltjen, Sophie Semaj, Kyle T. Sramek, Christopher J. Lindeman, Eric C. Daubach, Oladapo O. Akinmoladun, Paul J. Wurtz, and Yuhong Yang, University of Nebraska Medical Center.

Disclaimer. The content is solely the authors’ responsibility and does not necessarily represent the official views of the National Institutes of Health.

Ethical approval. This work was evaluated and classified as exempt nonhuman subjects research by the University of Nebraska Medical Center Institutional Review Board.

Patient consent. This work does not include factors necessitating patient consent.

Financial support. None.

Author notes

Potential conflicts of interest. R. K. reports grants from Boston Children's Hospital, Brigham & Women's Hospital, and the Joel and Barbara Alpert Endowment for Children of the City; consultant fees from the New York City Department of Health and Mental Hygiene; and travel/lodging reimbursement for service on The Lancet's Commission on Antiracism and Solidarity outside the submitted work. E. H. R. is a stock shareholder of various companies, including Pfizer and Moderna, unrelated to this project. N. F. reports consulting/advisory work for Merck, Janssen, AstraZeneca, Novocure, Regeneron Pfizer, and Johnson & Johnson and grant funding from Genentech and Daiichi-Sankyo unrelated to this project. J. C. A.-M. receives support from the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health under award K12HD113189. J. C. A.-M. has also participated in advisory boards for Shionogi Entasis Therapeutics, CSL Sequiris, Abbvie, GSK, and Novavax and has received honoraria. She is an appointed member of the CVS Health National Health Equity Advisory Board. All other authors report no potential conflicts.