Gender differences in bullying among internal medicine residents

Abstract

Purpose

To describe gender differences in experienced types of bullying, and resulting personal consequences, among internal medicine (IM) residents.

Methods

Participants in this cross-sectional study included 21 212 IM trainees who completed a voluntary survey with their 2016 in-training exam that assessed bullying during residency training. The 2875 (13.6% of) trainees who reported experiencing bullying on a screening question were asked for additional details about types of bullying experienced and resulting personal consequences.

Results

Female and male trainees experienced bullying at similar rates (47% versus 53%, P = .08). Gender differences were seen in both the type of bullying experienced and the resulting personal consequences. Female trainees were more likely than their male counterparts to report bullying characterized as verbal (83% versus 77%, P < .001) and sexual (5% versus 2%, P < .001), whereas male trainees were more likely to experience physical (6% versus 4%, P = .03) and “other” bullying types (27% versus 22%, P < .001). Female trainees were more likely to report negative personal consequences than male trainees, and the most common resultant sequela reported was feeling burned out (63% versus 51%, P < .001).

Conclusion

Gender differences exist in both the types and consequences of bullying experienced among this national sample of IM residents. These results should be considered by programs and institutions that are hoping to optimize the culture of their workplace and enhance safety in the learning environment.

Introduction

Bullying in the workplace is commonplace in many organizational cultures and creates a hostile work environment [1, 2]. Bullying is defined as repeated harassment by an individual perceived to be in a position of greater power [2–4] and is known to negatively affect mental and physical health with social, financial, and professional implications [5, 6].

At all levels of the medical education continuum, bullying is common and associated with a wide range of negative consequences including depression and suicidal ideation [7, 8]. A recent study of all US internal medicine (IM) trainees uncovered that close to 14% of trainees reported perceiving bullying in their training; the most common type of bullying was verbal harassment, and the ensuing consequences ranged from feeling burned out to illicit drug use [7].

In the school-age literature, gender differences in bullying have been observed. These data show that boys are more likely than girls to experience physical bullying, whereas girls are more commonly subjected to sexual harassment and emotional aggression [9, 10]. Recently, the #metoo movement [11] has highlighted the enormity of this problem systemically across our society; the healthcare industry has not been spared from these problems and transgressions [12, 13]. Shedding light on the differences in experienced bullying by gender may allow for tailored interventions to prevent it and support those who are affected by bullying [14].

This paper will describe gender differences in types of bullying experienced, and the personal consequences of this bullying, among a cohort of IM residents. Our a priori hypotheses were that female trainees experience sexual harassment more frequently than male trainees and that they would report more personal consequences of bullying than their male counterparts.

Materials and methods

Data collection

The methods for this study have been described in a previous publication [7]. Data were collected from 21 212 IM trainees (response rate: 88%) who completed a voluntary supplementary survey following the 2016 Internal Medicine In-Training Exam (IM-ITE), a test administered by the American College of Physicians [15]. This supplementary survey assessed experiences of bullying during residency training. Bullying was defined as “harassment that occurs repeatedly (>once) by an individual in a position of greater power” [2, 4, 16] and residents were asked, “During your time at your residency program, were you ever bullied?” The 2875 (13.6%) trainees who reported bullying were asked for additional details including the type of bullying experienced (verbal, sexual, physical, or “other” type of harassment). Further inquiry explored possible sequelae of said bullying in the following specific way: “What were the consequences of the bullying you experienced?” There were no questions investigating characteristics of the perpetrators. Respondents selected from a list of items, and were able to select multiple choices, if applicable. This study was deemed exempt by a Johns Hopkins Medicine Institutional Review Board.

Data analysis

First, we computed odds ratios (ORs) with confidence intervals (CIs) to assess the odds of each type of bullying for male compared to female residents. To examine gender differences in consequences across the different types of bullying experienced, logistic regression models tested main and interactive effects of gender and bullying categories for each of the sequelae.

Because survey respondents were able to endorse multiple categories of bullying, we created a categorical variable to classify respondents into substantively meaningful categories. The “verbal-only bullying” category was designated as the the reference category (n = 1950). Respondents who experienced sexual harassment alone or in combination with any other bullying type were assigned to the “sexual” category (n = 103); those who reported physical harassment with verbal or “other” type of bullying were assigned to the “physical” category (n = 128); and those who experienced “other” alone or in combination with verbal harassment were assigned to “other” (n = 694).

Interaction terms for the product of gender by bullying category provided tests of whether gender differences in personal consequences of bullying significantly varied depending on the type of bullying experienced. We used chi-square tables and computed ORs to further probe significant interaction terms. We performed sensitivity analyses to ensure that our findings were not dependent on methods of categorization. All descriptive and inferential statistics were computed using SPSS version 25 (IBM Corp).

Results

Table 1 describes the demographic characteristics of the sample. As described in the prior study [7], higher proportions of non-native English speakers and international medical graduates reported having experienced bullying. Table 2 shows gender differences in odds of reporting each bullying type among our subsample of 2875 residents that experienced having been bullied during their residency. We present the raw frequencies for each type of bullying included in the survey, where respondents were able to endorse multiple types.

Distributions across bullying categories

Most respondents (n = 2506, 87%) endorsed having experienced only one type of bullying during residency; 13% (n = 369) endorsed two or more types. The majority of respondents reported that the bullying they experienced was verbal harassment (n = 2306, 80%), including 68% of respondents (n = 1950) that experienced verbal harassment only. Endorsement of physical (n = 151, 5.3%) and sexual (n = 103, 3.6%) harassment was comparatively less prevalent. The majority of trainees who experienced sexual or physical harassment also endorsed other categories, but very few respondents endorsed that they experienced both sexual and physical harassment (n = 23, <1% of the sample).

Consequences associated with the bullying

Almost one-quarter of the sample (n = 677, 24%) indicated that they did not experience any of the personal consequences named in the survey, instead endorsing “none of these.” The majority of respondents (n = 2198, 76%) affirmed at least one personal consequence, and 42% (n = 1210) endorsed two or more personal consequences.

Gender differences in personal consequences

Table 3 presents ORs representing the odds for women compared to men of reporting each of the consequences. Women trainees were at greater odds of reporting depression, burnout, and changes in weight (all P < .05). Women were also significantly more likely to experience multiple consequences. Male residents were more likely to report consequential substance use or none of the specified consequences after being bullied than were their female counterparts (both P < .05).

Differences in personal consequences across bullying categories

Compared to those who experienced verbal harassment only, those who experienced sexual harassment endorsed four out of the eight personal consequence outcomes at significantly higher rates. As detailed in Table 4, experiencing sexual harassment carried a 2- to 6-fold increase in odds of experiencing certain negative consequences: depression, substance use, weight change, and leaving the program. Women in the sexual harassment bullying category were less likely to experience none of the specified consequences (OR = 0.2).

Of those bullied with physical harassment, the negative consequences were not significantly different from the verbal harassment group. In the physical harassment group, subsequent burnout was less frequently endorsed and improvement in performance was more common (both P < .05), see Table 4.

Gender by bullying category interaction

Tests of interaction terms retained the null hypothesis for all seven of the named personal consequence outcomes. These tests indicate that compared to the reference group, verbal-only harassment, gender differences were statistically similar across bullying categories for burnout, depression, weight change, substance use, decline in performance, improvement in performance, and leaving the program. The patterns of personal consequences experienced by women and men were similar within bullying type categories (see online supplementary material, Supplemental Table A for a summary of frequencies of all personal consequences endorsed by category).

The exception to this pattern relates to the endorsement of none of the specified consequences, resulting from the bullying. A significant gender by bullying type interaction term (Wald χ² = 7.8, df = 1, P = .005) suggested that gender differences significantly varied between those who experienced physical harassment compared to those who experienced verbal harassment only. Men in the verbal harassment only category were somewhat more likely than women to endorse none of the specified consequences of bullying (OR for men compared to women = 1.4; 95% CI: 1.1, 1.7; P = .0264), whereas women in the physical harassment category were significantly more likely than men to endorse none of the specified consequences (OR = 2.7; 95% CI: 1.2, 6.6; P = .026).

Discussion

This cross-sectional study of a national sample of IM trainees illustrates that even though overall female and male trainees endorsed experiencing bullying at similar rates, there are gender differences with regard to type of bullying experienced and resulting personal consequences. Female trainees were more likely than their male counterparts to experience bullying in the form of verbal and sexual harassment, and male trainees more likely to experience physical and “other” bullying types. Additionally, female trainees were more likely to report negative personal consequences (depression, burnout, weight change) than male trainees, who were more likely to report “none of these” or even a secondary improvement in performance. These noted differences may have implications for how we inquire about mistreatment among our trainees and the support offered when bullying is discovered.

Similar to previous studies, our results found that female trainees are more likely to experience sexual harassment and experience burnout [17–19]. Somewhat different from a recent cross-sectional national survey of general surgery residents in which female trainees were equally likely as male counterparts to experience burnout in models that were adjusted for mistreatment, our findings note increased rates of burnout reported by female trainees as a consequence of bullying [8]. Because studies cannot precisely quantify the severity of the bullying or mistreatment, it is impossible to know whether male or female trainees would suffer differential levels of burnout if they were subjected to comparable bullying.

Navigating medical training as a female resident can be particularly challenging. It is highly probable that sexual harassment, and bullying in general, remains underreported in medical education given concerns related to questioning the veracity of the victims’ experiences and possible retaliation. Increased rates of burnout may also result from an increased vulnerability among female trainees in the learning environment due to lack of organizational support focused on navigating work–life issues, especially for female trainees with young children [20, 21]. In, e.g. surgical training, it is postulated that female trainees are more susceptible to burnout because of fewer gender-concordant mentors and leaders without whom it may be difficult to challenge existing stereotypes and expectations [18]. Women may also be less likely to report bullying because of concerns for reinforcing gender stereotypes [22, 23] or causing a perceived gender role incongruity [24]. It may also be that some trainees view residency as time-limited and, due to fear of retaliation, tolerate bullying until it comes to an intolerable point.

It should also be noted that a large proportion (25%) of bullying victims reported that they experienced a form of bullying that was not adequately described as verbal, sexual, or physical. This “other” category of bullying was more likely to be endorsed by male trainees and showed a different pattern of personal consequences than the other types. Notably, the proportion of trainees that endorsed this “other” type of bullying was larger than the proportion that endorsed sexual and physical types of bullying combined. Apparently, we may have uncovered a prominent type of distinct bullying that warrants further exploration. This may be a specific type of bullying in training that remains uncharacterized and may include aspects of digital bullying [25], psychological abuse, discrimination, and other forms of aggressions.

For there to be a culture shift leading to the eradication of bullying from medical education, many strategies including the publication of empiric research are necessary. Additional studies to further characterize perpetrators and those comparing bullying types across specialties may be of benefit in further tailoring specific interventions. Bystander training is one promising approach being implemented at many institutions that empowers intervention at the grassroots level when mistreatment occurs [26, 27]. Some believe that effectively engaging bystanders can be even more effective than establishing a mechanism that encourages the victims to come forward in creating educational or work environments that are authentically anti-bullying [21]. For this to happen in medical education, there must be an ethical imperative that trainees and physicians be “upstanders” who actively stand up against any witnessed harassment [28]. Additional studies exploring the effects of these types of interventions are needed. There is a lack of awareness among program directors, as evidenced by the fact that less than one-third of IM program directors attested to the presence of any bullying of their residents [29] despite the noted pervasiveness of bullying as described by their residents. Unambiguous complaint and resolution processes including outside mediation should be considered for implementation by accreditation bodies like the Accreditation Council for Graduate Medical Education (ACGME) and may lead to increased awareness and a culture change in training programs [30]. It is imperative that the goal be “zero-tolerance” policies for bullying. Informed by the results of this study, such policies may take into account the differential trainee consequences from bullying by gender and the consideration about whether distinct institutional resources may be needed for supporting male versus female trainees. Once it is the expectation that witnesses are to come forward when they see bullying, with protection against retaliation and respect for anonymity if requested, there may begin to be a culture change with the elimination of unacceptable behaviors.

Several limitations of this study are worth considering. First, the data are cross-sectional and thus the consequences of experienced bullying were not followed over time. Second, bullying types were not specifically defined as there are no standardized definitions of what constitutes each type of bullying. However, during pilot testing, we saw that responders understood the constructs of the different bullying types (e.g. verbal, physical, sexual) in the same ways that they were considered by our team. Third, binary sex descriptors (male/female) are used in this paper to describe gender, which reflects how the information was collected and reported by the IM-ITE at the time of the data collection. Fourth, the data were collected in 2016, though they remain relevant as bullying continues to remain an active problem in medical education and training. Finally, there may have been additional consequences not included in the list we provided that participants may have experienced.

Conclusion

This national survey of IM residents who took the 2016 IM-ITE establishes that there are gender differences in types of bullying experienced and resulting consequences. There must be legitimate and concerted attempts to eliminate bullying from medical education, and such efforts will benefit all learners.

Acknowledgements

The authors would like to acknowledge the American College of Physicians IM-ITE survey team for their assistance with this project.

Funding

The authors alone are responsible for the content and writing of this article.

Conflict of interest statement: None declared.

Contributors’ statement

R.R. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors contributed to the concept and design of the study. All authors were involved in the acquisition, analysis, or interpetation of the data. R.R. and S.M.W. conducted the statistical analysis of the study. M.S.A. and S.M.W. drafted the manuscript. All authors were involved in the critical revision of the manuscript for important intellectual content.

References