Organizational Factors Contributing to Incivility at an Academic Medical Center and Systems-Based Solutions: A Qualitative Study.

PURPOSE: A rise in incivility has been documented in medicine, with implications for patient care, organizational effectiveness, and costs. This study explored organizational factors that may contribute to incivility at one academic medical center and potential systems-level solutions to combat it.
METHOD: The authors completed semistructured individual interviews with full-time faculty members of the Department of Medicine (DOM) at the University of Toronto Faculty of Medicine, Toronto, Ontario, Canada, with clinical appointments at six affiliated hospitals, between June and September 2016. They asked about participants' experiences with incivility, potential contributing factors, and possible solutions. Two analysts independently coded a portion of the transcripts until a framework was developed with excellent agreement within the research team, as signified by the Kappa coefficient. A single coder completed analysis of the remaining transcripts.
RESULTS: Forty-nine interviews with physicians from all university ranks and academic position descriptions were completed. All participants had collegial relationships with colleagues but had observed, heard of, or been personally affected by uncivil behavior. Incivility occurred furtively, face-to-face, or online. The participants identified several organizational factors that bred incivility including physician nonemployee status in hospitals, silos within the DOM, poor leadership, a culture of silence, and the existence of power cliques. They offered many systems-level solutions to combat incivility through prevention, improved reporting, and clearer consequences.
CONCLUSIONS: Existing strategies to combat incivility have focused on modifying individual behavior, but opportunities may exist to reduce incivility through a greater understanding of the role of health care organizations in shaping workplace culture.

Institutions across many sectors, including medicine,[1] are working to counter a rise in reported incivility. Professionalism—a core tenet of medical practice—is a broad concept rooted in responsibilities to the patient, other health care providers, and the profession.[2] Incivility, a subset of unprofessionalism,[3] is defined as low-intensity behaviors in either words or gestures that may confer indirect harm to colleagues[4] and cultivate a toxic work culture. Uncivil behavior includes, but is not limited to, bullying, verbal abuse or rudeness, excessive competitiveness, social exclusion, inaction (e.g., not responding to e-mails), and other forms of disrespect. The impact of incivility on organizational operating costs, employee turnover, performance, cohesion, and organizational commitment has been documented in many industries, as have the negative consequences of incivility for individuals.[5]

In medicine, a recent systematic review and meta-analysis showed that almost 60% of medical trainees reported experiencing harassment or discrimination over the course of their education,[6] a prevalence that is unacceptable. Staff physicians are likewise affected, with one study citing weekly exposure to rude, dismissive, or aggressive behaviors by over 30% of physicians.[7] Aside from the obvious moral dimension to civil behavior emerging from basic tenets like the Golden Rule (i.e., “treat others as you wish to be treated”),[8] there are practical reasons to promote civility in the workplace, especially in work environments with heightened complexity such as medicine.[4] There are stresses unique to medicine—for example, the existential nature of the work (i.e., looking after patients with acute or life-threatening conditions) and long work hours including overnight call shifts—that may amplify the potential for incivility.[9] In the health sector, an important consequence of incivility is the potential adverse impact on patient care.[10] Poorer patient outcomes may result from decreased consultation between terse colleagues, worsened communication, poor collaboration and teamwork, and physician burnout.[11] Indeed, burnout may itself contribute to a further “incivility spiral”[4] owing to the callous behaviors that may be attendant to the emotional exhaustion experienced by physicians.[12] These behaviors can affect the mental health and productivity of others, including interprofessional colleagues and trainees.[13] Academic medical centers present further catalysts of incivility including pressures around publications and academic promotions.

There is a dearth of literature exploring experiences with incivility between physicians, especially at academic medical centers. This is concerning given the decades of evidence that has correlated the “hidden curriculum” of medicine with the depersonalization of trainees and a subsequent loss of empathy toward patients and colleagues.[14] There is also very little evidence demonstrating the effectiveness of existing interventions to counter incivility,[15] which, within health care, have included strategies such as faculty feedback[16] and education and awareness campaigns.[17-22] These strategies are largely focused on modifying individual behavior rather than addressing workplace culture.

This study was conducted in an attempt to understand the organizational factors that may contribute to incivility and devise potentially testable systems-based strategies to counter it.

Method

Study design

We conducted a qualitative study using thematic analysis[23] and a positivist research paradigm.[24] We conducted semistructured, individual interviews with members of a single university department of medicine (DOM) that is affiliated with six hospitals to explore this issue and identify potential strategies to address it.

Sampling and recruitment

We used both a purposive sampling approach and a snowball sampling strategy to recruit DOM members from the six hospitals fully affiliated with the Faculty of Medicine, University of Toronto. This is the largest DOM in Canada, with almost 1,000 full-time faculty members who hold clinical appointments at their hospitals and academic appointments at the university. Given the universal health care system in Canada, a physician’s clinical appointment at a hospital facilitates work as a consultant, rather than as an employee, and most of a physician’s clinical pay is in the form of fee-for-service (pay for “acts” of patient care), remunerated by the provincial ministries of health, with overhead costs deducted by the hospital.[25] If physicians also hold administrative roles within the university, they may be supported by stipends from the university DOM. A small number of clinician–scientists (who are not funded through other research salary support mechanisms) may also be remunerated by the university for protected time for scholarly work.

Our goal was to recruit participants until saturation of themes was achieved and sampling continued until no new themes were identified.[26] We anticipated recruiting 18 to 36 male and 18 to 36 female participants who were current and full-time faculty members, positing that 6 to 12 interviews from each hospital may allow us to reach saturation of themes.[27] Statistical analysis of data obtained from a DOM survey that we conducted in 2015[28] suggested that experiences with incivility varied by hospital, division, and academic position description; therefore, we aimed to recruit individuals across these demographic variables. We included faculty members from all 15 specialties as defined by the Royal College of Physicians and Surgeons of Canada—namely, cardiology, clinical immunology and allergy, clinical pharmacology and toxicology, critical care, endocrinology, gastroenterology, general internal medicine, geriatrics, hematology, infectious diseases, medical oncology, nephrology, respirology, rheumatology, and occupational medicine. We also sampled across the six academic position descriptions which vary on the basis of the time spent on clinical versus scholarly work (Table 1), where scholarly work is composed of teaching, research, administrative activities, or quality improvement. We also wanted to explore the perceptions of incivility by university rank, and we sampled from each, including lecturer, assistant professor, associate professor, and full professor.

Table 1

Academic Position Descriptions of Faculty Members Within the Department of Medicine, University of Toronto Faculty of Medicine

We obtained a publicly available list of current, full-time faculty members and recruited via a personalized e-mail invitation. Interested participants were invited to contact a research coordinator. Prior to the interview, the study was described, and informed consent was obtained from each participant. We used a snowball sampling strategy by asking participants if they felt comfortable supporting recruitment and encouraging other DOM members to participate. To maintain confidentiality, all e-mail communication with prospective and actual participants was deleted from the recruitment team’s e-mail server upon study completion.

Data collection

Individual semistructured interviews lasting 45 to 60 minutes were completed by telephone between June and September 2016 by two authors (A.M.J., S.J.). We developed domains of inquiry based on our 2015 survey results and from consultation across the study team and DOM leadership. Questions included items about experiences with uncivil behavior, potential factors contributing to these behaviors, and strategies to prevent or manage incivility within the DOM. As data were analyzed, we iteratively modified the interview guide and added new questions to address evolving themes. The interview guide is provided in Supplemental Digital Appendix 1, available at http://links.lww.com/ACADMED/A567. Participant demographic information, including academic position description and gender, were collected.

After the first round, which included eight interviews, we submitted an ethics amendment to offer participants the choice between in-person or telephone interviews; this was in response to requests by participants, given the sensitive subject matter. Interviews were audio recorded, transcribed verbatim, anonymized, and deidentified. The primary data were composed of the transcripts along with secondary sources such as field notes and memos (drafted by the experienced interviewers) that were shared within the research team. We offered all participants the option to review their own transcripts prior to analysis for two reasons: Though transcripts were deidentified, we wanted to give participants the option of removing any potentially identifying stories or anecdotes; and we wanted to ensure that the data reflected the intended perceptions and experiences of participants.[29] Participants who were quoted were also given the option to review and comment on a draft copy of this manuscript. We sought permission from participants to use their direct quotes to describe the themes identified.

Data analysis

We used NVivo software, version 11 (QSR International Pty Ltd., Victoria, Australia) to facilitate the data analysis using a thematic analysis approach.[23] We established a coding framework using memoing, wherein interviewers (A.M.J., S.J.) collaboratively recorded their impressions of both repeated and new themes immediately after each interview. Two analysts (A.M.J., S.J.) and two other authors including the principal investigator (S.G., S.S.) reviewed and edited the coding framework for clarity and comprehensiveness. We then used the following modified coding consensus approach[30]: First, two members of our research team (A.M.J., S.J.) independently coded 10% of the transcripts using the initial coding framework and used NVivo software to determine interrater reliability, measured using the Kappa coefficient. Any discrepancies with a Kappa coefficient ≤ 0.6 were discussed and resolved by making changes to the coding framework as needed. We undertook a second round of coding, reviewed interrater reliability, and made further changes to the framework. This process continued iteratively until the number of discrepancies (i.e., Kappa coefficients ≤ 0.6) decreased dramatically. A single analyst (A.M.J.) then completed the analysis on the balance of the transcripts. We used written memos to provide a record of the analytic process. The memos captured the decisions and results of the analysis and helped to develop propositions.

Ethical approval

We obtained approval for the study from the Research Ethics Boards at St. Michael’s Hospital (REB no. 16-032C) and the University of Toronto (REB no. 32725).

Results

We conducted 49 interviews between June and September 2016. Two participants requested in-person interviews, and the remaining interviews were conducted by phone. All 15 specialty divisions were represented, with each division comprising 2% to 14% of participants. Five out of six of the fully affiliated hospital sites were represented. Remaining demographic variables are noted in Table 2. To ensure participant anonymity, we are not able to provide additional demographic information.

Table 2

Demographic Characteristics of 49 Participants, From a Qualitative Study of Incivility Within the Department of Medicine, University of Toronto Faculty of Medicine, 2016

Because of the complexity of the data set, six rounds of coding were required before arriving at a framework with near-universal agreement within the research team and a Kappa coefficient ≥ 0.6. Complexity was conferred by the sensitive nature of the subject matter; given the topic, several participants were not always forthcoming in their responses. The analysts needed to balance the need for fidelity to what was articulated against the need to draw conclusions by inference from the transcripts. Therefore, two analysts reviewed 40% of the transcripts over six iterations, and a single analyst reviewed the remaining 60% of transcripts.

Among all participants, 25 individuals reviewed their transcripts, and 5 participants edited their quotes. Among the 8 quoted participants, 1 requested that a quote be removed.

Overall, all participants described collegial relationships with their colleagues. However, all participants had observed, heard of, or had direct experiences with uncivil behavior. We obtained information regarding the types of incivility faced by DOM members, its effects, and participants’ perceptions of the potential formal consequences to perpetrators. We further identified themes focused on organizational factors influencing uncivil behavior and potential strategies for preventing and managing these behaviors at a systems level. Deidentified quotations from participants are provided to illustrate the themes.

Types of incivility, effects, and perceived consequences

Participants indicated that incivility within the DOM varied in its form and content. With regard to form, participants said that they had witnessed or experienced incivility furtively, face-to-face, and online. Furtive behaviors included eye rolling, gossip, the muttering of derogatory comments, and social exclusion. Face-to-face behaviors included public ridicule, personal attacks, blatant interruptions and overtalking, yelling, and cutting short conversations or phone calls, as well as more physically aggressive actions like throwing objects. The toxic online behaviors identified included nonresponse to e-mails, rude e-mails, passive–aggressive e-mails, and disparaging posts on social media platforms. With respect to content, participants described incivility based on disagreements with regard to consultations, treatment recommendations, negotiations around promotions or remuneration, or workload distribution. They highlighted incidents of discrimination based on subspecialty, academic position description, or academic rank.

The participants described several effects of incivility, including a negative impact on patient safety, on the mental health of affected individuals, and on the culture of the work environment. Participants felt that there were very few moderate-level disciplinary actions that could be taken to address incivility, and they indicated that they were only aware of extreme measures, such as revoking physician privileges at a hospital, to formally address incivility.

Organizational factors

The participants outlined organizational factors that they thought incited or perpetuated workplace incivility. We identified two broad themes: the work environment and human resources. In terms of the work environment, participants identified physician nonemployee status and silos within the DOM as explanatory factors. In terms of issues related to human resources, participants highlighted poor leadership, a culture of silence that leaves incivility unaddressed, and power cliques as potential catalysts.

Work environment.

Several participants felt that a challenge faced by academic medical centers in improving organizational effectiveness was the nonemployee status of physicians. Specifically, they work in hospitals but are considered self-employed. The perception is that their nonemployee status places them outside the purview of traditional hospital human resources policies and renders them immune to conventional protocols for disciplinary action. This is the perception despite the fact that these self-employed physicians are subject to the occupational health and safety policies of the hospital and university and they are accountable to the regulatory bodies responsible for credentialing. As a result, some participants perceived that only extreme behaviors by physicians are flagged. In the words of one participant:

The only lever you have is withdrawal of privileges, which is obviously a severe issue…. I mean, if the physicians were employees of the hospital it would be totally different.

In short, unprofessional clinicians operating in a professional “gray zone” were felt to be able to act with impunity.

The physical environment itself was felt to facilitate, and even exacerbate, workplace incivility. There are various silos within the DOM, specifically between the hospital and the university, and between divisions. Therefore, geographic and existential separations existed between colleagues, as one participant noted:

I think that that landscape is a lot more complicated given the number of subspecialties, physicians … it can be very difficult to navigate, more cold, more impersonal and more a sense that the first interaction is coming from a place of conflict.

Relative anonymity within the large university DOM was felt to lower the threshold for curt behaviors. Several participants correlated infrequent interactions between colleagues with the formation of inaccurate assumptions and prejudices.

Human resources.

Some participants felt that there were several organizational factors related to human resources that fostered a climate of incivility in the DOM. They felt that leaders that modeled or condoned divisiveness and competition may have been a driving force. In these cases, some participants suggested that a change in leadership often resulted in palpable changes to culture. In particular:

And then a new [leader] came who had a very, very different mind-set and was very honest, and very transparent and did not do business that way at all, and worked to get rid of the inequities that had resulted on the basis of the previous [leader].… I saw that people were embarrassed to be petty and suspicious and uncollegial because they were always treated with respect by the [new leader].

Furthermore, the inability of leaders to recognize, acknowledge, and confront incivility was thought to perpetuate a culture of silence, which appeared to further institutionalize the hidden curriculum. Several participants offered hypotheses as to why leaders may have remained complacent about incivility including inadequate training / skills / resources; unclear personal or organizational consequences; apathy toward culture change; fear of disciplining senior or accomplished faculty members; and lack of success in previous attempts at conflict resolution, allowing “historical bullies” to thrive. One participant reflected:

I think that in large institutions that have a great deal of inertia related to these issues it can become even more disempowering at the level of an individual interaction around the prospect of, you know, authentic change.

Some participants thought that unaddressed incivility contributed to the cultivation of power cliques. They characterized these power cliques as a small number of individuals, including some leaders, who banded together and excluded those who were perceived as a threat to the group’s collective success. Some participants felt that these groups enjoyed special privileges and opportunities, and protected each other from external concerns about uncivil behavior, as this comment illustrates:

Subtly, perniciously, systematically, you create a power clique that is allowed to do whatever they want to do, treat people however they want to treat them, have their own rules.

Systems-based solutions

The participants offered several potential systems-based solutions to curb or mitigate incivility within the DOM. Broadly, their suggestions related to preventing, reporting, and addressing incivility. Across this spectrum of activities, common themes emerged: the need for confidentiality, clear consequences, and an emphasis on rehabilitation for the victim and the perpetrator.

Preventing incivility.

Most participants believed that preventing incivility requires increased awareness and accurate identification of uncivil behaviors. Some participants noted that there may be subjective differences in the interpretation of uncivil behavior, and if there is a “gray zone” for professional behavior, they believed that the tendency would be to allow incivility to persist. They furthermore suggested that definitions are required so that individuals do not report incivility when none has occurred—for example, if an individual is trying to undermine constructive feedback that he or she has received by identifying it as bullying. Therefore, several participants felt that any definition must be accompanied by examples. Some participants also noted that a definition of incivility must also outline the roles and responsibilities of the hospitals and the university. Given the nonemployee status of physicians, they suggested that role clarification might identify the jurisdiction for addressing disruptive, uncivil behavior. Participants worried that local politics in hospitals might attenuate the action taken to address uncivil behavior, indicating a possible role for the university. Some participants also suggested that sharing anonymous data about incivility (including examples) across the DOM might serve as an educational tool and a deterrent.

Several participants suggested that leaders need to be provided with training and supports to role model professional behavior and identify incivility. They also thought that rotating leadership structures—in which a leader serves for a fixed number of years or terms—may permit the containment of power cliques, especially in circumstances in which an outgoing leader is, him- or herself, a member of such an exclusive group. They suggested that changing leaders may renew energy to foster culture change without the baggage of prior failed attempts at disciplining historical bullies. Most participants reported that leaders should be empowered to factor “citizenship” into decisions around hiring, performance evaluations, and academic promotions. This would make expectations about professional behavior explicit across the academic life cycle. It was suggested that data regarding a faculty member’s behavior could be obtained through the regular use of tools like 360-degree feedback from peers and anonymous feedback from trainees. However, for the use of these sources to be meaningful, faculty members should be supported by mentors or coaches to interpret the data received and identify steps toward behavioral change. Increasing transparency, particularly with regard to physician salaries, may also increase citizenship in the DOM by ensuring the equitable distribution of resources and minimizing competition.

Reporting incivility.

Participants uniformly agreed that reporting mechanisms need to be improved. They suggested that the processes should be clear; confidentiality must be ensured; and there needs to be a clear expectation of steps, timelines, and potential consequences. They felt that there were several potential pathways for reporting: anonymous online reporting, reporting to an ombudsperson or specially appointed clinical faculty advocate, or reporting to an external party. In favor of the latter, one participant noted:

I would like to have a really effective independent board outside of the institution not composed of people who have any invested interests in the protecting of the career of anybody else, who can hear this stuff and make sure that the people are appropriately disciplined.

Most participants noted the importance of strategies to protect individuals from potential retribution for filing a complaint but also noted challenges to practically achieving this goal.

Addressing incivility.

In addressing incivility, most participants emphasized a fair and rehabilitative approach. They indicated that investigations should seek to understand the perspectives of all involved parties and prevent the individuals raising concerns from experiencing further victimization. They highlighted the need to understand the root causes of incivility, which often stem from issues outside of personality and may be symptomatic of burnout, depression, or other stresses,[31] so that perpetrators of incivility can be connected with mentorship, coaching, or other resources as indicated. They emphasized that recurrent or remorseless incivility should be confronted with progressive disciplinary action; otherwise, incivility will persist. One participant observed:

It just seems to me that the people who do it are, like everyone knows who they are and they never seem to have, suffer any consequences. They get promoted, you know they get the same resources. I don’t see any consequence.

Several participants recommended that clear consequences of disruptive behavior be articulated so that there can be a true zero tolerance mandate and a uniform application of policies. They noted that equipping the leadership with training, skills, and resources to enact disciplinary action may be of benefit, as one participant commented:

All these mechanisms you might have, unless someone is willing to actually act on it … I don’t know that it’s really going to do anything.

Discussion

Despite the presence of many collegial, supportive, and positive working relationships at academic medical centers, our study shows that uncivil behaviors unfortunately still exist, and they have direct and indirect consequences for individuals and the collective good. Organizational factors that may facilitate or exacerbate incivility included physician nonemployee status, silos within the department, poor leadership (in the hospitals and university), a culture of silence, and power cliques. At the systems level, many potential solutions were suggested to improve the prevention, reporting, and addressing of incivility.

Among the health disciplines, the literature to date on incivility has stemmed largely from nursing. This literature has identified several potential causative factors of incivility between nurses, including individual behavior, unclear responsibilities, dissatisfaction with the performance of colleagues, failure to adhere to protocols or assignments, limited resources, and high workloads,[32,33] with certain leadership styles having a modulating effect on incivility.[34] Our study builds on this existing literature by examining the perspectives of physicians working at academic medical centers in Canada and by focusing on contributing factors at the organizational level. Professionalism is a core competency for physicians, with explicit mention of civil conduct as an imperative in the widely embraced Physician Charter, under the Principle of Social Justice.[35] However, some still argue that it is a nebulous concept,[36] with existing frameworks often overlooking the social and cultural context of professionalism in health care organizations.[37,38] A recent advance that has elevated the role of organizational culture in promoting citizenship is the Charter for Professionalism in Health Care Organizations.[39] In view of the growing appreciation of the organization’s role, our exploration of possible systems-level solutions marks an important advance on this issue. Although interventions for reform have largely focused on individual awareness and behavior, the role of institutions in shaping workplace culture has been granted insufficient consideration.[40] A systematic review examining prevention strategies yielded low-quality evidence with equivocal impact on behavior change.[15] A recurring theme in our interviews was the need for clarity around definitions, roles and responsibilities, processes, jurisdictions, and consequences to ensure more robust systems-level solutions. A key strength of our study is that it was conducted in the largest DOM in Canada and consequently had a large sample size, representing a plurality of opinions. By including five different sites affiliated with a single large DOM, we were able to capture a variety of local work cultures (with inherently diverse governance structures and politics) which share similar pressures due to the academic mandate of the university; this strengthens the transferability of the study.

This study has several limitations. Firstly, it was conducted within a single DOM. However, our sample size was large for an interview study, and we included participants from all divisions, ranks, academic position descriptions, and, as noted earlier, several affiliated teaching hospitals—allowing us to capture diverse viewpoints emerging from different local sociopolitical work environments. We did not explicitly ask participants if they ascribed a hospital versus university source of the organizational factors that may contribute to incivility. Nonetheless, similar themes emerged across the sites, highlighting the need for hospitals to work in partnership with the university to address incivility—something that is already happening at the University of Toronto through regular meetings between the hospital physicians-in-chief and the DOM chair.

We used a nonprobabilistic sampling technique, so our respondents may differ from nonrespondents in meaningful ways, indicating potential selection bias. For example, individuals may have been more likely to participate if they viewed incivility as a problem. We had a gradient of respondents by academic rank (2 lecturers, 12 assistant professors, 15 associate professors, and 20 full professors), suggesting that junior faculty members may have abstained from participating, either because they had less exposure to incivility or because they feared the potential consequences of disclosing their experiences. Indeed, the role of power differentials in shaping uncivil behaviors (and the response to such behaviors) was not directly examined in our study and warrants further exploration. Future studies will also benefit from taking an intersectional approach to examining incivility at academic medical centers, so long as safeguards can be provided to protect the anonymity of faculty members from underrepresented or marginalized (and potentially identifiable) groups.[41,42]

Organizational factors that contribute to disruptive and uncivil physician behaviors may be malleable to change. With greater understanding of the role of health care organizations, we can design and test systems-level solutions to reshape workplace culture and combat incivility. Within our own DOM, we will use these findings to develop a multimodal strategy to promote safe reporting, fair investigations, and a rehabilitative approach to disciplinary actions.

Acknowledgments:

The authors thank the members of the Department of Medicine who participated in this project.