Advancing Culture Change in Radiology Through Communication: The Value of Large- Versus Small-Group Discussions

In the midst of a global pandemic and nationwide calls for social justice, the urgency to advance health equity, combat bias, and eliminate racism within our health care system has increased substantially. The importance of diversity, equity, and inclusion and its implications in improving workplace culture, community engagement, patient satisfaction, and health outcomes have been well documented in the literature [1Cohen J.J. Gabriel B.A. Terrell C. The case for diversity in the health care workforce.Health Aff (Millwood). 2002; 21: 90-102Crossref PubMed Scopus (376) Google Scholar]. To better understand how to address these issues, health care organizations have had to ask themselves how to unequivocally reverse decades of inequitable systems, policies, and beliefs. One method that has been effectively utilized to properly address these issues is culture change [2Miller F.A. Strategic culture change: the door to achieving high performance and inclusion.Public Pers Manag. 1998; 27: 151-160Crossref Scopus (68) Google Scholar]. Culture change, described as transformations in values, beliefs, and assumptions of people with shared organizational membership, uses specific interventions that aim to overhaul them. This is generally done through a combination of identifying the reason for the change, mobilizing a diverse group of advocates, creating avenues for success, leadership coaching, and tangible change through evolutions in policy, norms, practices, and values [3Dreachslin J.L. Weech-Maldonado R. Dansky K.H. Racial and ethnic diversity and organizational behavior: a focused research agenda for health services management.Soc Sci Med. 2004; 59: 961-971Crossref PubMed Scopus (39) Google Scholar]. Mobilizing and amplifying employee voices in particular has emerged as an important step in this process to properly evaluate and rectify deficits in workplace culture. Group dynamics, particularly group size, have been studied extensively and have drastic effects on how groups ultimately function and achieve their goals. Larger groups tend to have a small number of people making decisions but many do not participate or fully interact with the group at all, resulting in lower satisfaction and engagement and higher dropout rates [4Borek A.J. Abraham C. How do small groups promote behaviour change? An integrative conceptual review of explanatory mechanisms.Appl Psychol Health Well Being. 2018; 10: 30-61Crossref PubMed Scopus (33) Google Scholar]. Alternatively, smaller groups allow for a greater degree of social interaction and relationship development, which generally results in more optimal outcomes. The effect of group size on health care–focused culture change specifically has not been adequately studied and may have valuable implications in future diversity, equity, and inclusion work. With renewed call and urgency for action, our department as part of a concerted effort within our large, urban academic institution adopted a strategic framework to address health inequities, bias, and racism and begin the process of culture change. Understanding the importance of employees’ voices and departmental culture, we deployed two forms of communication-based initiatives to solicit ideas, identify critical action items, and generate feedback from employees within our department. A large departmental town hall event and seven small focus groups were conducted, both in virtual format. The town hall was a 60-min session to which all employees and staff were invited. After ground rules were established, several reflections detailing experiences of bias or racism were shared to the group, some of which were heard directly from the victims. Ideas for addressing systemic and overt racism were also shared, including both ideas shared directly by creators and indirectly by presenters bringing previously submitted ideas to the floor. At the end of the session, 15 min were dedicated to open discussion and questions. The focus groups were 90-min sessions each with four facilitators and ranging from 15 to 27 participants. Two sessions were conducted with medical and research trainees only, which we characterized as homogeneous small groups (group A), and the remaining sessions were conducted department-wide with heterogenous groups of staff and faculty at all levels (group B). These were conducted as open forums in which discussion was elicited through a list of prepared questions. For both event types, participants could communicate verbally or by chat message with the option to be anonymous, as provided by the virtual platform BlueJeans. The chat function was monitored throughout the session by at least two individuals who were not participating as moderators or facilitators. Before and after the sessions, employees were allowed to ask questions and provide comments anonymously through an online survey. De-identified transcripts of verbal and written comments were recorded for the pre-established goal of data gathering and data sharing that would protect participant anonymity. Our primary outcome was employee satisfaction and perception of effectiveness of the large-group versus small-group sessions, assessed through an anonymous survey. Survey questions were generated based on our LISTEN framework, which we propose as a tool to maintain effective communication at the departmental and organization level with potential to transform culture around diversity, inclusion, and equity (Table 1). There were five possible survey options from strongly disagree to strongly agree. Differences in each question response between groups were assessed using χ2 trend test. Because of instances of low data counts, Fisher exact test was also performed. Subgroup analysis of homogeneous versus heterogeneous small groups was done.Table 1The LISTEN framework: a proposed tool to maintain effective communication at the departmental and organization level as we look to transform culture around diversity, inclusion, and equityAcronymStrategyApproachLProvide time to listen and learn.•Try to understand employees’ concerns.•Ask the right questions.•Schedule regular sessions.•Maintain an open line of communication.IPromote constant flow of ideas and information.•Share your ideas for action.•Elicit and embrace ideas from the team.•Send frequent updates and progress reports.SCreate safe spaces for discussion.•Provide different types of environment for conversation.•Allow anonymity.•Create a judgment-free zone.TTake action.•Use ideas to generate action plans.•Highlight how ideas will be implemented•Have a clear timeline for the plan of action.•Follow through on the action plan.ENEncourage engagement.•Provide other opportunities to be involved in departmental initiatives.•Highlight dedication and hard work of those involved. Open table in a new tab Our secondary outcomes were the effectiveness of small versus large groups and homogeneous versus heterogeneous small groups in generating action ideas. To assess this, we developed a scale to assess the strength “grade” of each action proposal from the sessions (Table 2), loosely modeled based on the early stages of Design Thinking [5Sándorová Z. Repáňová T. Palenčíková Z. Beták N. Design thinking—a revolutionary new approach in tourism education?.J Hosp Leis Sport Tour Educ. 2020; 26: 100238Google Scholar]. The blinded transcript comments were reviewed and scored by three primary reviewers. The final score was assigned if two or more reviewers were concordant. Cases without concordance were reviewed by a fourth blinded reviewer. χ2 analysis was performed.Table 2Scale established to assess the strength of an action proposalGrade0—Not applicable1—Recognition of a significant problem without an action proposal2—Big picture ideas without a specific action item3—Poorly defined idea for action toward a specific outcome or goal4—Well-defined idea for action without a measurable outcome5—Well-defined idea for action with a measurable outcomeEach reviewer assigned a score from 0 to 5 to each comment, with 0 indicating that a comment was not related to the problem-solving process. Open table in a new tab Each reviewer assigned a score from 0 to 5 to each comment, with 0 indicating that a comment was not related to the problem-solving process. A total of 156 employees participated in the town hall, and 155 participated in the small-group sessions. There were 22 survey responses from the large-group and 38 from small-group participants, corresponding to a response rate of 14.1% and 24.5%, respectively. There were 11 participants in the homogeneous (group A) and 27 in the heterogeneous (group B) small groups. Demographic and identity characteristics of survey participants are detailed in Table 3.Table 3Demographics and identity characteristics of survey participantsCharacteristicLarge Group (n = 22), %Small Group (n = 38), %All Participants (n = 60), %Age (y) 26-359.129.021.3 36-4536.415.823.0 46-5522.723.723.0 56-6531.823.726.7 Other or not disclosed-7.83.3Race or ethnicity Asian9.113.211.7 Black or African American4.615.811.7 Hispanic or Latinx-7.95.0 White63.660.561.7 Other or not disclosed22.72.610.0Gender identity Cisgender female54.660.558.3 Cisgender male40.934.236.7 Other or not disclosed4.65.35.0Sexual orientation Heterosexual72.784.280.0 Lesbian, gay, bisexual13.810.510.0 Other or not disclosed13.65.310.0Survey included broader spectrum of identities than could be individually listed in table. Open table in a new tab Survey included broader spectrum of identities than could be individually listed in table. Of 16 survey questions, 9 showed significant results using χ2 analysis with 5 remaining significant using the Fisher exact test (Fig. 1). The trend was favorable toward small groups with results suggesting that although both were valued by participants in culture change, small groups were perceived as more effective in having purposeful conversations in which ideas and concerns could be shared with less judgment and bias. In our subgroup analysis, only one survey question demonstrated a significant difference between group A and group B with group A, composed of either medical or research trainees only, demonstrating greater concern for judgment from department leadership. For our secondary outcomes, there were 412 comments total with 400 from the small groups and only 12 from the town hall. Given the significant difference in number of comments generated between the groups, a statistically useful comparison of the strength of action proposals between large and small groups could not be made. Of small group comments, 36% were not considered related to the problem-solving process (rated 0), 43% were rated 1 to 2, and 21% rated 3 to 5. All large-group comments were rated 0 to 2. There was no difference in strength of comments between group A and group B (P = .643). The study has several limitations. First, we had a total survey response rate of 19% and more responses from small-group than large-group participants, which may contribute to sampling bias and limit external validity. This limitation was partly addressed by employing both the χ2 trend and Fischer exact tests with results consistently favoring small groups. Second, we compared several small-group sessions with a single town hall event that did not allow as easy opportunities for comment and idea sharing. Nonetheless, our study helps to highlight the fundamental limitation of large groups in effectively hearing employees’ voices, even with opportunities for written and online comments. Third, the comments generated from the sessions were scored using a nonstandardized, untested scale and using a collection of comments that may not have captured the full context of a conversation, leaving room for interpretation by reviewers. Nonetheless, the small percentage of comments requiring fourth review (44 of 412, 11%) suggests some validity to our scale. Quy Cao, PhD, provided statistical consultation and analysis. Catherine Oliva provided data collection support.