Collective Compassion: Responding to Structural Barriers to Compassion With Agentic Action in Healthcare Organizations

A Case for Observing Compassion in Healthcare

Some scholars argue that all organizations have the capacity to foster compassion and virtuousness (e.g., Cameron et al., 2003), but evidence of compassion is less visible in certain contexts. As an example, healthcare organizations generally operate under a biomedical discourse that emphasizes rationality and minimizes emotions (Bisel & Zanin, 2016; Dutta, 2008). Additionally, most healthcare organizations in the United States prescribe to a “managed care” model—a healthcare system and ideology driven by capitalistic logics, cost-cutting measures, and insurance-driven healthcare (Harrill & Melon, 2021; Ray & Apker, 2010). Managed care reflects bureaucracy—one of the most pervasive organizational forms in Western society—and features formal hierarchies, rigid structures, and standardized policies (Lillis & Varetto, 2020). In managed care, communication from leadership is commonly described as top-down, one-way, and unsupportive (Apker et al., 2021).

Under the influences of managed care, healthcare workers are generally socialized to be emotionally distant at work (Underman & Hirshfield, 2016), even though they are regularly placed into delicate situations where emotion management skills are necessary (Carminati, 2021). Medical students receive competing messages to both value patients as people and value efficiency and cost-saving (Harter & Kirby, 2004). This emotional labor double-bind is common in healthcare (Launer, 2010), leading many physicians to fear situations requiring empathy. Often, they view this emotional work (i.e., expressing authentic emotions as a part of their job) as a “Pandora’s box” that is difficult to control (Hardee & Platt, 2010). Moreover, many structural constraints within healthcare organizations create and exacerbate stress (e.g., hospital restructuring, high turnover, staff shortages), often with insufficient resources and pathways for support (McSherry & Pearce, 2018; Ray & Apker, 2010).

Given these conditions, burnout rates are high among all types of healthcare workers and were exacerbated during the global COVID-19 pandemic (Morgantini et al., 2020). Burnout is connected to several consequences including negative health outcomes (Maslach et al., 2001), poor physician-patient rapport (Ratanawongsa et al., 2008), and ineffective communication with patients (Passalacqua & Segrin, 2012). To combat burnout, some healthcare organizations attempt to cultivate compassion through formal workplace initiatives (Hewison et al., 2018), but many healthcare leaders simply leave the need for compassion unaddressed (McClelland & Vogus, 2021). In extreme cases, organizational leaders may reject or neglect the value of compassion in the workplace, even when healthcare workers directly express a need for support (Egan et al., 2019).

The perception that collegial compassion does not belong in healthcare organizations is problematic, especially considering the positive benefits that compassion may bring to healthcare workers suffering from stress and burnout (Gerber & Anaki, 2021). Without structural support for compassion and endorsement from leadership, compassion in healthcare is relegated to an individual responsibility. As such, expressions of social support mostly occur in the backstage of healthcare, such as in informal conversations (Ellingson, 2003) and meetings (Wittenberg-Lyles et al., 2013) between coworkers. In many healthcare organizations, compassion largely manifests as a form of individual and interactional work.

A Communicative Lens: Compassion as Interactional

Early conceptualizations positioned compassion as effortful, individual work—a three-step process of (a) noticing an individual’s suffering, (b) feeling their emotional pain, and (c) responding with an action that could alleviate their suffering (Kanov et al., 2004). Building on and challenging such works, Way and Tracy (2012) reimagined compassion as a communicative and interactional process of: (a) recognizing, (b) relating, and (c) (re)acting.

Whereas noticing was primarily a cognitive skill in previous models, recognizing requires “understanding and applying meaning to others’ verbal and nonverbal communicative cues, the timing and context of these cues as well as cracks between or absences of messages” (Way & Tracy, 2012, p. 307). Recognizing involves attention and interpretation of the meanings of spoken and unspoken pain. As such, recognizing suffering in others necessitates being physically and emotionally present (Durkin et al., 2019). Recognizing involves stopping, “turning toward the other … with one’s eyes, face, and body” (Huffman, 2017, p. 159), and being attentive to the ways that others wear their suffering whether or not they explicitly call it out.

Next, extending Miller’s (2007) concept of connecting as a relational process, Way and Tracy (2012) describe relating as “identifying with, feeling for, and communicatively connecting with another to enable sharing of emotions, values, and decisions” (p. 307). Relating moves beyond the internal notion of “feeling” to emphasize the importance of communicative interaction in building deeper connections and establishing common ground between two parties. For example, relating between a healthcare provider and patient requires intentional efforts to see each other as humans with unique needs; this involves soliciting and appreciating a person’s “broader life story” (Sinclair et al., 2018, p. 7). Such a process may include mutual self-disclosure and employee vulnerability so that patients realize they are in a space that is safe for sharing their own suffering. Relating also manifests through empathic consideration (i.e., imagining and inquiring about the experiences of people in different positions) (Way et al., 2015).

Finally, (re)acting is “engaging in behaviors or communicating in ways that are seen, or could be seen, as compassionate by the provider, the recipient and/or another individual” (Way & Tracy, 2012, p. 307). The parentheses around the “re” in (re)acting signify that proactive compassionate action can happen even before suffering has been recognized. Compassionate action can range from grandiose gestures of support, to intentionally leaving someone alone, to subtle acts of kindness such as a friendly hand on a shoulder (Sinclair et al., 2017). By this logic, people engage in compassion by exercising skills such as listening, perspective-taking, vulnerable self-disclosure, and supportive messaging which may build hope (Tracy & Huffman, 2017). Ideally, individuals also account for relational context and history when employing skill sets related to compassion (Dutton et al., 2014).

A Structurational Lens: Compassion as Organizational

While compassion is indeed individual and interactional (Way & Tracy, 2012), it is also important to consider the broader context in which interactions occur. Everyday interactions are grounded in “social structures of meaning, norms, and power” (Canary, 2017, p. 1688) that provide guidelines for behavior. The communication of compassion is shaped by and shapes organizational forces that indicate what it means to recognize, relate, and (re)act to pain.

Structuration theory (Giddens, 1984) highlights the inherent complexity of organizing by centering on the duality of structure, where “structure is both the medium and the outcome of the human activities which it recursively organizes” (Giddens, 1986, p. 533). Structures (i.e., rules and resources) constitute meaning and sanction modes of social conduct, providing guidelines for how to navigate social interactions (Giddens, 1984). Structures in healthcare can be as overt as the Hippocratic oath (Olufowote, 2008) and formal medical laws in the judicial system (Zanin & Piercy, 2019). However, many rules of organizational life are tacitly known, learned through routine procedures and habitual encounters at work. Employees often embrace the ontological security of routine by (re)producing existing structures, but they also have agency (i.e., the ability to act otherwise) to challenge and transform structures (Giddens, 1984). For example, the ideological commitment to “do the right thing” for patients (Carmack, 2017, p. 36) can be clearly defined in workplace directives (e.g., policies, training), but such meanings can also be spontaneously challenged in situations where standard procedures may endanger patient safety (Groves et al., 2011).

To act otherwise or challenge the standard routines in an organization, agents must draw upon various structural resources. However, while all actors have agency, not all actors have access to all the options and resources of structures, particularly those in low authority positions (Zanin, 2018). Indeed, policy as written is often starkly different than policy as practiced (Kirby & Krone, 2002). Beneficial resources (e.g., work-family initiatives, paid time off) can seem inaccessible for employees, especially when there is peer pressure to refrain from utilizing such resources (Kirby & Krone, 2002). When there is a greater concern for preservation of routine, employees are more likely to (re)produce existing structures (Giddens, 1984). Presumably, agents try to act in ways that would benefit them, but the outcomes of any action are not guaranteed and may result in unintended, contradictory consequences (Giddens, 1984, 1986).

Like any other communicative process, compassion can be simultaneously enabled and constrained by employees’ (re)production of organizational structures. Especially in the emotionally-complex organizing of healthcare, there may be unique structures that influence how employees understand and express compassion to each other. Many workplaces institute feeling rules that discipline employees’ emotions, consequently categorizing some emotions as productive (e.g., patience, resilience, stoicism), others as feminized and low status (e.g., empathy and care), and others as problematic (e.g., stress and burnout) (Tracy & Malvini Redden, 2019). Feeling rules can be especially blatant in healthcare, where expressions of regret or grief are permissible in apologies for medical mistakes but not beyond (Carmack, 2010). In healthcare, compassion is often allowable toward patients but masked among coworkers due to prevailing emotion norms of professionalism and disidentification (Tracy & Malvini Redden, 2019). Some healthcare workers (e.g., nurses) also receive mixed messages about the appropriateness of emotions at work due to incompatible structures that prioritize either relationships or tasks (Nicotera & Clinkscales, 2010); this may result in further immobilization, burnout, and depersonalization (Glasberg et al., 2007; Nicotera, 2015).

Given that structure and agency are interdependent and inextricably tied together (Giddens, 1984), it is vital to investigate how healthcare workers communicatively draw upon structures to (re)produce or transform rules and resources related to compassion. This study brings to bear how current understandings of compassion as dyadic may be extended through a lens of structuration theory and, in turn, illuminates how compassion is communicatively and collectively constructed in the recursive space between structure and agency. Therefore, we pose the research question:

RQ

How do healthcare workers exercise agency to (re)produce or transform structures related to the communication of compassion in healthcare organizations?

Data Collection

Data Analysis

The present study utilized a phronetic iterative approach, where data collection and analysis were both guided by a combination of preexisting theory and emergent qualitative data (Tracy, 2020). Although the analytic steps described may suggest a sequential order, a phronetic iterative approach is nonlinear (Tracy, 2020). To begin, the first author checked the accuracy of all transcriptions. While reviewing the interview data, the first author loosely held onto sensitizing concepts related to organizational compassion (e.g., Way & Tracy, 2012) and structuration theory (Giddens, 1984). Second, the first author revisited the audio recordings of interviews and engaged in memo-writing as “an interactive space and place for exploration and discovery” (Charmaz, 2014, p. 170). Third, the first author regularly met with the second and third authors to discuss notable stories and emergent findings, creating an additional space for data immersion and sensemaking (Tracy, 2020). Finally, the first author engaged in primary-cycle and secondary-cycle coding to distill and reorganize data into meaningful codes.

Guided by the research question, coding began through a data reduction process, particularly focusing on concepts of structure and agency (Giddens, 1984). Next, each reduced piece of data was compared with the next. If a piece of data did not fit within a previously established code, a new, clearly defined code was created. This process continued until no new codes emerged in relation to the research questions. In the secondary cycle of coding, the first author employed axial coding to reassemble the fractured data (Charmaz, 2014) and organized initial codes under second-level hierarchical codes (Tracy, 2020). This analysis revealed nuanced relationships among primary and secondary cycle codes.

Research Context: Structural Constraints to Compassion in Healthcare

Echoing prior research on managed care (e.g., Harter & Kirby, 2004; Ray & Apker, 2010), participants confirmed that “money” and “business” were the main priorities of their organizations. George, a physician assistant, explained, “Sometimes all [the managers] see is dollar signs, and then I’m concerned about the patient in front of me.” Like George, many participants felt that financial concerns were in direct conflict with a patient-centered approach in healthcare emphasizing patient-centered communication and patient autonomy (Gray, 2011).

The drive to stay financially solvent can be seen in the design of healthcare workers’ workloads and schedules. Most participants described their workload as “overwhelming” and “exhausting,” both in quantity of tasks and expertise needed. Emma, a registered medical assistant, described her typical workday in which she would: “see 30 patients, call in all the prescriptions, do all the referrals, do all the prior authorizations, return all of [the] phone messages” and more. For most participants, workday expectations also dramatically increased during COVID-19. Overpacked schedules limited employees’ time and ability to meaningfully interact with patients and coworkers. Participants also struggled to find time to seek out mental health resources. Although many healthcare organizations offered a variety of resources (e.g., counseling, wellness coaches, dieticians, crisis hotlines), most participants did not use them.

Caroline, a paramedic technician, explained: “They have all of these things, but they don’t necessarily make time for us to go there. Nobody wants to be there longer than 12 h, so I don’t know a lot of people that use them.” Overwhelming workloads represent a domination structure in healthcare, which describes the power dynamics that affect the mobilization of resources (Olufowote, 2008). Healthcare organizations rely on both material resources (e.g., monetary support, staff, medications) and nonmaterial resources (e.g., mental health resources, support groups, education) to survive, but the allocation of such resources depend on how power is distributed in the organization (Zanin & Piercy, 2019). Caroline’s explanation for why she cannot take advantage of helpful resources is linked to the perception that management does not make utilization of resources possible. A structure does not need to be supported by all parties in an organization to have an impact on daily life, but endorsement from organizational leaders is more likely to sustain and entrench particular structures (Giddens, 1990).

Additionally, participants were sensitive to the ways that mental health was stigmatized at work. Consistent with prior research, data indicated that participants were socialized to hide any negative emotions (e.g., fear, sadness, anxiety) (Underman & Hirshfield, 2016) and display neutrality instead (Lammers & Garcia, 2009). Participants disclosed feeling socially and professionally obligated to suppress their emotions. Will, an emergency medical technician (EMT), explained:

Amongst each other we try to be very tough, but also in front of the patients. I’m not sure it would be good for them to see us express those emotions, but we still feel the need to be strong and not show those emotions.

Participants also described this emotional repression norm as “bottling up,” “powering through,” “pushing aside,” and “keeping the lid on” emotions at work. By stifling their emotions both verbally and nonverbally, participants likely hid the communicative cues that could signal a need for compassion from others. Together, the managed care model and biomedical discourse privilege profit and rationality, leaving little space for emotion-sharing and compassion. Despite these structural constraints, though, healthcare workers in this study exercised their agency to create opportunities for compassion in their workplaces.

Agency in Recognizing: Earnest Script-Breaking

While participants described several organizational barriers to engaging in and receiving compassion in their workplace, they also engaged in creative agency to cultivate compassion in their workplaces. An interesting technique participants employed for recognizing suffering (Way & Tracy, 2012) was to ask coworkers “how are you?” twice in a row. Daisy, a nurse practitioner, made this repeated question part of her daily routine:

If I ask a patient or if I ask one of my co-workers, “How are you?” and their answer is a little bit withdrawn, then I ask again, “How are you?” And then typically in that second question, I’ll get more truth to their response with either tears and “I’m not okay.”

Daisy often paired her words with physical touch, “whether it’s just a pat on the shoulder or on their back or actually giving them a hug.” Like Daisy, many participants noted that persistence was important when checking in with their coworkers. Lucas, a resident physician, described a time when he was reading his team’s work schedule and noticed that one of his coworkers was on a bereavement leave. Lucas said:

When she came back into town, I saw her and I was like “Hey, how are you doing?” And she’s like “I’m doing good.” And I’m like, “I saw the schedule. Like really, how are you doing?” She was like, “Very sad and feeling very tired … I’m feeling very burnt out.”

Here again, asking “how are you?” twice in a row is key to inviting a sincere response. With their overwhelming workloads and chaotic schedules that were especially exacerbated during COVID-19, many participants were worried about missing nonverbal cues of distress in coworkers, so they verbally checked in with their peers instead.

Madeline, a nurse practitioner, also committed to checking in with others as part of her routine: “I’ll check in with [nurses] that go in and out of the hospital, like, ‘Hey, how are you doing? Can I help you in any way?’ … Just a check in. Make sure their heads [are] above water, basically.” Madeline highlights the importance of quick, subtle check-ins with her coworkers, indicating a mindfulness of the ways her actions might be considered unprofessional to others. A check-in allows workers to break norms of emotional suppression in a way that poses few risks to the asker. As previous scholars have suggested, however, “small moves” of compassion do not necessarily translate to a small impact (Frost et al., 2006, p. 18). Sometimes small acts of compassion may be relatively unseen, but often these acts were remembered and (re)produced by other organizational members.

We call these expressions of agency earnest script-breaking, where an employee recognizes the scriptedness of social interaction (e.g., asking “how are you”), disrupts routine with intentional verbal and/or nonverbal communication, and creates space for genuine conversation and emotion-sharing. In the context of healthcare, participants largely follow a script of emotional suppression and small-talk niceties. Participants indicated that the everyday script of a single “how are you” rarely evoked emotion or sincerity out of others. Due to the norm of emotional suppression, workers’ first instinct may be to mute themselves or withdraw from conversation. By asking a second “how are you,” participants tactfully (re)structured organizational norms to create space for compassion. This agentic (re)action challenges structural constraints on emotion-sharing. As Giddens argues, constraints do not “push” an individual to do something if the individual has not already been “pulled” to act purposefully (Giddens, 1984). In other words, healthcare workers in this study choose to recognize suffering despite structural constraints. Yet, they also choose the most feasible option for doing so in an organization where emotional displays are not always welcome.

Agency in Relating: Creating Spiral Time

In an organization where emotional problems are rarely addressed, employees may struggle to identify situations where they can demonstrate empathy for coworkers while simultaneously maintaining ideals of professionalism and emotional stoicism. In the compassion process, relating necessitates communicative connection where people can share their emotions freely (Way & Tracy, 2012). To this end, participants exercised agency by creating “spiral time,” a discursive space where coworkers can vent to each other about various stressors in their day. The term “spiral” originates from an interview with Samantha, a psychotherapist:

We call them spirals. We tell each other, “You get two spirals a day, and that’s it.” We jokingly tell people to “get off of their spiral, because you're spiraling too much, and you need to get off of that one” … And like, “You’ve already had a spiral today or you’ve had one big one, so that’s it for you.” It brings us back down to earth a little bit and tries to redirect to “What’s important here?”

Spiral time provides a sanctioned but compartmentalized opportunity for healthcare workers to share their emotions and identify with one another. For many, spiral time is a means of compartmentalizing difficult emotions that may affect their ability to work. Participants’ desire to compartmentalize emotionality and relationship-building apart from rationality and work indicates an awareness of contradiction in the workplace. The knowledge of structural constraint can become a springboard for structural resistance and transformation, but as shown in this study, employees often choose to monitor and (re)produce existing structures in ways that (seemingly) serve them (Giddens, 1984).

Privacy is also an important element of spiral time. For example, when Will and his fellow EMT spent almost 2 hours doing CPR on a young patient, Will explained that he felt the need to hide his emotions. Following the event, Will and his coworker took a brief “moment” in the locker room to allow themselves to become “teary-eyed” with no one around, but “only a moment.” Will shared: “Then we came out and just had to pretend that everything was cool … We don’t want people to think we can’t handle what’s going on around us.” Like Will, many participants were worried that emotional displays would signal incompetence to coworkers and patients. Spiral times serve as safe spaces to feel the stress and heartache that often manifest in healthcare work.

Given that most healthcare workers cannot easily access mental health resources or feel too exhausted to do so after long shifts, spiral times are vital spaces that help many healthcare workers cope during difficult days. However, the inherently limiting nature of spiral time also poses a potential cost to healthcare workers. Cecilia, a physician assistant, explained:

[Spiral time] costs me later. If something was rough or hard, it will come up when I don't expect it. So I can definitely turn it off or compartmentalize or whatever you want to call it to get through the shift, right? But then on a day off or … when it’s safe because people aren’t depending on you, then you feel it.

In Cecilia’s example, it is evident that spiral time is not a long-term solution. Instead, spiral time may function as a stress valve, releasing just enough emotional distress to allow healthcare workers to survive busy and difficult days at work. Spiral time represents the efforts of healthcare workers caught between the desire to transform structure and the pressure to (re)produce it.

Agency in (Re)acting: Coordinating Compassion as a Collective

Behavioral or communicative (re)action is at the core of compassion (Way & Tracy, 2012). Despite overwhelming workloads and schedules, healthcare workers often look for ways to (re)act compassionately when their peers need support. Participants collaborated with their coworkers to coordinate actions that could alleviate suffering. These collaborative acts ranged from informal and spontaneous practices to formal and planned efforts. For example, participants frequently shared workloads and took on tasks outside of their job description when they noticed a peer struggling. Fiona, a nurse practitioner, explained that sharing workload was about acknowledging each other’s humanity: “At the end of the day, everyone looks out for each other … If somebody’s struggling or somebody can’t show up or something, you should never say, ‘That’s not my job.’” The firmness of Fiona’s claim that one should “never” say no to helping others indicates how normative sharing workload is in healthcare settings, but it also reveals how structures that constrain compassion are (re)produced and maintained. Even though sharing work is a compassionate act for overwhelmed workers, the norm also sustains overwhelming workloads that were of the employee’s own making. Indeed, when employees choose to share work, this “choice” can serve to disincentivize organizational power holders in otherwise addressing healthcare workers’ overwhelming workloads. Despite this (re)production of problematic structures, sharing workloads represents an important pathway for building rapport and trust so that compassion might continue to flourish between coworkers in the future.

Participants also discussed the ways they intentionally organized to support each other, particularly in times of exceptional hardship (e.g., death of a family member, illness, surgery, loss of home). Losing someone to COVID-19 also constituted a serious hardship, according to participants. Common examples of collective compassion include pooling paid time off (PTO), coordinating meal trains, and fundraising to support a coworker. Although these acts can be done individually, most participants described these actions as a collective effort. Participants claimed that working together often led to “bigger” or more “memorable” outcomes. For example, the effects of a PTO donation may only be memorable when a certain number of hours accumulates for the person in need. Helen, a pharmacist, described a time when she and her coworkers pooled their PTO together to help a technician. Helen had learned that her coworker would not receive a US$1000 bonus due to a technical error in accounting. Recognizing that her coworker was “a single mom with two kids … who depended on that paycheck,” Helen and her peers investigated possible solutions and resources. After exploring several options at her organization, Helen explained:

We found out that employees can gift her from their PTO … So we planned to give her maybe an hour each. Of course, nobody’s going to give [her] a thousand dollars all of sudden, but at least, an hour or two from the three of us…we could give it to her.

Although each worker in Helen’s example could only donate one or 2 hours, their collective efforts resulted in a greater amount of PTO. Similarly, participants described other feats that were more sustainable when a group worked together, such as meal trains. Meals trains are a coordinated effort where people sign up to provide meals to a colleague and their family, often over an extended period of time.

Elaina, a registered nurse, described how her coworkers came together to support a colleague who had just given birth and was soon after diagnosed with cancer:

After she had the baby, she found out it was breast cancer. And literally everyone on our unit signed up to babysit her kids and send her food. We had a meal train going for months for her, while she was getting chemo … That was really incredible to see: everyone just kind of came together to support her.

Elaina and other participants noted that meal trains were often suggested by one person, but there was not always one decision-maker. Collective duties were usually divided among peers. For example, Shelby, a registered nurse, described how an informal committee was formed to sell bracelets and fundraise for a coworker with cancer: “There were a couple girls that just kind of took charge. One of them kind of took over the food drive part of it, and the other one got the bracelet organized and went around … selling those.”

Shelby’s committee was initially formed outside of their organization’s jurisdiction, but they later announced their efforts to management and received permission to solicit donations in the workplace. Perhaps as a result of seeing compassion in action, participants were confident that their coworkers would likely help them if it ever became necessary. As Annie, a family nurse practitioner, explained: “I think it’s give and take too … If I can, I’m going to help you out because I truly feel that what goes around comes around. And if you’re going to be helpful to them, when you need help, they help.”

Together, these findings have significant implications about how compassion is both enabled and constrained in workers’ (re)production of structures in healthcare. Compassion is not merely an individual process of recognizing, relating, and (re)acting to pain. The meaning and practice of compassion is collectively constructed (and challenged) in communication.

The Structuring of Collective Compassion

First, this study offers insight into compassion as a communicative and collective process. At the individual level, compassion is an effortful form of work (Frost et al., 2006), but compassion does not need to be defined as an individual responsibility. Compassion is inherently interactional (Way & Tracy, 2012) and, as shown in this study, compassion can be a collaborative endeavor. We define collective compassion as a collective capacity where organizational members collaborate to recognize, relate, and (re)act to suffering in the organization. More specifically, collective compassion involves a co-constructed process where multiple people influence how a problem is interpreted, how empathy is generated, and how (re)actions are formed. Practically, collective compassion means that the labor of compassion can be shared among many. For example, in this study, healthcare workers collaboratively organized a meal train for a coworker with cancer. This effort would have required, at minimum: (a) information-sharing to increase recognition of a coworker’s issue, (b) collective reflecting on an emotionally charged topic, (c) debating potential actions that could alleviate the pain, and (d) coordinating and executing a specific plan of compassion over an extended period. At every stage of the compassion process, multiple members coordinate to scale up dyadic compassion to collective compassion.

Given the inextricable tie between structure and agency, collective compassion necessarily shapes and is shaped by the (re)production and/or transformation of structures related to the communication of compassion in healthcare. Existing structures that discipline expressions of pain, empathy, and professionalism affect how healthcare workers co-construct meaning around compassion at work. Agentic expressions of compassion and possibilities for collaboration are inherently linked to structural constraints in healthcare, such that the influence of managed care, bureaucracy, and capitalism on compassion cannot be ignored. Prior research has indicated that the dominant value sets of healthcare are regularly negotiated in communication (Harter & Kirby, 2004; Olufowote, 2008), and this study highlights collective compassion as a site where such negotiations occur.

The nature of collective compassion is arguably a symptom of “the contradictory nature of the capitalist state … expressed in the push and pull between commodification, de-commodification, and re-commodification” (Giddens, 1984, p. 315). Under managed care, compassion has an ambiguous (fiscal) value, which can cause confusion among employees. Organizational powerholders sometimes value and commodify compassion if the target is a patient and if the worker serves a more relational role (e.g., nurses) (Nicotera & Clinkscales, 2010). Generally, however, collective compassion among coworkers is not valued in a system that privileges rationality and efficiency. Therefore, it is within reason that healthcare workers often collaborate to manage the emotional weight, time, and professional risks of compassion. By coordinating compassion as a collective, healthcare workers can find creative and meaningful ways to support one another, which can simultaneously sustain and stretch existing structures.

The (Re)structuring of Suffering

Second, by employing structuration theory as a lens to study compassion, our findings highlight the duality of structure such that participants’ agentic acts of compassion simultaneously created compassion and (re)enforced current structures often related to the capitalistic managed healthcare models (Bisel & Zanin, 2016). For example, “gifting” personal PTO to a suffering coworker is indeed an opportunity for employees to (re)act to suffering. However, when workers adhere to a PTO donation policy—rather than questioning the framing of PTO as a scarce and fixed organizational resource—this adherence habituates current problematic structures. Ironically, a “compassionate” PTO donation policy structures further worker suffering by obligating altruistic co-workers, who are likely also suffering, to sacrifice their limited PTO to a coworker.

Similar to Nicotera and Clinkscales’ (2003; 2010) extension of structuration theory, structurational divergence theory (SDT), these findings demonstrate how divergent structures become entrenched, resulting in conflict, burnout, and employee turnover (Ford et al., 2022; Nicotera, 2015; Nicotera et al., 2015; Nicotera & Mahon, 2013). For example, the creation of “spiral time” as an organizational norm is an example of employees collectively and agentically recognizing emotional suffering. However, constraining and monitoring the time allowed to coworkers to feel emotion in response to work, also reinforces capitalistic orientations to time in healthcare (e.g., “time is money; ” Harter & Kirby, 2004, p. 56). This norm creates sanctions for workers who need to express emotion and suffering outside of their allotted “spiral time.” While employees may believe they are agents of change by resisting structures of emotional stoicism, in reality they often become further “suspended in webs of significance” that they themselves have spun (Geertz, 1973, p. 5; Nicotera, 2015).

Importantly, this study offers an alternative to the implementation of “compassionate” organizational policies (Carmack, 2010, 2017; Olufowote, 2008), which often reinforce individual, formalized acts of compassion rather than collective, improvisational acts of compassion towards reducing suffering. Similar studies on “compassionate” organizational policies, such as medical mistake disclosure, informed consent policies, and employee leave policies (Carmack, 2010, 2017; Kirby & Krone, 2004; Olufowote, 2008), have demonstrated that these policies appear to create structures that allow for emotion sharing among employees and patients. However, in practice, these structures constrain candid emotional expression and mainly function to benefit the organization though legal protections.

As Giddens (1984) explained, constraints, such as compassionate organizational policies, do not “push” an individual to do something if the individual has not already been “pulled” to act purposefully. In short, people are already pulled to act compassionately within everyday work. Therefore, compassion organizing should focus on reducing structural constraints to improvisational acts of compassion rather than creating “compassionate” policies. We argue that improvisational acts of compassion, both individual and collective, can be cultivated. A key theoretical contribution to structuration theory and SDT is the focus on “improvisation” as worker agency to change divergent structures, which often strip workers of agency (Nicotera, 2015). In the current study, earnest script breaking (e.g., asking “how are you twice”) offers a small example of an improvisational act of compassion. However, this improvisation toward recognizing suffering would not have been possible if employees did not have the backstage time and space to interact, which is often the case for overscheduled workers and understaffed healthcare organizations in managed care. By reducing structural constraints to collective and improvisational compassion, employees have greater opportunity to challenge existing structures and cultivate compassion in everyday talk.

Practical Recommendations for Cultivating Improvisational Compassion

Burnout is a pervasive issue in healthcare that poses consequences to both healthcare workers and patients (Carayon et al., 2019). Moreover, burnout financially hurts organizations over time due to costs of turnover and job withdrawal (Maslach et al., 2001). As such, it behooves organizational powerholders to pursue structural changes that foster collective and improvisational compassion. Currently, healthcare workers’ time and energy are monopolized by their schedules, leaving little chance for workers to build meaningful relationships and systems of support with coworkers. Healthcare leaders may not be able to radically change existing schedules to stay financially solvent in a capitalist society (Harrill & Melon, 2021), but strategic scheduling of PTO may offer some reprieve for exhausted workers. Additionally, scheduled breaks during the day (beyond mealtimes) may encourage healthcare workers to rest and visit with one another, which may become a stage for emotion-sharing and collective compassion. Given that surviving workloads is a major concern for healthcare workers, structural change toward this end may allow workers to redirect their efforts of collective compassion to other sources of burnout.

Just as a flexible schedule communicates the importance of rest, easy access to mental health resources would communicate care for healthcare workers’ wellbeing. By promoting mental health services for employees—ideally, at free or reduced costs—healthcare organizations can move away from norms of emotional suppression that stunt the compassion process. To accomplish such goals within budget, healthcare leaders may need to reevaluate which existing resources are helpful and which are merely cosmetic. It makes better fiscal sense to invest in benefits that healthcare workers can readily use at any time regardless of their schedule constraints (e.g., access to local gyms, meditation rooms, workshops related to mental health). Working to destigmatize mental health may also create a greater sense of freedom around collective compassion, especially as healthcare workers recognize and co-construct meanings of suffering and how to (re)act.

Finally, leaders in healthcare organizations should be encouraged to model compassion and emotion-sharing. The findings of this study indicate that healthcare workers often regulate each other’s emotions (e.g., spiral time), which feeds a recurring cycle of norms that mute emotional expression. However, healthcare workers’ interactions can also be promising pathways for modeling compassion. Seeing an organizational leader engage in earnest script-breaking and acts of collective compassion would positively sanction such activity. Given the lack of emotional training and preparedness in healthcare (Iannarino, 2023), healthcare workers learn what it means to be compassionate by observing key role models and replicating their behaviors. Prior research indicates that compassion can be communicated through embodied aboutness, or “making one’s body about the other” through presence, immediacy, and acts of service (Huffman, 2017, p. 159). As such, leaders can demonstrate compassion through their presence, active listening, and empathetic responses to suffering. In doing so, healthcare workers are more likely to find and mitigate sources of stress before they evolve into burnout.

Limitations and Future Directions

The limitations of this study offer directions for future research. First, this study utilized a maximum variation sample to seek out a wide range of perspectives on compassion in healthcare. Although this sampling technique provides a variety of intersectional perspectives, the findings presented in this study likely differ in other cultural and professional contexts. Future research should investigate how compassion differs based upon the healthcare discipline (e.g., chaplains versus nurses) and their organizations (e.g., a faith-based care center versus a city hospital). Additionally, this study was conducted during COVID-19, as participants were exposed to extreme situations and higher rates of patient death (Morgantini et al., 2020). The pandemic certainly influenced participants’ interpretations of their organization, sensitizing them to certain issues that may or may not remain in their workplace post-pandemic. To address this possibility, the interview guide contained multiple questions directly inquiring about changes that have emerged because of COVID-19. Future research should continue to explore the nature of transitions and challenges related to COVID-19 in healthcare.

Finally, this study’s sample was largely White (66.67%) and female (77.78%). The nature of this sample may be linked to the use of snowball sampling, especially if participants were recommending colleagues who were demographically similar to them. Certain identities (e.g., racial minorities, men) are underrepresented in this sample. Future research should diversify the voices represented in investigations of compassion, especially because “compassionate” policies may be used as a mode of power against marginalized groups (Simpson et al., 2014). Male participants can be difficult to recruit for studies related to delicate topics, especially if the content of interviews could lead to negative perceptions of participants (Tracy & Rivera, 2010). For this study, it is possible that potential male interviewees were cognizant of societal expectations about the appropriateness of emotion-sharing for men and women. Future research should consider the ways in which compassion might be a gendered expectation in healthcare settings.