The interplay between teamwork, clinicians’ emotional exhaustion, and clinician-rated patient safety: a longitudinal study

The study was conducted in intensive care units (ICUs) across all language regions in Switzerland. To investigate causal effects, longitudinal data were collected from medical and nursing staff using an online survey that included three assessments at 3-month intervals. We contacted nursing and medical leaders of each unit, informing them about the purpose of the study and asking them to discuss participation within their teams. Unit leaders of participating units were sent a link to the survey and forwarded it to their team.

Out of the 81 ICUs in Switzerland, 55 chose to participate (67 %). The ICUs comprised all specialties (medical, surgical, pediatric, and mixed) and were distributed across 48 hospitals. Participants were 2100 nurses and physicians (month 1: n = 1460; month 4: n = 1007; month 7: n = 807; n = 493 clinicians participated at all three measurement occasions). Nurses comprised registered nurses (RNs), registered nurses with specialization in intensive care, registered nurses in training for intensive care specialization, and nurse unit managers. Physicians specialized in intensive care, anesthesia, surgery, pediatrics, internal medicine, or trauma medicine and included resident physicians, senior physicians, and head physicians (see Table 1 for descriptive statistics).

To investigate the potential impact of attrition, differences on study variables were tested between participants who completed the month 7 assessment and participants who dropped out of the study before month 7 (see Table 2). Participants who completed the month 7 assessment were slightly older, had more professional experience and longer tenure. They also reported lower emotional exhaustion. This phenomenon is commonly found in longitudinal studies assessing aspects of work stress and interpreted as dropout of highly stressed individuals. We found no significant group differences for the other variables included in the statistical model. Thus, it is unlikely that changes in the sample structure influenced study results – if anything, the effects of emotional exhaustion might have been more pronounced.

Reliability statistics for all measures at all measurement occasions are reported in Table 3.

Figure 2 shows the significant cross-lagged and reverse cross-lagged effects of the structural equation model. Analyses revealed that cognitive-behavioral teamwork, interpersonal teamwork, emotional exhaustion, and clinician-rated patient safety were interrelated. Cognitive-behavioral teamwork (β = .17, p = .03), but not interpersonal teamwork (β = .03, p = .30) predicted an increase in clinician-rated patient safety (see Table 4). In turn, clinicians’ safety perceptions predicted an increase in cognitive-behavioral teamwork (β = .08, p = .03).

Moreover, there was a reciprocal lagged relationship of cognitive-behavioral teamwork on interpersonal teamwork (β = .13, p = .03) and vice versa (β = .09, p = .03). Thus, cognitive-behavioral teamwork predicts an improvement in interpersonal teamwork and interpersonal teamwork predicts an improvement in cognitive-behavioral teamwork.

Cognitive-behavioral teamwork (β = -.01, p = .02) and interpersonal teamwork (β = -.03, p = .02) had no effect on later emotional exhaustion. However, emotional exhaustion predicted a deterioration of the quality of interpersonal teamwork (β = -.07, p = .02). In addition, there was a tendency for emotional exhaustion to predict a decrease in clinician-rated patient safety (β = -.05, p = .09).

In general, physicians reported better cognitive-behavioral teamwork (β = .08, p = .02), interpersonal teamwork (β = .15, p = .02), clinician-rated patient safety (β = .06, p = .02), and higher emotional exhaustion (β = .04, p = .01) than nurses.

Based on the job demands-resources model, we assumed that effective teamwork may act as a job resource (i.e., aspects of a job that can buffer work demands, achieve work goals, and foster employee well-being) that positively affects patient safety [37].

According to the conservation of resources theory’s (COR) gain cycles, individuals or teams who initially possess plenty of resources are more likely to gain additional resources [38]. Positive interpersonal teamwork may act as a resource fostering the exchange of safety-relevant information between professions, thus increasing patient safety. Similarly, positive cognitive-behavioral teamwork may act as a resource that leads to procedures being carried out more smoothly, fewer errors, and thus higher patient safety. Effective cognitive-behavioral teamwork may be fostered by a good safety climate, which has been shown to be associated with fewer self-reported errors [39]. Teamwork as a resource may also buffer the impact of daily stressors and thus prevent clinician emotional exhaustion.

Lastly, we expected clinician emotional exhaustion to have a negative effect on patient safety. Emotional exhaustion can develop in individuals whose resources are insufficient to meet the cognitive, emotional, or physical demands of their job [39, 40]. Emotionally exhausted clinicians are less vigilant; their motivation to exhibit safe work practices may decrease; and thus errors are more likely to occur [41, 42].

Our results showed that cognitive-behavioral teamwork had a positive effect on clinician-rated patient safety, but it was not related to clinician emotional exhaustion. Tangible cognitive-behavioral team processes are instrumental to complete patient care-related tasks, such as pooling collective expertise to solve problems. These processes contribute to higher patient safety, and clinicians may be required to participate in such team processes despite being emotionally exhausted. However, they are not required to invest resources into interpersonal relationships, which might explain the finding that interpersonal teamwork suffered as a consequence of high clinician emotional exhaustion.

Interpersonal teamwork, a global evaluation of the quality of interprofessional collaboration, had no immediate impact on clinician-rated patient safety. It did, however, facilitate cognitive-behavioral teamwork, which in turn positively affected clinician-rated patient safety.

The results highlight the importance of longitudinal studies that go beyond mere cross-sectional associations between teamwork, patient safety and emotional exhaustion, but actually test assumptions concerning causal directions between these constructs.

Furthermore, our findings demonstrate that the interpersonal and cognitive-behavioral dimensions of teamwork are mutually dependent: On the one hand, better teamwork between professions facilitates cognitive-behavioral teamwork, such as coordination, communication, and cognitive functioning. This is in line with previous work suggesting that interpersonal teamwork forms the foundation on which cognitive-behavioral teamwork components are executed [2]. Trust and mutual respect foster a positive team climate that encourages individuals to contribute their expertise to the common goal, to speak up and voice their concerns in situations where they might deviate from the majority, or to report errors [42]. On the other hand, effective cognitive-behavioral teamwork improves the interpersonal quality of teamwork between professions.

Our analyses showed that clinician emotional exhaustion and patient safety do not evolve independently. First, our results suggest that emotional exhaustion may have a direct effect on clinician-rated patient safety. These findings complement a previous cross-sectional study conducted in intensive care units showing that emotional exhaustion was not only related to clinician-rated patient safety, but also to standardized mortality ratios [43].

Second, our results imply that clinician emotional exhaustion and clinician-rated patient safety are connected via the reciprocal relationships between interpersonal teamwork and safety organizing: clinicians with low emotional exhaustion may possess more resources to invest in interprofessional relationships. These, in turn, facilitate cognitive-behavioral teamwork and vice versa. Finally, cognitive-behavioral teamwork contributes to higher clinician-rated patient safety.

Finally, our results highlight the importance of including multiple professions when investigating teamwork. We confirmed that nurses’ and physicians’ ratings of teamwork, emotional exhaustion, and clinician-rated patient safety differ. Previous survey studies that investigated relationships between teamwork and clinician emotional exhaustion or patient safety rarely included multiple professions or explicitly addressed interprofessional teamwork, particularly in intensive care settings [44]. However, interprofessionalism is a defining feature of teams. Our study shows that even in highly technical environments, such as intensive care, the quality of interprofessional collaboration is an important aspect of teamwork that complements the cognitive-behavioral dimension.

The results of this study should be interpreted with some limitations in mind. We focused on the emotional exhaustion subscale from the Maslach Burnout Inventory because the sample size on the unit level did not allow for testing a more complex model that included the depersonalization and reduced personal accomplishment subscales [24]. Nevertheless, we believe that our results are representative and reliable. Emotional exhaustion is the core dimension of burnout and the most reliable and valid subscale across languages and cultures [45]. Moreover, 55 out of 81 Swiss ICUs and a total of 2100 clinicians constitute a high participation rate and large sample size at the individual level.

For economic reasons, patient safety was measured with a single-item indicator that assessed clinicians’ perceptions of overall safety in their unit and may therefore be less reliable than detailed surveys or objective indicators. However, previous research has shown that subjective safety ratings are indicative of objective patient safety such as standardized mortality ratios, as subjective and objective safety measures partly overlap [46]. In addition, our data show a high level of agreement regarding patient safety between team members. This illustrates that safety perceptions are a unit attribute and not only an individual rating impacted by emotional exhaustion and associated negative cognitions. In this study, we focused on developing a general understanding of the causal relationships between two aspects of teamwork, emotional exhaustion, and patient safety. Future research might include a multi multifaceted faceted conceptualization of patient safety into this model.

Interpersonal and cognitive-behavioral aspects of teamwork build upon one another and are thus both important for effective team functioning. Even in high-technology environments such as the ICU setting, good interpersonal relationships can facilitate cognitive-behavioral teamwork. Thus, interventions targeting teamwork should be designed with both teamwork aspects in mind, as such interventions carry the potential to reinforce each other: inclusion of the entire, multi-professional team; focusing on similarities and shared goals; building of shared mental models; and improving communication and coordination.

Observational studies in intensive care settings have shown the significance of cognitive-behavioral teamwork for immediate team performance outcomes [2]. Our study complements these findings by highlighting long-term effects. Long-term investment in teamwork is likely to build routine on which team members can rely in stressful situations. Previous research has shown that burnout (including emotional exhaustion) can spread from one intensive care clinician to another [47]. It is thus important to prevent the development of clinician emotional exhaustion before it becomes a problem for the entire team, as emotionally exhausted clinicians are less likely to possess the resources to engage in or benefit from team trainings.

Ethics permission was granted from the university and the cantonal ethics committees (75, 2013-06-03; 024/13-CER-FR, 2013-24-06). Written consent to participate was obtained per unit from the unit leaders. Upon accessing the online survey, participants were asked for their consent to participate, and assured complete anonymity and confidential handling of their data.