Clinicians’ Perceptions of How Burnout Affects Their Work

Clinicians in this mixed-methods study were participating in a randomized controlled trial of a burnout intervention. For the parent study, clinicians were recruited from a Veterans Affairs (VA) medical center, a veteran-focused social services provider, or a community mental health center. Recruitment was open to anyone providing mental health services, regardless of current burnout level. Data for this study were collected from the baseline assessment, prior to any intervention, between January and June 2013; 123 participants were recruited and completed an online survey. Of these, 120 completed the open-ended questions and were included in this study. All procedures were approved by the Indiana University Purdue University Indianapolis Institutional Review Board and Roudebush VA Medical Center Research and Development Committee.

Most participants were white (N=92, 77%) and female (N=83, 69%), and the mean±SD age was 46.3±11.7. The educational background included 12 (10%) participants with less than a bachelor’s degree; 22 (18%) with a bachelor’s degree, 63 (53%) with a master’s degree, and 22 (18%) with a doctorate. The mean length of time in the current job was 4.6±5.7 years, and length in the mental health field was 14.5±10.7 years. Participants reported spending 75.9%±26.3% of their time in direct care.

Burnout was assessed with the Maslach Burnout Inventory for Human Services (9), a widely used measure of three components of burnout: emotional exhaustion, depersonalization, and personal accomplishment. The subscales have shown good internal consistency, stability over time, and convergent validity with related constructs (9).

Perceptions of the impact of burnout were assessed through the following open-ended questions: “How, if at all, do you think burnout affects the way you work with clients?” “How, if at all, do you think burnout affects the outcomes of the clients you work with and which outcomes?” “How, if at all, do you think burnout affects the way you work with colleagues?” “If you felt more energized and engaged in your work, how would your clinical work be different?” “If you felt more energized and engaged in your work, how would it affect the outcomes of the clients with whom you work?” “If you felt more energized and engaged in your work, how would it affect the way you work with colleagues?”

Responses to open-ended questions were collated into a spreadsheet for content analysis. Using emergent, consensus-based coding for conventional content analysis (10), the research team independently examined responses from approximately ten participants to create an initial list of categories. We met as a group to discuss, merge, and refine categories. We repeated this process until we agreed on a list of codes (three iterations). Pairs of raters independently coded all responses, determining whether each category was present in a particular participant’s answers and if so, whether the target of the impact was consumer, colleague, both, or other or unspecified. Raters met to review codes and reach consensus.

Using consensus codes, we examined the types and frequency of impacts that clinicians described. We then explored, using point-biserial correlations, whether the level of burnout was related to perceptions that burnout affected their work with consumers or colleagues or affected consumer outcomes.

More than half of clinicians (58%) described burnout as having a negative impact on the quality or amount of their work (Table 1). Additional negative impacts included withdrawing from others, having less patience and less energy, poor communication and listening skills, worse consumer outcomes, and poor attitudes of others (consumers and colleagues). Although our survey questions were focused on clinical work, a few participants spontaneously described how burnout was affecting their home life; for example, “I try not to allow the way I feel impact the veterans. I take it out on my family.” Surprisingly, a few clinicians mentioned positive aspects of burnout, for example, bringing staff closer together or serving a tempering function (for example, in response to the impact of being more engaged, one participant responded, “They would probably find [me] even more annoying as I would be intrusive in their time/space”).

Overall, 101 of 116 (87%) answered open-ended questions, with responses indicating that burnout has some type of impact on how clinicians work with consumers, and 104 of 112 (93%) described an impact on how they work with colleagues. In addition, 75 of 111 (68%) reported that burnout has an impact on consumer outcomes. However, this overall category included responses that we later categorized as processes rather than as outcomes. For example, a clinician responded to the question of whether burnout affects consumer outcomes with the following: “I hope my burnout is not affecting them, but we are having tremendous turnover right now, and I know that affects them. They get accustomed to working with someone, developing trust, then that person leaves.” This response reflects an impact on poor alliance and other relationships in our coding system, rather than on consumer outcome. Clinicians who did not think burnout affected their work often responded with “I don’t think it does,” but some elaborated; for example, “Ultimately my clients determine their own outcomes” or “I try very hard not to let it affect my interactions with clients.”

Overall impact responses were examined in relationship to level of burnout reported in the same survey. Burnout was not related to perceived impact on colleagues (correlations ranged from .11 to −.09). Higher depersonalization was associated with reporting an impact on how clinicians work with consumers (r=.21, p=.024). Clinicians with higher levels of emotional exhaustion were less likely to report an impact on consumer outcomes (r=−.24, p=.010).

Most clinicians described negative impacts of burnout on mental health services, particularly in interactions with colleagues (93%) and consumers (87%). Fewer, but still a sizable number, reported impacts on consumer outcomes (68%). These findings are consistent with empirical studies linking burnout to reduced consumer satisfaction (5), lower perceived quality of care (4), and worse consumer outcomes (6).

This study extends prior work, however, by pointing to a variety of different aspects of care that could be studied in relation to burnout. Clinicians’ ability to provide care with enthusiasm, patience, empathy, effective communication skills, attention, and creativity was influenced by burnout. Process indicators, such as therapeutic alliance and consumer engagement, appeared to be important elements of care negatively affected by burnout. All of these factors could conceivably lead to worse consumer outcomes and are worthy of future study.

Although most clinicians (68%) reported that burnout affects consumer outcomes, the coding system revealed fewer clinicians (45%) reporting specific consumer outcomes. That is, when asked about “outcomes,” clinicians sometimes responded with items that coders included as process variables instead. It may be that clinicians have a different view of outcomes, have a harder time conceptualizing outcomes as distinct from clinical processes, or perhaps intended to describe a process that might eventually lead to an outcome without specifying which outcome. This lack of differentiation may also reflect reality in that some constructs can be conceptualized as both processes and outcomes (similar to recovery [11]). When asking about outcomes in the future, it may be helpful to provide greater specification.

The exploration of level of burnout by seeking answers to open-ended questions led to some interesting findings. Clinicians with higher levels of depersonalization appeared to recognize the impact of burnout on how they work with consumers. This may reflect validity of the coding system rather than a unique finding. For example, the depersonalization subscale includes items such as “I feel I treat some recipients as if they [are] impersonal objects.” Interestingly, the most emotionally exhausted clinicians, however, were less likely to report that burnout affected consumer outcomes. Although these correlations represent small effect sizes (12), they are notable in the context of mixed-methods research, with less risk that common method variance will inflate correlations. Exhausted clinicians may not be able to see the impacts of burnout, perhaps because of reduced energy or attention or perhaps through self-protective denial. Alternatively, clinicians who are working the hardest to affect consumer outcomes or to keep burnout from affecting outcomes may become more emotionally exhausted. With few exceptions (6), burnout has not been studied in relation to consumer outcomes and appears to be a ripe area for future study.

Although this study provides greater depth of understanding the ways in which professional burnout may affect clinical care and outcomes, we also recognize its limitations. This study relied solely on self-reported impact, and although clinicians can provide useful insights, the field also needs studies linking burnout to actual consumer outcomes. It is possible that participants seeking an opportunity to address burnout may be more likely to endorse ways in which burnout could affect care. This could result in higher levels of perceived impact than might be found in a more general population of clinicians. Yet even in this sample of clinicians who were interested in reducing burnout, we note that almost a third of them reported that burnout would not affect client outcomes—and they were the clinicians with higher emotional exhaustion.

Clinicians perceived a variety of ways in which burnout may affect how they work with people and, ultimately, consumer outcomes. The level of burnout may also affect how clinicians view that care is being affected. Future research is warranted in linking clinician burnout with processes of care (such as the therapeutic alliance and consumer engagement) and consumer outcomes.

This study was supported by grant IIR 11-030 from the U.S. Department of Veterans Affairs Health Services Research and Development. The views expressed in this report are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs.

The authors report no financial relationships with commercial interests.