As the role of a health care system’s influence on nurse burnout becomes better understood, an under-standing of the impact of a nurses’ work environment on burnout and well-being is also imperative.
To identify the key elements of a healthy work environment associated with burnout, secondary trauma, and compassion satisfaction, as well as the effect of burnout and the work environment on nurse turnover.
A total of 779 nurses in 24 critical care units at 13 hospitals completed a survey measuring burnout and quality of the work environment. Actual unit-level data for nurse turnover during a 5-month period were queried and compared with the survey results.
Among nurses in the sample, 61% experience moderate burnout. In models controlling for key nurse characteristics including age, level of education, and professional recognition, 3 key elements of the work environment emerged as significant predictors of burnout: staffing, meaningful recognition, and effective decision-making. The latter 2 elements also predicted more compassion satisfaction among critical care nurses. In line with previous research, these findings affirm that younger age is associated with more burnout and less compassion satisfaction.
Efforts are recommended on these 3 elements of the work environment (staffing, meaningful recognition, effective decision-making) as part of a holistic, systems-based approach to addressing burnout and well-being. Such efforts, in addition to supporting personal resilience-building activities, should be undertaken especially with younger members of the workforce in order to begin to address the crisis of burnout in health care.
Health systems strive to achieve the triple aim—an enhanced patient experience, population health, and lower costs—and to recognize clinician well-being as an unaddressed issue that can compromise patient safety.1,2 A clinician’s professional quality of life can be described as a balance between the compassion fatigue (negative aspects) and compassion satisfaction (positive aspects) they experience in their work.3 Compassion satisfaction is the joy one derives from their work, whereas compassion fatigue comprises secondary traumatic stress and burnout. Secondary traumatic stress results from witnessing or experiencing events associated with a trauma during patient care.3
Much of the focus on clinician well-being has emphasized the detrimental role of burnout, the emotional and physical exhaustion that comes from chronic exposure to stressors in the work environment.4,5 Burnout is associated with lower patient satisfaction and worse health outcomes (eg, higher mortality and more health care–associated infections).4 Nearly half of critical care professionals experience burnout.6 Critical care nurses, who provide care to high-acuity patients, have been a focus of research on the prevalence of burnout, the factors that cause it, and its potential consequences.5,7 Most studies have focused on prevalence rates and the mitigation of burnout; few have examined how aspects of a healthy work environment (HWE) relate to nurse burnout and its impact on nurse and organizational outcomes.
Burnout comes from the work environment.
Recent initiatives have been directed at under-standing how we perceive and address burnout. In 2016, the Critical Care Societies Collaborative released a call to action to address the harmful effects that burnout can have among health care providers, including the deleterious consequences for the patients for whom they care.5 The Collaborative’s call to action provides an important framework for guiding research into risk factors for burnout; this framework includes the examination of associations between personal factors, organizational features, and outcomes. Notably, the intensive care unit (ICU) work environment is called out as a factor contributing to burnout in nurses. In 2017, the National Academy of Medicine launched the Action Collaborative on Clinician Well-being and Resilience,8 an initiative to bring together experts, organizations, and resources to address burnout and well-being. Through these efforts, a distinct shift in thinking has occurred: we now understand that burnout is not an individual’s issue, but rather an organizational phenomenon resulting from more job demands, role strain, and excessive stress from poor work environments.2 One significant step is the World Health Organization’s reclassification of burnout from a medical condition to an occupational phenomenon.9 Although these initiatives have prompted awareness and calls to action, clinicians continue to report ongoing burnout and few responses to address the issue.10 A 2019 report outlined a clear recommendation that to decrease burnout, organizations must create positive work environments that foster well-being and support quality of care.2
Recognizing the importance of an HWE, in 2005 the American Association of Critical-Care Nurses (AACN) established 6 essential standards: (1) skilled communication, (2) true collaboration, (3) effective decision-making, (4) appropriate staffing, (5) meaningful recognition, and (6) authentic leadership.11 All are considered necessary and are linked to excellent nursing practice and patient care. After the introduction of these standards, reports of the overall health of critical care nurses’ work environments initially declined, but they have recently shown improvement.12–14 Research demonstrates a relationship between burnout and poorly rated work environments.15–17 Strong connections seem to exist between characteristics of nurses’ work environments and outcomes for both nurses (communication, teamwork, safety) and patients.18–22 Implications from research have not addressed specific aspects of the work environment that organizational leaders can prioritize for improvement. Small studies examining only a single facility or specialty have, however, begun to identify areas of focus. Nurses report that authentic leadership (managerial support) is associated with less compassion fatigue,23 and meaningful recognition (acknowledging a person’s valuable contribution) can serve as a powerful countervailing force that reduces compassion fatigue and promotes compassion satisfaction.24 In a pilot study using 1 hospital as the sample, we found associations between all HWE standards and burnout.25 Findings that help researchers examine the interplay between components of an HWE and other important individual and organizational factors may help provide a better understanding of which key elements of the work environment relate to professional quality of life outcomes. For leaders and staff to successfully improve clinician well-being and create positive work environments, they must prioritize crucial elements of the environment associated with burnout, secondary traumatic stress, and compassion satisfaction, as well as organizational outcomes such as turnover.
The purpose of this study was to examine the relationships among nurses’ perceived work environment and outcomes on the Professional Quality of Life (ProQOL) measure (ProQOL outcomes: burnout, secondary traumatic stress, and compassion satisfaction). In addition, we examined the relationship between nurses’ perceived work environment and ProQOL outcomes and their effects on unit-level nurse turnover.
We conducted a prospective survey of critical care nurses working at 13 hospitals within a single health system. We obtained institutional review board approval for the study. Survey directions included a statement that participation in the survey implied consent.
Sample and Setting
Nurses were recruited from 24 adult ICUs— medical, surgical, cardiac, trauma, and burn ICUs— in 13 hospitals within 1 health system. We invited all nurses who provided direct patient care and were employed full-time, part-time, or per diem to participate in the survey, giving us a potential sample size of 1206 nurses. Because we hypothesized that burnout and secondary traumatic stress experiences among nurses with a primary role outside of direct patient care may differ from those of the targeted group, we excluded the former.
To ensure that reports of individual nurses reliably represented the work environment of each ICU, we set a goal for a minimal response rate of 40%; this was consistent with response rates used in prior work.26 The projected sample size (n = 482) afforded a power of 0.80 to detect small associations (r2 > .03) between study variables in multivariable regression analyses. This calculation incorporates an anticipated mean unit response of 30 nurses and an estimated intraclass correlation of .04 for clustering by ICU unit.27
We used version 5 of the ProQOL instrument to measure 3 distinct subscales: burnout, secondary traumatic stress, and compassion satisfaction. Levels of these outcomes are described as low when scores are 22 or less; scores between 23 and 41 indicate a moderate level, and those greater than 41 indicate a high level.3 Researchers have used the ProQOL instrument extensively and found it to be reliable when used with the nursing population.28–30 Each of the 3 ProQOL subscales demonstrated good reliability in both previous studies3 and in the current sample (Cronbach α = 0.79 for burnout, 0.84 for secondary traumatic stress, and 0.89 for compassion satisfaction).
We used the 18-item AACN Healthy Work Environment Assessment (HWEA) to measure nurses’ ratings of their work environments according to AACN’s 6 standards, each assessed with 3 items. Aggregated mean scores are calculated for each subscale and standardized cutoffs are categorized as needs improvement (score 1.00-2.99), good (3.00-3.99), or excellent (4.00-5.00). Item composites showed good reliability of the HWEA in previous work31 and in the current sample (Cronbach α = 0.72- 0.81). Permission to use the instrument in the study was obtained through AACN.
We also asked respondents questions about their demographic and clinical background (Table 1); our analysis included age, highest degree obtained, and whether they had been recognized with a DAISY Award nomination. Our definition of turnover included turnover by both organization (ie, nurses who left the health system) and position (ie, nurses who left their position for another position within the health system).32
Data Collection and Analysis
We collected data during February and March 2018. An email invited nurses to access a unique website created for survey participation. Nurses completed the anonymous survey and then were given the option to receive a $20 gift card. Two registered nurse research assistants and 2 of the study investigators (L.A.K., K.L.J.) facilitated recruitment efforts; they joined rounds in the ICUs, attended staff meetings, gave presentations to leaders, and provided feedback on response rates throughout the data collection period. The survey was open for 3 weeks. We also queried turnover data from each ICU in the health system across 5 calendar months (February through June 2018).
We analyzed data using SPSS version 26 (IBM Corp.). We downloaded, checked, and cleaned the data. A few participants completed the survey more than 1 time, as evidenced by duplicate email entries for the incentive. We deleted duplicate responses.
We initially conducted descriptive analyses to assess relationships between predictors of nurses’ demographic characteristics and measures of HWEA components and ProQOL outcomes. Through an evaluation of the collinearity of predictors and the inclusion of variables in the model, we identified a set of covariates that we included in the final analyses: age, degree (diploma/associate’s degree versus bachelor’s degree and higher), and DAISY Award nomination.
In the primary analyses, we used generalized estimating equations to examine associations between HWEA subscales and ProQOL outcomes. Individual respondents were treated as being clustered within ICUs. We modeled each of the ProQOL outcomes (burnout, secondary traumatic stress, and compassion satisfaction) using the 6 HWEA subscales as predictors and adjusting for the back-ground covariates. To equilibrate variances across measures, we converted HWEA subscale scores to standardized scores (z scores). We selected an exchangeable working correlation matrix for the generalized estimating equations.
We analyzed separately the predictive association of ProQOL and HWEA measures with turnover rates. We used a negative binomial regression approach to predict ICU-level turnover (among 24 ICUs) from the estimated ICU-level means of burnout, secondary traumatic stress, compassion satisfaction, and each of the HWEA measures. That approach yielded incidence rate ratios (IRRs), which in this case characterize the incidence of turnover in terms of the percentage difference in turnover for every 1-unit change in the predictor (ICU-level mean score for each ProQOL outcome and HWEA measure). We used the number of turnovers during the 5-month reporting period as the outcome variable, and the natural logarithm of the number of nurses on the unit at baseline as the offset value. Because of the small number of ICUs, we estimated only single-predictor regression models.
The analytic sample included 779 surveys (a 64.6% response rate) with complete outcomes data. Demographic characteristics of the respondents are summarized in Table 1. On average, participants were around 40 years old and had worked almost a decade as a registered nurse, spending nearly half that time in their current ICU. Most respondents were women, held a bachelor’s degree, and worked full-time. Less than 20% worked an additional position outside their current unit.
Mean scores for burnout, secondary traumatic stress, and compassion satisfaction (from the Pro-QOL instrument) are listed in Table 2, as are mean scores for each HWEA subscale. Table 2 also shows nurses’ ratings of their work environment in relation to each subscale. Notably, more than half of nurses demonstrated moderate burnout with scores ranging from 23 to 41; however, no nurses reported scores higher than 41, the cutoff for “high” burnout.
Our models using generalized estimating equations, which evaluated the HWEA measures as predictors of ProQOL outcomes (see Table 3), showed that burnout was negatively related to 3 HWEA subscales (P < .05 for all 3 subscales): appropriate staffing, meaningful recognition, and effective decision-making. Compassion satisfaction was positively related to 2 HWEA subscales: meaningful recognition and effective decision-making. Secondary traumatic stress was negatively related to appropriate staffing.
We found age to be a significant predictor of all 3 ProQOL outcomes (Table 3). Older age predicted less burnout, less secondary traumatic stress, and more compassion satisfaction. Recognition with a DAISY Award nomination was positively associated with higher compassion satisfaction.
After we controlled for variance in the ProQOL outcomes attributable to age, educational degree, and recognition (DAISY Award nomination), the HWEA scales, in combination, accounted for 30.2% of the variability (r2 change) in burnout, 6.4% of the variability in secondary traumatic stress, and 23.6% of the variability in compassion satisfaction (all P < .001).
Our analysis of turnover initially showed strong correlations between ICU-level means and ProQOL outcomes and HWEA measures. In the negative binomial regression models, however, only the mean for skilled communication was significant (Table 4). Notably, the mean score for burnout approached significance. We calculated the IRR between the mean burnout and the turnover rate, and we found that for a given ICU, the turnover rate would be 1.22 times as large as, or 22% higher than, that of an ICU with a burnout score that is 1 point lower. The IRR for skilled communication, however, demonstrates that a given ICU’s turnover rate would be 0.09 times as large as, or 90% lower than, that of an ICU with mean skilled communication measuring 1 point higher. Table 4 shows the IRR values that suggest the relative risk of turnover to HWEA essential standards and ProQOL outcomes. Prediction of turnover was limited to ICU-level measures of predictors and outcomes, resulting in a small sample size for analysis.
Appropriate staffing, meaningful recognition, and effective decision-making have the most impact on burnout.
Our results demonstrate that the environment in which critical care nurses work is indeed a significant predictor of nurses’ well-being, with specific elements of that environment (appropriate staffing, meaningful recognition, effective decision-making) having the most impact on burnout. In addition, we provide evidence to demonstrate that in relation to the work environment and personal characteristics, burnout has a significant effect on nurse turnover. As calls to action make organizations aware of these issues,10 organizational leaders must understand where to focus their efforts to support nurse well-being.
Our results support the proposition that burnout is a workplace phenomenon, evidenced by the fact that appropriate staffing influenced ProQOL outcomes. The importance of appropriate staffing remains a significant and complex issue for nurses, health care organizations, and policy makers. Our findings further emphasize the need to prioritize appropriate staffing as a main element of a healthy work environment. In this study, we assessed staffing through 3 questions from the HWEA tool asking nurses whether their unit has (1) enough staff to maintain patient safety, (2) the right mix of staff to ensure optimal outcomes, and (3) adequate support services to prioritize patients and families.11 Collectively, the staffing questions negatively predicted burnout and secondary traumatic stress (ie, lower scores for appropriate staffing, higher scores for burnout and secondary traumatic stress), and positively predicted compassion satisfaction (ie, higher appropriate staffing, higher compassion satisfaction). To that end, a holistic approach to improve appropriate staffing should include an assessment of nurse well-being.
In addition to appropriate staffing, our study showed that effective decision-making was associated with less burnout and more compassion satisfaction. We must note that efforts to improve clinician well-being that are associated with effective decision-making are not aimed at individuals. The HWE standards from the AACN describe effective decision-making through the use of data, ensuring that the appropriate departments, professions, and groups are involved in important decisions, and that all perspectives are incorporated, including those of patients and their families. Organizations should enhance strategies to improve effective decision-making in their work environments; such strategies include shared governance, multidisciplinary rounds, creation of interprofessional educational opportunities, and improved family-centered care.33
As previous research has indicated, meaningful recognition is a powerful tool for reconnecting nurses to their work,34,35 and our findings support the value of emphasizing meaningful recognition. In addition to having formal reward and recognition programs, leaders must also understand what nurses value and use recognition strategies appropriately. Our study shows that DAISY Award nomination36 specifically stands out as a form of individual recognition that is associated with higher compassion satisfaction. The aspects of the DAISY Award program seem to carry significance in increasing compassion satisfaction, including the recognition that comes directly from the “nominator,” who is usually either a patient or a coworker, and the delivery of that recognition in a meaningful manner.
We also found that a nurse’s age is predictive of ProQOL outcomes. Whereas younger nurses are more likely to experience more burnout, older nurses experience more compassion satisfaction. This discrepancy continues to present a workforce-related challenge, because nurse turnover is an issue for new graduate nurses, as nurses often leave their positions and the profession within their first 3 years of practice.32,37 This suggests the need for a stronger emphasis on supporting nurse well-being earlier in a nurse’s career through strategies such as education, preceptorship, and new-graduate nurse experiences.
We identified a significant relationship between burnout and nurse turnover. Recent research has demonstrated that burnout is a predictor of nurses’ intentions to leave their current position, and this intention is moderated by the nurse’s job satisfaction.38 This finding is important in the context of the relationship between an unhealthy work environment and burnout. Although our results are limited by aggregated measures, they demonstrate the impact that improving burnout and unhealthy work environments can have on the risk of turnover. More research is needed—specifically studies using longitudinal data and linking individual-level outcomes— in order to evaluate the impact of burnout on workforce, quality, and financial outcomes.
Our study is not without limitations. Although the survey included objective measures of turnover, the anonymous survey data could not be linked to individual-level employment records. Accordingly, turnover data were aggregated to the ICU level for our analyses. Future analysis with matched survey data (eg, using an employee identification number to link individual survey responses with individual employee turnover) would enhance the outcomes analysis, as would additional individual-level outcomes influenced by burnout and the work environment. Funding to support incentives for survey completion, research assistant salary, and increased meeting attendance for study awareness is not usual for hospital-based research and may not be feasible for future studies. Despite the large number of hospitals and nurse respondents, our sample is limited to critical care nurses.
As research continues to generate evidence on burnout and well-being, organizations must address workforce issues in a strategic and comprehensive manner. Recommendations from the report “Taking Action Against Clinician Burnout: A System’s Approach to Professional Well-Being,” published in 2019 by the National Academy of Medicine,2 include creating positive work environments that reduce burnout and foster well-being. Our findings identify 3 key elements of the work environment—meaningful recognition, appropriate staffing, and effective decision-making— that have a statistically significant impact on burnout, secondary traumatic stress, and compassion satisfaction. In addition to ongoing valid measurement of burnout and well-being, we recommend prioritizing initiatives and interventions in these areas.
This research was performed within the Banner Health System. The authors gratefully acknowledge Jasmine Bhatti, MS, RN, (Arizona State University Edson College of Nursing and Health Innovation) and Lauren Leander, BSN, RN, CCRN, (Banner University Medical Center Phoenix) for their help with data collection during this research, We also thank Jill Evans, MS, RN, NEA-BC, (Banner University Medical Center Phoenix) for her advice about the research and her leadership expertise.
The research presented in this manuscript was funded by an American Association of Critical-Care Nurses 2017 Impact Grant (Drs Kelly, Johnson, and Todd).
For more about burnout, visit the AACN Advanced Critical Care website, www.aacnacconline.org, and read the article by Rushton and Pappas, “Systems to Address Burnout and Support Well-being: Implications for Intensive Care Unit Nurses” (Summer 2020).
To purchase electronic or print reprints, contact American Association of Critical-Care Nurses, 27071 Aliso Creek Road, Aliso Viejo, CA 92656. Phone, (800) 899-1712 or (949) 362-2050 (ext 532); fax, (949) 362-2049; email, firstname.lastname@example.org.