. 2020 Jul-Sep;32(3):458–467. doi: 10.5935/0103-507X.20200076

Table 3.

Findings retrieved from the included studies and deemed of interest, with particular focus on factors associated or correlated with burnout

Study	Variables found to be associated or correlated with burnout
Colville et al.⁽³⁴⁾	The analysis on the overall staff (physicians and nurses) showed that burnout significantly overlapped with post-traumatic stress disorder and anxiety. Multivariate analyses on variables correlated with burnout were performed according to two models. Resilience and being a doctor were the strongest predictors of reporting burnout. Attending a debriefing was correlated with halving the risk of burnout, whereas venting emotion and using alcohol were correlated with increased burnout reporting.
Lederer et al.⁽³⁵⁾	The authors also included four questions in the MBI that were related to consternation, which is defined as fear resulting from the awareness of being susceptible to psychological trauma. This element was not included in the burnout definition and thus did not alter the results on burnout. The authors did not find significant differences in burnout between the subgroups for age, gender, level of training, years of employment or family status. ICU personnel with fully established burnout planned to change professions more frequently than participants with no burnout.
See et al.⁽³⁶⁾	In the multivariate analysis of this Asian continental survey, protective factors against burnout for physicians were religiosity, years of experience in the current department, shift work and number of stay-at-home calls. The number of days worked per month was positively correlated with higher burnout.
Barbosa et al.⁽³⁷⁾	The study found that 50% of the participants who did not practice physical activity had high levels of EE. Several sources of stress were investigated, but their influence on burnout was not directly evaluated.
Fumis et al.⁽³⁸⁾	Moral distress (evaluated by the Moral Distress Scale-Revised questionnaire) correlated moderately with EE and weakly with PA (inversely) and DP.
Garcia et al.⁽³⁹⁾	The authors found higher burnout among pediatric intensivists than among general pediatricians. No other demographics or personal characteristics were associated with burnout in the univariate analysis.
Tironi et al.⁽⁴⁰⁾	Study including only physicians who were working in adult or pediatric/neonatal ICUs. Functional characteristics and occupational stress factors were reported, but their association with burnout was not analyzed. When considering high scores in all the three dimensions simultaneously, burnout was only observed in doctors working in adult ICUs (7.1%).
Embriaco et al.⁽⁴¹⁾	In this study, 50% of physicians with high levels of burnout wished to leave their job. The univariate analysis showed higher levels of burnout in females, in younger staff and in those not married and not having children. Burnout was also associated with withholding or withdrawing treatment, workload, and recent conflicts with nurses, families and colleagues. In the multivariate analysis, the factors remaining correlated with burnout were female gender, workload, and conflicts. Protective effects were good quality relationships with the chief nurse and nurses. The authors subsequently published other results of this study highlighting that, in the same cohort of physicians, depressive symptoms were correlated with high levels of burnout (Embriaco N et al. Annals of Intensive Care 2012).
Garrouste-Orgeas et al.⁽⁴²⁾	Burnout was correlated with depression scores, as evaluated by the Centre of Epidemiologic Studies Depression scale, but not with safety attitudes, as evaluated by the Safety Attitude Questionnaire - ICU version. The study focused on the association between burnout and medical errors in the ICU.
Malaquin et al.⁽⁴³⁾	Prevalence of burnout was not different between physicians and non-physicians or among the three different ICUs in this single center study. Severe burnout was more likely due to low PA than to high DP or EE. However, severe burnout was observed only in the cardiothoracic and vascular ICU (9%). After multivariate analysis, only the prevalence of depressive symptoms, low well-being and absence of a hobby were correlated with burnout.
Giannini et al.⁽⁴⁴⁾	The study evaluated burnout and other outcomes, such as anxiety, in nurses and physicians regarding the liberalization of visiting times in the ICU. Staff was surveyed at three time-points, and nurses always had a significantly greater predominance of high burnout levels. Staff with favorable opinions regarding liberalization had lower burnout levels. Burnout level increased during the surveyed period in both nurses and physicians.
Raggio et al.⁽⁴⁵⁾	The study evaluated the prediction of burnout according to results of the "profile of mood state" questionnaire that studies the profile of the state of mood in the previous week (58 specific sensations). Apart from the state of mood, the study showed a higher degree of DP in male physicians and a higher degree of EE in female physicians.
Shenoi et al.⁽⁴⁶⁾	Approximately two thirds of the investigated population of physicians recently considered leaving their job in the pediatric ICU. Burnout and severe burnout were significantly associated with willingness to leave the job (4 and over 9 times higher risk, respectively). Severe burnout was significantly associated with psychological distress (over 8 times higher risk). The correlation between the EE score and the psychological distress score was moderate to high, while it was low to moderate for DP and PA.
Ntantana et al.⁽⁴⁷⁾	In the overall study evaluating nurses and physicians, female sex was associated with higher EE and lower PA scores. Regarding EE, the multivariate analysis found a correlation with job satisfaction, satisfaction with end-of-life care, feelings of isolation after providing end-of-life care, neuroticism and extraversion traits.
Teixeira et al.⁽⁴⁸⁾	Physicians and nurses were included in this study, and the data were mostly reported as pooled outcomes. The only significant difference found was the lower scores for EE in physicians compared with nurses (17 vs 20, respectively). The authors subsequently published other results for the same cohort highlighting that nurses' burnout (and in particular EE) was associated with ethical decisions (withdrawing or withholding treatments, terminal sedation), while this was not the case for physicians (Teixeira C et al J Med Ethics 2013).
Merlani et al.⁽⁴⁹⁾	The study evaluated stress and burnout in a mixed population of physicians, nurses and nurse assistants in Swiss ICUs. The latter healthcare workers had significantly higher burnout (41%)than nurses (28%) and physicians (31%). The multivariate analysis in the overall population showed a higher risk of burnout according to individual factors (males, having no children, being younger than 40 years old), patients' related factors (higher ICU mortality), and organizational factors (working in German-speaking ICUs and having a lower proportion of females nurses). Moreover, a positive answer to the question about "Feeling stressed" was the predominant independent factor increasing burnout risk.
Galván et al.⁽⁵⁰⁾	The score in the PA domain was independent from the scores in the EE and DP domains, while the latter scores had significant associations between them. In the multivariate analysis, being certified as a pediatric ICU physician and working in a public practice was protective against burnout, while a higher workload (more than 36 hours/week as on-call duties).

MBI - Maslach Burnout Inventory; EE - emotional exhaustion; PA - personal accomplishment; DP - depersonalization; ICU - intensive care unit.