Burnout syndrome and resilience in medical students from a Brazilian public college in Salvador, Brazil

Abstract Introduction Burnout syndrome is highly prevalent among medical students. Whereas burnout syndrome has been associated with negative outcomes, like suicidal ideation, protective factors are still unknown. Objective To evaluate if there is an association between burnout syndrome and resilience in medical students, assessing covariates such as depressive symptoms, suicidal ideation, and religiosity. Method This cross-sectional study was carried out with a sample of 209 students from a medical school in Brazil. Burnout syndrome was assessed using the Maslach Burnout Inventory – Student Survey. Potential protective factors and aggravators to burnout syndrome were investigated using appropriate scales. Results Fifty-nine students (28.2%) presented burnout. Multivariate analysis showed that resilience was a protective factor (p < 0.001), along with being older, married or having better academic performance. Depressive symptoms were positively associated with burnout. Religiosity was not a protective factor and suicidal ideation was not associated with burnout when adjusted for depressive symptoms. Conclusion Burnout is frequent among medical students, impacting mental health and academic performance. Resilience seems to be a protective factor, and the relationship between burnout and suicidal ideation is possibly mediated by depressive symptoms. Prospective studies are needed to further investigate the associations found in this study.


Introduction
Burnout syndrome is a three-dimensional concept defined as a reaction to chronic emotional distress linked to work conditions. 1 Among health care students, it is characterized by: 1) weariness in relation to academic activities; 2) disbelief and detachment from studies and patients; and 3) feeling of incompetence. 2 Considering this definition, a systematic review with meta-analysis published in 2019 showed a 44.2% frequency of burnout in medical students around the world. 3 This number has grown in the last decade and is notably increased in medical residency programs, reaching 60%, and suggesting that doctors possibly carry a load of burnout symptoms since graduation. 4 A cross-sectional study showed that medical students, in general, have a higher frequency of burnout than students from other fields. 5 Another http://dx.doi.org/10.47626/2237-6089-2020-0187 study revealed strong associations between burnout and suicidal ideation (SI) in this population. 6 However, it was shown that newly enrolled medical students have lower rates of burnout than students from the same period in other areas. 7 This raises the hypothesis that, over the course of medical education, factors related to the academic context are added to individual predispositions, which together become responsible for the emergence of high levels of burnout throughout the medical program. 4 From a different standpoint, support from family, friends, and colleagues, as well as high levels of religiosity/spirituality have been inversely related to the frequency of burnout in medical students. 8,9 These findings are in line with what has already been proposed in the literature regarding the development of life-long resilience mechanisms, among which social support, including the relationship with family and community (e.g., active participation in a church, supportive friendships, marital relationship), is highly relevant among health protection factors. 10 Recently, a systematic review has analyzed studies conducted in 43 countries and showed that at least one in nine medical students (11.2%) has SI and about 27.2% have depressive symptoms (DS), 11 i.e., higher numbers than those described in the general population. 12 In fact, many associations have been established with SI in medical students, and it is possible to synthesize them into two main groups: 1) factors related to the individual, such as heredity, personality characteristics, psychiatric morbidities, 13,14 consumption of tobacco, alcohol, and other drugs 15 ; and 2) factors related to the academic context, such as dissatisfaction with academic performance, 16 perception of low performance, coming from other cities (being far from the family nucleus), 17 and having thoughts of abandoning the medical program. 18 Based on the above, we have hypothesized that: We used the three-dimensional criteria (high score for emotional exhaustion and cynicism + low score for professional efficacy) to define the presence of burnout.

Beck Scale for Suicide Ideation (BSS)
The BSS is a 21-item self-administered questionnaire developed to assess the presence and severity of SI in the general population. 20  rates of religiosity in each of the three subdomains.

Statistical analysis
The groups of students with and without burnout were compared in relation to the variables collected.
The descriptive analysis of the continuous variables was performed using median and interquartile range (IQR) or mean and standard deviation (SD), as appropriate.
Categorical variables were described as absolute and relative frequencies of the valid cases. The measure of association was expressed as prevalence ratio (PR) and 95% confidence interval (95%CI).
Multivariate analyses were conducted using a binary

Results
The protocol was sent to 534 students via email, of which 100 responded completely; it was also shared via WhatsApp ® in class groups, which yielded another 109 complete responses (total of 209 students). Of the total sample, 59 students (28.2%) presented burnout. The mean age of the total sample was 23.8 (±4.0) years; 57.9% were women, 54.1% were brown skinned, 79.4% were heterosexual, and 53.6% declared to have some religion (Table 1).

Relationship between burnout syndrome and resilience
Burnout was more frequent in students who showed    Table 3).

Multivariate analysis
Factors associated with burnout syndrome Table 4 shows the results of the binary logistic regression for SI, after adjustment for DS.
The multivariate analysis conducted using binary logistic regression and having burnout as dependent

Discussion
The present study found a prevalence of 28.2% of burnout in medical students at a federal university in the Northeast of Brazil, using the diagnostic criteria of MBI-SS. This result was lower than that found in a previously published meta-analysis (44.2%), 3  It is also worth mentioning that there is evidence that the frequency of burnout is higher in medical students than in students attending other undergraduate programs. 5 Also, even though the problems reported by all university students are similar, their intensity, as experienced by medical students, seems to be higher. 30 The relevance of this finding becomes more evident when taking into consideration the association between burnout and negative outcomes, such as less empathy, carelessness with the patient, medical errors and greater stigma in relation to mental health problems. 4 Confirming our initial hypothesis, scoring higher on the RS-25 was associated with a lower chance of having burnout. In fact, we observed that students who had low or moderate levels of resilience had a prevalence of burnout about four times higher than those with high levels of resilience, which demonstrates the protective factor of this variable in this population. In this sense, it is reasonable to conclude that students who have a greater ability to adapt to negative situations also tend to cope better with stress. However, in comparison with the general population, the literature points out that medical students have lower levels of resilience, 31 which would make them more vulnerable, possibly due to a lack of psychological resources to confront or limit the effect of stressors. 32 Thus, it becomes important to provide this population with opportunities to increase their levels resilience, so as to improve their quality of life, medical education, and, ultimately, patient care. 33  In relation to religiosity, although the students without burnout presented higher levels of religiosity, the potential protective effect of this variable was not maintained in the multivariate analysis. In our sample, such effect was more moderate when compared to the findings of another study, in which a different scale was used to measure religiosity; that study reported that religiosity was a protective factor against burnout in this population. 34 We also found that older students presented lower  (Table 4).
This confirms our initial hypothesis, namely, that the higher prevalence of burnout among students with SI may be due to the presence of DS.
In this sense, a multicenter American study involving 4,287 medical students showed that burnout could be an independent predictor for SI, 6 this being the worst outcome that burnout could have. However, there is published evidence of the association of SI with other variables, such as personality characteristics, 13 consumption of tobacco, alcohol and other drugs, 15 and psychiatric morbidities, 14 which can influence the association found in our study.
In our sample, the frequency of moderate to severe DS was 47.1%, much higher than the numbers found in a previous meta-analysis involving medical students (27.2%), 12 and was an independent predictor of burnout (OR = 2.03; 95%CI 1.42-2.92; p < 0.001). This result is ratified in the literature, which also shows that burnout is responsible for the maintenance of DS throughout the medical degree program. 39 Knowing that DS is one of the main predictor variables of SI among university students, 40 we researched whether this could be a confounding factor in the burnout × SI relationship. In fact, our findings suggest that much of the SI found among students with burnout is a result of the presence of DS, since this condition behaved as an independent predictor of SI (OR = 3.53; 95%CI 2.54-4.90; p < 0.001) ( Table 4).

Limitations
Our main limitations are: the cross-sectional study design, which does not allow to assess causality; the use of self-applicable scales, naturally with limited precision; having considered only the finished questionnaires and not the entire classes, generating possible selection bias, which may overestimate the frequencies found; and finally, the reduced size sample (n = 209) if compared to other Brazilian studies on burnout. [27][28][29] Moreover, because of differences in curriculum content and teaching methodology across the several medical schools available in the world, caution is needed when trying to extrapolate our results.

Strengths
To the authors' knowledge, this is the first study to evaluate the relationship between SI and burnout among medical students in Brazil. It is also one of the first to evaluate variables such as religiosity and resilience through scales, in the context of burnout.

Final considerations
Burnout is a frequent problem that has important consequences in terms of mental health and academic performance in the population assessed. Having greater resilience levels proved to be a protective factor against burnout. Burnout did not prove to be an independent predictor of SI when adjusted for DS in the multivariate analysis. Likewise, religiosity did not prove to be a protective factor. Therefore, it is suggested that interventions with the potential to increase resilience be prioritized by medical college institutions to reduce burnout among their students, in addition to providing reception services for students with DS.