Factors behind burnout increase in medical students. Are the criteria so important?

Introduction: Medical students’ mental health has been a concern to the scientific community, especially as a result of the epidemic of mental comorbidities that have become commonplace among the various social groups in modern society. Objectives: To evaluate the prevalence of Burnout among firstto fourthyear medical students and compare different criteria to define the syndrome. Methods: A total of 511 students from three universities in Brazil answered validated instruments to assess burnout (The Maslach Burnout Inventory) and quality of life (WHOQOL-BREF), and a questionnaire prepared by the authors to assess sociodemographic data Results: There was a prevalence of 31.1% of three-dimensional burnout, 37% of two-dimensional burnout and 44,8% of one-dimensional burnout. There were worse levels of emotional exhaustion among students with two-dimensional burnout, compared to those with only one-dimensional criterion, and worse levels of depersonalization among students with three-dimensional burnout, compared to those with two-dimensional criterion only. The same assessed variables showed correlation with all three criteria. Considering the four domains of quality of life – psychological, physical, environment and social relationships, the psychological and physical domains were the ones that correlated the most to all three aspects of the threedimensional criteria. Emotional exhaustion was the dimension with the strongest association with three of the four quality of life domains. Conclusions: We question whether the three-dimensional criterion should really be the gold standard to define Burnout.


INTRODUCTION
Medical students' mental health has been a concern to the scientific community, especially as a result of the epidemic of mental comorbidities that have become commonplace among the various social groups in modern society 1 . In addition to medical school-related burdens, many demands and responsibilities are placed on these students, since this profession is focused on patient care, and it has minimal tolerance for errors 2 .
The education period is a complex and laborious process, characterized by several stressors such as high school performance requirements, reduced leisure time, high study load, high amount of information to assimilate, among others 3,4 . Thus, the development of nonpsychotic psychiatric disorderse.g., depression, anxiety, and burnout -are commonly found in this population 5,6 .
An important condition associated with the exhaustive nature of medical education is burnout, a concept introduced by psychologist Herbert Freudenberger in 1974. He defined burnout as "a progressive loss of idealism, energy, and purpose experienced by people in the helping professions as a result of the conditions of their work" 7 .
The burnout syndrome is associated with varying degrees of emotional exhaustion, depersonalization of the physician-patient process, loss of idealism and low sense of self-fulfillment 8 . Thus, to diagnose it, changes in dimensions related to emotional exhaustion, cynicism, and professional disbelief/low efficacy should be considered. Exhaustion, the main manifestation of burnout, is described as physical and emotional stress caused by high academic demand. The dimension of cynicism consists in the student's withdrawal from learning, which can be a form of self-defense. Finally, inefficiency reflects the student's feeling of unhappiness and dissatisfaction with their academic knowledge 9 .
The first instrument created for the assessment of burnout, and the most used to date, is the Maslach Burnout Inventory (MBI), initially developed to assess the syndrome in human service professionals using the three-dimensional criterion 9 . From the emergence of studies describing burnout in other areas, there have been variations of the MBI, such as the Maslach Burnout Inventory-Student Survey (MBI-SS), an adaptation made to assess the syndrome in students 10 , validated in Brazil by Carlotto 11 .
However, there are controversies in the literature regarding the diagnostic criteria. Some studies have chosen a two-dimensional approach, considering only emotional exhaustion and cynicism, such as the Oldenburg Burnout Inventory (OLBI), suggesting that ineffectiveness would be a personality pattern, rather than a component of the syndrome 12 .
In addition, other studies have chosen a one-dimensional criterion based only on emotional exhaustion, such as the Copenhagen Burnout Inventory (CBI), which follows a line of thought similar to that of OLBI: cynicism and low personal achievement would be different phenomena and should be analyzed and studied separately 13 .
This study aimed to investigate the prevalence of burnout among medical students from three universities in Brazil, evaluating and suggesting ideas about which diagnostic criteria of the syndrome can contribute to the screening of the condition and an earlier intervention.

Type of study, data collection, population, and sample
This is a multicenter, analytical, cross-sectional study conducted in three northeastern Brazilian universities. Data collection was performed using two scales and a questionnaire in students from the first to the fourth year of medical school. The study sample represents around 50% of the total of students from the three universities. The rest of the students were excluded for one of two reasons: the students either chose not to participate or they answered incorrectly (e.g. left blank answers).
During the data collection period, at least one of the researchers was a student in each university. We sent emails to the students explaining the research and, subsequently, they were approached in person, after or in-between classes, to fill out the questionnaires. The data were collected between September and October of 2018.
When the study results were analyzed and showed to be a matter of concern, the students could not be approached individually because of the anonymity, but the faculty members were informed about the results, so that institutional measures could be taken. The WHOQOL-BREF instrument is used to assess quality of life, and there have been studies that validated its use in medical students 14 . The instrument addresses four domains: physical, psychological, social relationships, and environment.

Data collection instruments
The score for each domain is obtained from an average score, according to the number of questions in the domain.

Data analysis
Categorical data were expressed as absolute counts with frequency and percentages and were compared by chisquare test. All quantitative variables were tested for normal distribution using the Kolmogorov-Smirnov test. Variables with normal distribution were presented as mean ± standard deviation and non-normal data were shown as median and interquartile range. Student's t-test, one-way analysis of variance (ANOVA) with Tukey post-test, or the Kruskal-Wallis test with Dunn's post-test were used to compare means/ medians of continuous variables accordingly to the distribution of data between the groups.
Correlations were evaluated by Pearson's correlation. Additionally, univariate logistic regression was used to assess the association between variables with the presence of burnout.
All analyses were performed using IBM SPSS Statistics for MAC OSX, version 23.0 (IBM, USA).

Ethical aspects
The project was approved by the Research Ethics Committee of Universidade Estadual do Ceará. Consent was requested from participants through a signed Informed Consent Form and anonymity was ensured.

RESULTS
A sample of 511 students attending the first to the eighth semesters of medical school participated in the research, 58% of which were males. Students in the basic cycle of medical school (first and second years) represented 56% of the sample, while the remainder were in the clinical cycle. About 90% of the sample was in the age group between 17 and 24 years. Table 1 shows the general sociodemographic data and according to the participants' gender. Regarding the students' daily hours of sleep, it was found that 54% slept 3 to 6 hours per night, while 46% claimed to have more than 6 hours of sleep. Forty students (8%) reported regular use of sleep medications, namely: alprazolam (11%), amitriptyline (5%), clonazepam (18%), zolpidem (16%), and others (50%).
The use of psychoactive substances was reported by 153 students (30%), and energy drinks showed a significant percentage (69%). Of the sample studied, 18% and 13% reported drug use and smoking, respectively. Of the smokers, 46% started smoking during medical school. Twenty percent of the smokers considered that attending medical school stimulated the smoking habit.
A total of 174 students (29%) stated they needed psychotherapeutic support at some point during medical school. The majority of students (90%) do not feel they receive enough emotional support from the faculty at their universities; among students with three-dimensional burnout, this number was 96.2%, with a statistically significant difference (p <0.05). Table 2 shows the general aspects of burnout. It was found that 37% of students had a diagnosis of burnout according to the two-dimensional criterion, 31% considering the three-dimensional criterion, and 44,8% considering the one-dimensional criterion. There was a higher level of cynicism/ depersonalization among the students with three-dimensional burnout compared to those with only the two-dimensional criterion, and a higher level of emotional exhaustion among students with only the two-dimensional burnout, compared to those with only the one-dimensional criterion. Table 3 shows the correlation of burnout (one-, twoand three-dimensional) with other parameters. A statistical correlation was found between burnout and age, and students under 25 years old were the least affected. The syndrome is also more often present in the students who sleep less than 6 hours per night, those who needed psychotherapeutic follow-up in medical school, and those who needed medications to regulate their sleeping patterns. There was no statistically significant correlation between burnout and drug abuse, harmful alcohol use (i.e., moderate risk or worse), smoking, psychoactive substance use and year of medical school the students were attending (basic or clinical cycle).

DISCUSSION
Considering the most traditional criterion adopted, the prevalence of 31.1% of three-dimensional burnout detected in the study population is high when compared to what was found in similar studies 15,16 . A Brazilian study conducted with students from the first to the sixth year and two studies conducted with students from the first to the fourth year of medical school -the   In the international literature, a systematic review of the syndrome in medical students showed that burnout is very prevalent during medical school 16 . Large multicenter studies found that at least half of medical students are affected during this period 17,18 . This higher prevalence of burnout compared to that of the present study could be explained mainly by the criterion adopted to diagnose the syndrome, since studies that use the one-dimensional or two-dimensional criteria show a higher prevalence. Considering the two-and one-dimensional criterion for our study, we found prevalence rates of 37% and 44.8%, respectively. Moreover, diverse cultural factors and individual characteristics of the students could also contribute to the development of the syndrome, as previously suggested 19-22 . Maslach (1993) has suggested that employees suffering from burnout first experience exhaustion, then depersonalization/ cynicism and reduced personal accomplishment. We compared three groups: the first one containing students who would fit only the one-dimensional criterion -those who only had emotion exhaustion, but not depersonalization or low academic efficacy; the second containing the students who would fit only the twodimensional criterion -those who did not have low academic efficacy but had emotional exhaustion and depersonalization -and the last containing those with three-dimensional burnout.
Comparing the first and the second groups, we observed that there was a higher level of emotional exhaustion in the second group.
Comparing the second and third groups, there was a higher level of cynicism in those with three-dimensional burnout. These data Considering that the same studied variables showed correlation with all the criteria -one-, two-and threedimensional burnout, as shown in table 3, we question whether the use of the three-dimensional criterion, which is more classically described, would only delay the detection of the syndrome, leading to diagnosis at more advanced stages of the process. In addition to that, even though these findings could reinforce the traditional theory related to the syndrome, there is an important discussion in the literature regarding the right criterion to define burnout. As previously mentioned, some authors suggest that a two-factor model that includes only emotional exhaustion and depersonalization might be more appropriate for some reasons. Firstly, because personal accomplishment is differentially related to other organizational outcomes (e.g., job satisfaction and organizational commitment). Secondly, because it is perhaps more appropriately conceptualized as a personality trait or a coping resource rather than a burnout component 12,23 .
The same logic has been applied to depersonalization. The depersonalization process was described by Maslach as a form of distancing oneself, which professional groups develop to cope with stress. Taking that into consideration, it has been suggested that depersonalization should be analyzed along with other coping strategies and it should not be necessary to define the syndrome 13 .
In addition to the concerns over the three-factor structure  (Table 4). This could reinforce that exhaustion related to burnout is not only emotional, but also physical and it should be more directly addressed.
We also observed that the emotional exhaustion aspect showed stronger correlations with three of the four quality of life domains when compared to depersonalization and low academic efficacy: physical, psychological and environment (Table 4). This could reinforce the theory that emotional exhaustion should be considered the most important, if not the only component to define burnout.
This study has limitations. Because this is a cross-sectional study, it is not possible to define a cause and effect relationship; only associations can be demonstrated. Moreover, with the use of self-administered instruments, there is the possibility of social desirability bias and answer distortion. Moreover, as four instruments were applied, it is possible that part of the students may have rushed into answering some questions not to prolong the time to complete the questionnaires, which may have impaired the information accuracy.

CONCLUSION
Although the three-dimensional criterion is the only one that addresses the three aspects of burnout present in the Maslach Burnout Inventory, we question whether it should really be the gold standard to define Burnout or if that definition needs to be revised. It is possible that there is a greater advantage in using the two-dimensional or onedimensional syndrome criteria, not only to attain a more correct and current concept of the Burnout, but also as a way to screen for the condition and intervene earlier, before the students develop harmful coping mechanisms. In addition, using a standardized criterion for the syndrome would allow standardizing its diagnosis between different studies, allowing a better comparison between different populations.