The relationship between the burnout syndrome and academic success of medical students: a cross-sectional study

This cross-sectional study included 760 full-time university medical students from the first to sixth year at the Kragujevac University Faculty of Medical Sciences of the invited 836 (90.9 % response rate), as described in our previous report (29). Their curriculum includes preclinical training (first three study years) and clinical training (study years 4–6).

The respondents completed the survey (on paper) anonymously, and participation was voluntary. They had about 15 (±5) min to fill out the survey and were supervised by a medical doctor specialised for burnout assessment. Students who did not cooperate and/or did not complete the survey were excluded from analysis.

The sample size was calculated using the Epi Info statistical calculator (Centers for Disease Control and Prevention, Atlanta, GA, USA) based on the population of 836, acceptable margin of error of 5 %, expected prevalence of 22.6 % (30), and confidence level of 99.99 %. The obtained sample size was corrected to account for non-response rate of 15 %, yielding the minimum sample of 538 participants.

This study is part of a research approved by the Ethics Committee of the Kragujevac University Faculty of Medical Sciences (Ref. No.: 01-1176). All participants provided a written informed consent prior to taking part.

The survey included an epidemiological questionnaire and the Maslach Burnout Inventory – Student Survey (MBI-SS) (2). The epidemiological questionnaire included questions about sociodemographic characteristics, such as age (years: ≤21 / 22–24 / ≥25), having children (No / Yes), overall GPA (≤8 / >8) on a 10-point scale (with passing score 6), study year (1–6), whether the study year was retaken (No / Yes), number of exams passed (dichotomised: ≤23 / >23), length of study (dichotomised: ≤6 year / >6 years), training (preclinical: first-three year of study; clinical: study year four through six), total time spent in classes per day (average number of hours of classes attended per day, dichotomised: ≤5 h / >5 h), total time spent studying per day (average number of study hours per day, dichotomised: ≤5 h / >5 h), study financing (state-sponsored / self-funded), housing model (own home / with parents / rent / student dormitory), habits (cigarette smoking, alcohol consumption: No / Yes), personal medical history of chronic disease (No / Yes), use of sedatives (No / Yes), and use of psychoactive substances (No / Yes). If the students regularly smoked at least one cigarette per day for one year they were considered smokers. If they smoked at least one cigarette every day over the past 12 months they were considered current smokers. Ex-smokers were those who had quit smoking for more than a year before taking the survey. All data are self-reported and were not verified.

The MBI-SS questionnaire is a version of the Maslach Burnout Inventory designed to measure the level of the burnout syndrome in students by Schaufeli et al. (2), which consists of 15 items/ statements in three subscales: emotional exhaustion (5 items), cynicism (4 items), and academic efficacy (6 items). Students were asked to score the accuracy of each statement in terms of frequency on a 7-point Likert-type scale ranging from 0 to 6 (0 – never; 1 – several times a year; 2 – once a month; 3 – several times monthly; 4 – once a week; 5 – several times a week; 6 – daily). Scores higher than 14 for emotional exhaustion, higher than 6 for cynicism, and lower than 23 for academic efficacy are considered indicative of high burnout risk (2, 31, 32).

After obtaining a license from the current owner, Mind Garden (Menlo Park, CA, USA) for the use of the MBI-SS questionnaire, the questionnaire was translated into Serbian, adapted, and validated. The three subscales (explaining the variance of 64.9 %) had satisfactory Cronbach’s α coefficients (0.869 for emotional exhaustion, 0.856 for cynicism, and 0.852 for academic efficacy) and good test-retest reliability (33).

All statistical analyses were run on the SPSS program version 19 (SPSS Inc., Chicago, IL, USA). Categorical variables are expressed in percentages and were compared with the chi-squared test to determine differences between medical students with a lower GPA and those with a higher GPA. Univariate and multivariate regression analysis were used to establish the relationship between the burnout syndrome and academic performance [odds ratio (OR) with 95 % confidence interval (95 % CI)].

The multivariate logistic regression model included all variables that univariate analyses determined as related to the burnout syndrome (P<0.10). In order to assess the effect of potential confounding variables, ORs were adjusted for all variables determined as related to the burnout syndrome by univariate analyses (P<0.10) and for other variables that had been reported to correlate with the burnout syndrome by literature (including age, study year, marital status, children, housing, completed secondary school, study financing, re-enrolment in the academic year, length of study, cigarette smoking, alcohol consumption, sports, recreational activity, and positive personal medical history of chronic diseases).

Training (preclinical and clinical) was not included in the multivariate model because of its collinearity with age and some of the study years. Model fit was assessed with the Hosmer-Lemeshow test of goodness of fit and Cox and Snell’s and Nagelkerke’s pseudo R-squared measures. The test for linear trend in risk was based on the logistic regression model. All statistical analyses were performed separately for medical students with a lower GPA (≤8) and with a higher GPA (>8). A value of P<0.05 was considered significant.

This study has several limitations. In addition to the known limitations of the cross-sectional study design, we did not determine the burnout risk before the students enrolled for the medical study. There is also the limitation of self-reporting, which cannot rule out information bias, especially in terms of self-reported GPA.

This study focused on personal variables rather than environmental factors (such as relationships with colleagues and teachers) and did not look into other potential predictors of the burnout syndrome (such as socioeconomic status, employment status of students, or parents’ occupation). In addition, although the response rate is high, there may still be some response bias, as students suffering from the burnout syndrome may not have been in class at the time of data collection or may have chosen not to participate.