Prevalence of empathy, anxiety and depression, and their association with each other and with sex and intended specialty in medical students

Introduction: Empathy and mental health are crucial for medical students’ self-care and performance as well as for patient care. Objective: to assess the prevalence of empathy, anxiety and depression, and their association with each other and sex, intended specialty and course semester. Method: Cross-sectional study with 405 of 543 students (74.6%) from odd semesters and from the 12 th semester of the medical course of two universities in southern Brazil. Data were collected using a self-administered questionnaire containing information on age, sex, medical course semester, intended specialty, Jefferson Scale of Empathy (JSE) and Beck Anxiety and Depression Inventories (BAI and BDI). The data were analyzed using descriptive statistics, Student’s t-tests, Chi-square, and bidirectional ANOVA between groups. The level of significance was set at p < 0.05. Results: The mean JSE score was 120.2 (SD = 10.6) [116.9 (SD = 11.0) in men and 123.4 (SD = 9.2) in women, p = 0.000], being higher among students who wanted to follow medical areas aimed at people [123.1 (SD = 10.1)], than among those whose intended areas aimed at techniques and procedures [118.5 (SD = 11.2)], p = .003. There was no difference between the course periods. The anxiety and depression mean rates were, respectively, 16.2 (SD = 11.3) and 11.9 (SD = 9.0) [13.1 (SD = 10.3) and 9.9 (SD = 8.3) in men and 19.1 (SD = 11.4), and 13.8 (SD = 9.4) in women, p = .000 for both]. The prevalence rate of moderate and high anxiety was 33.8% and, when including mild anxiety, it was 59%. The prevalence rate of dysphoria (BDI = 16 20) and depression (BDI > 20) was 26.4%, and 11.9% for suicidal ideation. An association was observed between severe anxiety and the JSE subscale ‘Walking in patient’s shoes’, more related to empathic stress. Conclusions: Empathy is high and stable throughout the medical course at the studied institutions and higher in women and students who want to follow people-oriented specialties. Anxiety and depression have higher prevalence rates in women. Severe anxiety is associated with the JSE subscale ‘Walking in patient’s shoes’.


INTRODUCTION
Medical students' empathy and mental health influence their self-care, academic performance, and patient care. However, several factors can affect them before and during academic training 1,2 .
When choosing Medicine as a profession, many students are still transitioning from adolescence to adulthood, with different adaptation demands that can worsen their mental health. The preparation for admission at the medical course usually requires great dedication to studies due to the extremely competitive selection process. After entering the course, it is necessary to adapt to several aspects, such as: socializing with new classmates and teachers; new teaching and learning methods; full-time course; usually frequent and overlapping evaluations that require continuous study of excessive new contents; activities that can undermine one's emotional capacity, such as anatomy classes with cadavers. For some, being away from one's family, having to live alone or with new roommates, and financial and/or access to food limitations are also added 3 . In addition to fatigue, stress and anxiety, time limitation due to the pressures of the course usually interferes with opportunities for leisure and interaction with family and friends. Throughout the course, the students deal with the suffering of patients and their loved ones, with death, and receive messages about professional values and attitudes from people with whom they live, which are often ambiguous and can make them vulnerable and confused about their deepest values.
In the final semesters of the course, there is also progressive responsibility for patient care, shifts, doubts about which specialty they intend to follow and stress regarding the tests for medical residency or for a job that will be the first source of income after six years of training, which limit economic independence 1,2,4-9 .
The diagnosis of anxiety and depression is associated with the duration, frequency and intensity of manifestations of mood and affective disorders 10 . Dyrbye et al., in a systematic review, found a higher prevalence of depression and anxiety in medical students from Canada and the United States than in the general population 1 . Other studies have also observed this same phenomenon [11][12][13][14][15][16][17][18][19] and some have also found a higher risk of suicide among medical students [12][13][14][15][16][17][18][19] .
A meta-analysis on anxiety among medical students including 18 studies from Asia, 21 from the Middle East, 13 from Europe, 10 from South America, 4 from North America, two from Oceania and one from Africa, totaling 40,348 medical students, found an overall prevalence of 33.8% (95%CI 29.2 -38.7%), higher than that in the general population 18 .
A recent analysis of 10 systematic reviews on depression among medical students, including 249 articles written in People with subclinical depression also seem to have impaired social performance even after depression remission, which can influence the persistence or worsening of depression because they tend to avoid social interactions with those who could help them to overcome difficult times 26,27 . There is also the possibility that certain brain areas that are affected in depression may lead to difficulties in social communication and the ability to perceive and understand others, which is crucial for empathy 28 . A systematic review including 37 studies written in English analyzed the association between empathy in adults with a primary diagnosis of major or subclinical primary depression who had not sought help, and who had no other mental or somatic diseases. It was verified that people with depression had no alterations in empathic concern but tended to have greater emphatic stress and less cognitive empathy. The study by Schreiter et al. showed that depression rates, in addition to being higher among women, could result in a greater decrease in cognitive empathy for women than for men with depression; however, the authors call attention to limitations in the studies analyzed, as all of them were crosssectional, with variation in the population included in each study and in the instruments used to assess depression. Also, some used instruments were filled out by the participant, which made them more subject to bias 25 .
As empathy is a component of medical professionalism [29][30][31] and one of the pillars of the doctor-patient relationship, it has been extensively investigated [32][33][34][35] . Studies show that more empathetic physicians, with more patient-  16. This variation occurred not only between countries, but also between different institutions in the same country. Empathy was greater in women in 18 of 27 studies and among students with an intended specialty focused on people in three of nine studies 46 . According to Hojat, people-oriented specialties are those in which the doctor has frequent meetings with patients to monitor their health over a long period of time, which includes family doctors, clinicians, pediatricians, gynecologists/ obstetricians and psychiatrists. Technology or procedureoriented specialties encompass the other specialties, including clinical specialists (such as cardiologists or gastroenterologists) and anesthetists, surgeons, pathologists and radiologists 43 .
Therefore, given the importance of medical students' empathy and mental health, the following research questions were raised: How is empathy, anxiety and depression among medical students and how are they related? Is there a difference in these variables between students attending a private institution and another attending a public institution, with different curricula? What are the associations between empathy, anxiety, depression with age, sex and intended specialty by these students?
To answer these questions, the aim of this study was to assess the prevalence of empathy, anxiety and depression, their association with each other and with sex, intended specialty and course semester.
Our hypotheses were: 1. There is no difference between students attending courses in private and public institutions; 2.
The prevalence of anxiety is around 34% and of depression is 27% among the students; 3. Empathy, anxiety and depression are greater among women; 3. Empathy is greater among students who intend to pursue people-oriented specialties; 4.
Anxiety and depression are negatively correlated with empathy.

Participants
The participants were medical students from the 1 st , 3 rd , 5 th , 7 th , 9 th , 11 th and 12 th semesters regularly enrolled in the course in the second semester of 2018 at UFSC and UNOESC. Students attending the first semester and 11 th and 12 th semesters were included to allow comparisons between the initial and final semesters of the course.
UFSC has an integrated modular curriculum and a federal administrative link, hereafter called University A. UNOESC has a problem-based curriculum and its administrative link is a private one, and henceforth will be called University B.
Inclusion criteria: students present in the classroom on the day of data collection who agreed to participate in the study.
Exclusion criteria: not filling out more than 4 questions on the Jefferson Scale of Empathy or filling them out using the same number, and not filling out Beck's Anxiety Inventory (BAI) and the Beck's Depression Inventory (BDI).

Data collection
Data collection was carried out using a self-administered questionnaire containing participant identification data (the 7 middle numbers of the ITIN -Individual Taxpayer Identification Number-, university of origin), course semester, sex, age, intended specialty, empathy assessed by JSE, version validated for Brazilian Portuguese for students 38 , BAI and BDI.
The JSE is a self-completed instrument comprising twenty items, answered on a seven-point scale, where 1 is equivalent to "strongly disagree" and 7 to "strongly agree". Its The BDI I aims to identify and assess the severity of symptoms of depression. It consists of 21 items with statements, from which the respondent selects the one(s) that best describe how they felt in the previous week. If the participant checks more than one option, the highest one is considered. The score is calculated by the sum of the scores on the items. The recommended cutoff points for the classification of the Brazilian version of the BDI I 47 are: 0 to 11: minimal depression; 12 to 19: mild depression; 20 to 35: moderate depression; from 36 to 63: severe depression. Kendal et al. warn that the BDI should be used as a measure of depression syndrome, but that its scores alone are insufficient as an index of nosological depression, as they may reflect other diagnoses such as schizophrenia, anxiety disorders and substance abuse, and may be affected by aspects such as stressful life events. The authors suggest that other affective states be aggregated to depression research to identify which effects are specific to depression in relation to other mood conditions. They criticize its classification because, even when the total score is equal to zero, the respondent is classified as having minimal depression, instead of being identified as having no depression. They also claim that cutoff points for depression such as 10 or 16 can generate false positive diagnoses. They recommend that individuals with scores up to 9 and mild depression between 10 and 20 be considered without alterations but consider that people with scores up to 17 could be considered dysphoric (non-specific negative affectivity).
Scores above 17 would increase the probability of having depression, which should be considered moderate when scores are between 20 and 30, and severe when above 30 48 . The recommendation of a score limit above 20 for depression and below this cutoff point for dysphoria has been adopted in Brazil in studies with university 49 and medical students 50 . Gorenstein et al. adopted the BDI cutoff points of less than 16 for normality, 16 to 20 for dysphoria, and more than 20 for depression in the non-clinical subgroup in their study to assess the psychometric properties of the Brazilian version of the BDI 49 . Therefore, in this study, in addition to the classification recommended in the Brazilian version 47 , this classification will also be considered 49 .

Data analysis
The data were entered and analyzed using the software The accepted level of significance was set at p < 0.05.

RESULTS
Of the 543 students attending the 1 st , 3 rd , 5 th , 7 th , 9 th ,  Table 1 shows the mean values of empathy, anxiety and depression per university and per attended course semester.
As observed, at university B, the mean value of empathy was lower in the 12 th semester than in the 3 rd and 5 th semesters, while anxiety was higher in the 11 th semester than in all other semesters and depression was higher in the 11 th than when compared to the 12 th semester. Figure 1 shows the overall mean of empathy, analyzed with the JSE, by sex, course period and intended specialty, assessed with the JSE. It is observed that the mean was high (shown in the observations), higher in the female sex and among students who intended to follow people-oriented specialties, regardless of the sex. There was no difference between the course semesters. Figure 2 shows the mean scores of anxiety, analyzed with the BAI, by sex, course semester and intended specialty among the study participants. As observed, the female sex shows higher means. Figure 3 shows the mean scores of depression, analyzed with the BDI, by sex, course semester and intended specialty. As observed, females have higher means of depression. Table 2 shows the prevalence of anxiety and depression by sex and by university. As it can be observed, the prevalence of anxiety and depression was similar at both universities and higher among females.  Abbreviations -n: number; numbers that appear followed by parentheses represent the mean and standard deviation. a. Four students did not fill out the sex variable. B. The mean score in the Beck Depression Inventory (BDI) was 11.9 (SD = 9.0) [9.9 (SD = 8.  Three students had had suicidal ideations in the week the questionnaire was applied, two from University A and one from University B, one male and two females.

DISCUSSION
In our study, the mean empathy score was high at both universities and higher among women. These findings are in line with articles included in the review by Andersen et al. 46 .
As for the higher empathy among students who intended to follow people-oriented specialties, these authors also indicated studies with similar findings 46 .
Our initial hypothesis of a decline in empathy throughout the course was not confirmed, with no difference in empathy throughout the course in the two assessed universities, which is also in line with some studies from the meta-analysis by Spatoula et al., which found studies with variable findings 45 .
We also found that women showed higher mean values for depression and anxiety than men, which is in line with the findings of several studies [12][13][14]16 .
In our study, the prevalence of anxiety was 59% when considering the presence of mild to severe anxiety (BAI >10). This prevalence is higher than the overall prevalence of anxiety in the meta-analysis by  49 , the prevalence of depression in our study is 26.4%, similar to the overall rate found by Tam and Pacheco 19 .
The prevalence of suicidal ideation in our study was 11.9%. This finding is of great concern, considering that approximately 12 of every 100 students had thought of suicide in the week preceding data collection. Similar findings were found by Rotenstein et al., which was 11.1% (95% CI = 9.0 -13.7%) 12 .
Our initial hypothesis that anxiety and depression would increase throughout the course was not confirmed, with an increase being found only in the 11 th semester of the course in one of the universities.
An association was found between anxiety and the JSE 'Walking in Patient's shoes' subscale, which is equivalent to empathic stress, with a higher mean observed in students with severe anxiety. However, we did not find an association between this subscale and depression, as indicated in the literature, but we found its association with anxiety 25 . Moreover, there is a trend (p = 0.05) towards the association between suicidal ideation and this component, which should be further explored in future studies with a larger number of participants.
We did not find any association of JSE and its 'Perspective taking' and 'Compassionate care' subscales with anxiety or depression. Our study has some limitations. One of them was the difficulty of contacting students in the clerkship, especially those attending the 11 th and 12 th semesters of the course, because they are distributed in smaller groups, with activities in different sectors inside and outside the university. As higher means of anxiety and depression were found among students in the 11 th semester of University B, we questioned whether this finding could have a participation bias, due to the adherence of more anxious or depressed students. Another limitation was the use of self-completed instruments, which are subject to social desirability. Additionally, it was not asked whether the participant used any drugs for depression or anxiety, which could indicate their presence among them and would increase the number of students with depression and anxiety diagnosis.
Also, only the biological sex was considered and not gender identity. Finally, a limitation that has occurred in studies on the prevalence of anxiety and depression is the variability of their prevalence resulting from different cutoff points in their classification, even in the same instrument. We draw attention to the high prevalence of anxiety, depression and suicidal ideation among medical students in our study. We hypothesize that students who put themselves in the patient's place, starting to feel like them and losing their own perspectives, may have greater anxiety, which can also generate greater stress and loss of therapeutic potential.
We therefore suggest greater attention to students' mental health, seeking strategies to promote it as well to prevent and treat its harm, as it is necessary to take care of these students, so they can feel well throughout the course and can, in the future, take care more efficiently of their health and the health of the population attended by them.

CONCLUSIONS
The prevalence of anxiety, depression and suicidal ideation among medical students was high, in line with the findings in the literature. The mean scores of empathy, anxiety and depression were higher in women than in men.
Empathy was higher among students who intended to pursue people-oriented specialties and there was no change in their means throughout the course.
Depression and anxiety remained stable throughout the course, except for higher anxiety and depression rates among students in the 11 th semester of one of the assessed universities.
There was an association between severe anxiety and the JSE 'Walking in patient's shoes' subscale, which was more related to empathic stress. More studies should be carried out to investigate this association.
The findings demonstrate the urgent need for attention and care to the mental health of medical students, showing that the responsibility of educators and managers related to this issue is high.