BMC Medical Education BioMed Central

Research article Open Access

The characteristics of depressive symptoms in medical students
during medical education and training: a cross-sectional study
Sergio Baldassin*1, Tânia Correa de Toledo Ferraz Alves1,2, Arthur Guerra de
Andrade1,2 and Luiz Antonio Nogueira Martins3

Address: 1Psychiatry and Medical Psychology Disciplines, ABC Regional Medical School, Santo André, Brazil, 2Psychiatry Department of the
Medical Faculty, University of São Paulo, São Paulo, Brazil and 3Psychiatry and Medical Psychology Department, Federal University of São Paulo,
São Paulo, Brazil
Email: Sergio Baldassin* - spbaldassin@uol.com.br; Tânia Correa de Toledo Ferraz Alves - tania_alves@hotmail.com; Arthur Guerra de
Andrade - aandrade@usp.br; Luiz Antonio Nogueira Martins - nogmart2004@yahoo.com.br
* Corresponding author

Published: 11 December 2008 Received: 4 April 2008

Accepted: 11 December 2008
BMC Medical Education 2008, 8:60 doi:10.1186/1472-6920-8-60
This article is available from: http://www.biomedcentral.com/1472-6920/8/60
© 2008 Baldassin et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract
Background: Medical education and training can contribute to the development of depressive
symptoms that might lead to possible academic and professional consequences. We aimed to
investigate the characteristics of depressive symptoms among 481 medical students (79.8% of the
total who matriculated).
Methods: The Beck Depression Inventory (BDI) and cluster analyses were used in order to better
describe the characteristics of depressive symptoms. Medical education and training in Brazil is
divided into basic (1st and 2nd years), intermediate (3rd and 4th years), and internship (5th and 6th
years) periods. The study organized each item from the BDI into the following three clusters:
affective, cognitive, and somatic. Statistical analyses were performed using analysis of variance
(ANOVA) with post-hoc Tukey corrected for multiple comparisons.
Results: There were 184 (38.2%) students with depressive symptoms (BDI > 9). The internship
period resulted in the highest BDI scores in comparison to both the basic (p < .001) and
intermediate (p < .001) periods. Affective, cognitive, and somatic clusters were significantly higher
in the internship period. An exploratory analysis of possible risk factors showed that females (p =
.020) not having a parent who practiced medicine (p = .016), and the internship period (p = .001)
were factors for the development of depressive symptoms.
Conclusion: There is a high prevalence towards depressive symptoms among medical students,
particularly females, in the internship level, mainly involving the somatic and affective clusters, and
not having a parent who practiced medicine. The active assessment of these students in evaluating
their depressive symptoms is important in order to prevent the development of co-morbidities and
suicide risk.

Page 1 of 8
(page number not for citation purposes)
BMC Medical Education 2008, 8:60 http://www.biomedcentral.com/1472-6920/8/60

Background programs and others are 5 or 6 years undergraduate pro-

Depressive symptoms are highly prevalent among medi- grams. Many programs have early clinical experience and
cal students. Several studies have revealed that medical the boundaries between clinical and preclinical are not
students are susceptible to high rates of morbidity during exactly clear. In Brazil, we have a 6-year program, and the
their undergraduate years [1-5] and this can be related to medical curriculum is divided into basic (1st and 2nd
impairment in the development of professional, aca- years), intermediate (3rd and 4th years), and internship
demic, and social skills [6-9]. In addition, this co-morbid- (5th and 6th years) periods. The curriculum for the basic
ity is associated with an increased risk of suicide, phase focuses on preclinical disciplines involving study-
evaluated by attempted and completed suicides [10,11]. ing both structure and function of the cell and the human
body (such as anatomy and biochemistry). The interme-
Helmers et al. (1997) compared the presence of depressive diate phase focuses on internal medicine disciplines and
and anxiety symptoms among medical students and other preliminary clinical experience, with a compilation of
disciplines in higher education [12]. The authors found courses in the main areas (general medicine, public
that medical students experienced less stress than law stu- health, surgery, pediatrics, and gynecology). The intern-
dents, graduate students, and the general population, ship period is developed in the general hospital and emer-
although medical students had elevated scores on stress gency units in a 2-year direct supervision learning action
and depressed mood inventories at the transition from program.
basic-to-clinical training. However, it does not seem to be
an adequate comparison, considering that they are differ- Even though there is an abundance of literature regarding
ent populations, with very different curriculum character- the presence of both depression and depressive symptoms
istics and methods of teaching-learning. More recently, among medical students, it can be difficult to distinguish
Dyrbye et al. [13] systematically reviewed the literature from the effects of the stress inherent in student life [28].
reporting on depression, anxiety, and burnout among There is scattered information regarding the specific
U.S. and Canadian medical students. The authors con- aspects of these symptoms, such as cognition, somatic,
cluded that medical school is a time of significant psycho- and affective dimensions. None of the above-mentioned
logical distress for physicians-in-training; however, the studies examined the specific pattern of depressive symp-
current available data was insufficient to draw firm con- toms during medical education and training. Better
clusions on the causes and consequences of student dis- knowledge and understanding of the symptoms involved
tress. in the depression in medical students could assist in the
development of specific target programs (like mentoring
Medical education and training can directly contribute to and tutoring), thus helping professors and medical educa-
the development of depression [13] and behavioral prob- tors to better understand and identify students at risk in
lems, such as alcohol and drug abuse [14,15]. During the each year of education or level of training and to reduce
first semester, there are significant changes in the student's the impact of any disturbances in attitudes and behavior
daily habits [16,17]. Other issues may lead to the develop- which are imperative in order to make the students aware
ment of depressive symptoms among medical students, of their own risk factors [29,30]. Therefore, the current
such as work overload, competitive environment, con- study was aimed at evaluating the specific characteristics
stant pressure of examination/assessment, as well as the of the depressive symptoms in medical students, as well as
vicissitudes of the coursework, which exposes students to to perform an exploratory analysis to survey possible cor-
several sources of distress from the admission process to relations with several risk factors.
graduation, including dealing with traumatic events, such
as death and dying, ethical dilemmas, dissecting cadavers, Methods
pathologic processes, the first physical examination on a The medical school's Ethics Committee approved the
patient [18], the fear of acquiring diseases, feelings of study, and written consent forms were obtained from all
inadequacy, medical hierarchies, and bullying and harass- participating students.
ment [19-21]. On the other hand, some authors have
focused their studies in identifying risk factors for devel- Students were potentially eligible if currently matriculated
opment of depression in medical students. Some known at the ABC Regional Medical School, a private medical
risk factors for developing affective disorders are gender, school supported by three cities (ABC) near the São Paulo
lack of family support [22,23], personal history of depres- State Capital. A total of 603 students were found to be
sive disorders [24], personal beliefs towards the medical potentially eligible. The questionnaires were anonymous
professional [25,26], and the number of years of school- and were distributed to all students who were present at
ing prior to entry into medical school [27]. the classrooms before their academic activities. Any stu-
dent who did not fill out or return the questionnaire was
The medical education accounts for diversity across differ- considered a "loss." There was an absence of 122 (20.2%)
ent countries; some programs are 4-year graduate entry students. There were no refusals to participate in the

Page 2 of 8
(page number not for citation purposes)
BMC Medical Education 2008, 8:60 http://www.biomedcentral.com/1472-6920/8/60

study. The comparison between absentees and respond- 4, dissatisfaction; 10, episode of crying; 11, irritability;
ents regarding gender and age were not statistically signif- and 12, social withdrawal. The second factor, cognitive
icant (p = .192). The remaining 481 (79.8%) students (21. cluster (items 2, 3, 5, 6, 7, 8, 9, 13, 14, and 20), assesses
9 ± 2.4 years old) from the basic (n = 163, 33.8%), inter- the following cognitive aspects: 2, pessimism; 3, sense of
mediate (n = 164, 34.1%), and internship periods (n = failure; 5, guilt; 6, expectation of punishment; 7, dislike of
154, 32.1%) were assessed using the Beck Depression self; 8, self-accusation; 9, suicidal ideations; 13, indeci-
Inventory (BDI) [31], and a questionnaire evaluating per- siveness; 14, change in body image; and 20, somatic pre-
sonal data (gender, age, having one first degree relative occupation. The third factor, somatic cluster (items 15,
working as a physician, years of study between high 16, 17, 18, 19, and 21), assesses the presence of the fol-
school and entering medical school [27], and living alone lowing characteristics: 15, slowness; 16, insomnia; 17,
or with family). fatigue; 18, appetite; 19, loss of weight; and 21, loss of sex-
ual interest. In addition, cluster scores were computed as
The presence of depressive symptoms among medical stu- the means across all items assigned to a particular cluster
dents was assessed using the BDI, a 21-item self-report (the mean was given preference over the sum of the indi-
inventory designed to measure the severity of depressive vidual item scores to account for differences in the
symptomatology. For the BDI, the answers were dichot- number of items per cluster). The possible score for each
omized between the presence and absence of major of these clusters were as follows: affective cluster, a range
depressive symptoms. The cut-off points for the BDI were of 0–21 and a maximum possible score of 30; cognitive
as follows: minimal or none (0–9), mild (10–16), moder- cluster, a range of 0–28 and a maximum possible score of
ate (17–29), and severe (30–63). The Portuguese version 60; and somatic cluster, a range of 0–14 and a maximal
of the BDI was validated by Gorenstein and Andrade [31] possible score of 36.
and a more detailed characteristic of BDI psychometrics
have been given elsewhere. The internal consistency of the Statistical analysis
Portuguese version of the BDI is in agreement with the lit- Statistical analyses were performed using the Window's
erature (0.81 for non-depressed subjects and 0.88 for Statistical Package for Social Sciences (version 13.0).
depressed patients) [31]. Between-group comparisons of depressive symptoms in
students from three medical period were divided in basic
Clinical studies indicate that the many different symp- (1st and 2nd years), intermediate (3rd and 4th years), and
toms of depression can be grouped into a limited number internship (5th and 6th years) using variance analysis
of clusters [32]. Such cluster analysis can also be helpful (ANOVA) with post-hoc Tukey corrected for multiple com-
in describing the particularities of depressive symptoms in parisons after the Kolmogorov-Smirnov test (normality)
a more differentiated manner than the total BDI score were carried out. The investigation of significant depres-
alone would permit. Studying such symptom clusters sive cluster symptoms in each year was separately carried
rather than individual symptoms has the advantage of out using variance analysis (ANOVA) with post-hoc Tukey.
minimizing the number of correlations examined and This was an exploratory investigation aimed at ascertain-
consequently the risk of committing a Type I error. There- ing whether or not depressive symptoms might be related
fore the study organized each item from the BDI into three to a determined period of the medical course of study in
different clusters in order to perform an exploratory anal- order to generate hypotheses for future studies. Pearson
ysis of the particularities of depressive symptoms during correlations between depressive scores were performed
the medical course of study. In our analysis, the empirical upon the entire medical student cohort and within the
construction of subscales of items from the pool of the 21- three periods. Finally, in order to verify which demo-
item BDI was given preference over a rational approach graphic and risk or protective factor variables most signif-
based on the individual items' face validity, in order to icantly influenced the BDI scores in the overall medical
avoid "inductive bias" caused by the analysts' perceptions student sample, we conducted a multiple-regression anal-
of the cognitive field. In order to avoid the intrinsic disad- ysis using the BDI total score as dependent variable was
vantages of factor analysis outlined above, the factors conducted, involving gender, age, course periods, years to
were derived by using cluster-analytical techniques. get into medical school, first degree parent as a physician,
Unlike factor analysis, cluster analysis groups the ele- living alone or with a family as independent factors
ments of the analysis (i.e., items and persons) into an ("enter method"), and a logistic regression for BDI > 16
appropriate number of mutually exclusive subsets, so that using the cut-off to create a pragmatic clinical parameter
a more straightforward interpretation can be achieved based in the those used during a Portuguese validation:
when the analysis is successful. The first factor, affective BDI > 16 to detect a disphoric range and BDI > 20 to detect
cluster (the sum of scores on items 1, 4, 10, 11, and 12 a considered a depression range [31]. Using this cut of we
from the BDI), represents the core symptoms of a depres- have obtained a ROC curve with a sensitivity of 75% and
sive mood, based on the following symptoms: 1, sadness; specificity at 53%. The significance level used was p < .05.

Page 3 of 8
(page number not for citation purposes)
BMC Medical Education 2008, 8:60 http://www.biomedcentral.com/1472-6920/8/60

Results mean ± S.D. = 2.9 ± 2.4, range 0–14, and maximum pos-
Study Sample sible score = 36. There was a significant difference
Of the 481 students participating in the study, 40.5% were between BDI total scores and gender distribution: female
males, 90.6% lived with family, 17.1% had at least a two- (n = 279, BDI, 9.8 ± 8.1) and male (n = 190, BDI, 8.1 ±
year gap between the end of high school and entering 6.8; t-test, T = 2.363, p = .019).
medical school, and 28.2% had a parent as a physician.
Table 1 summarizes the socio-demographic characteristics Cluster analysis of depressive symptoms in separate years
of the sample. among medical students
Table 2 shows the mean BDI scores in the medical course
In the basic period, there were 163 students with the fol- of study grouped according to the period of the course.
lowing characteristics: the mean age was 20.0 years The mean BDI scores for each period were as follows:
(range, 17–31 years); 34.8% were males; 97.1% lived with basic (8.6 ± 7.9); intermediate (7.0 ± 6.9); and internship
family; 20.3% had at least a two-year gap between the end period (11.7 ± 7.2). The ANOVA showed a significant
of high school and entering into medical school; and association between the year level and the mean BDI
28.2% had a parent as a physician. In the intermediate scores (F = 9.282, degrees of freedom = 5, p < .001). Post
period, there were 164 students with the following char- hoc Tukey for each period showed that there were signifi-
acteristics: the mean age was 21.9 years (range, 20–35 cant differences between the internship period in compar-
years); 44.3% were males; 86.0% lived with family; 16.9% ison to both basic and intermediate periods.
had at least a two-year gap between the end of high school
and entering into medical school (n = 28); and 33.8% had In the ANOVA investigating cluster analysis among the
a parent as a physician. Finally, in the internship period, three periods (basic, intermediate, and internship), signif-
there were 154 students with the following characteristics: icant differences with regard to the affective cluster (2.1 ±
the mean age was 24.0 years (range, 21–36 years); 41.1% 2.7, 1.8 ± 2.0, and 3.5 ± 2.4, respectively, F = 22.220, p <
were males; 88.8% lived with family; 14.7% had at least a .001), cognitive cluster (3.9 ± 3.9, 3.1 ± 3.5, and 4.7 ± 3.8,
two-year gap between the end of high school and entering respectively, F = 6.493, p = .002), and somatic cluster (2.8
into medical school; and 22.9% had a parent as a physi- ± 2.2, 2.1 ± 2.2, and 3.8 ± 2.6 respectively, F = 19.927, p <
cian. .001) were observed. Post hoc Tukey for each period
showed that there were significant differences between
Investigation of depressive symptoms among the entire basic and intermediate periods with the internship peri-
medical student population ods in the affective cluster (p < .001); between the inter-
The general mean of the BDI for the medical course of mediate period with the internship period in the cognitive
study was 9.1 ± 7.6. Considering the BDI cut-off points, cluster (p < .01); and between the basic period and the
184 (38.2%) were depressed (mild depression, n = 120 internship period (p < .01) and between the intermediate
[24.9%]; moderate depression, n = 53 [11%]; and severe period and the internship period (p < .001) in the somatic
depression n = 11 [2.3%]). With regard to the characteris- cluster.
tics of the depressive symptoms presented by the entire
medical student population, the cluster analysis showed Exploratory preliminary investigation of possible
the following mean score ± S.D., range, and maximum protective/risk factors for depression among medical
possible score for each of these clusters: affective cluster, students
mean ± S.D. = 2.4 ± 2.5, range 0–21, maximum possible A multiple regression analysis was conducted on the BDI
score = 30; cognitive cluster, mean ± S.D. = 3.9 ± 3.8, range scores in order to ascertain the effect of other demo-
0–28, maximum possible score = 60; and somatic cluster, graphic variables in the development of depressive symp-
Table 1: Demographic characteristics of medical students

Periods Basic Intermediate Internship Total p

Sample n (% of regular matriculated) 163 (79.1%) 164 (83.7%) 154 (76.7%) 481 (79.8%)
Age – m (sd) 20.0 (2.0) 21.9 (1.5) 24.0 (1.9) 21.9 (2.4) <.001a
Male n (%) 56 (34.8%) 72 (44.3%) 62 (41.1%) 190 (40.5%) .242b
Female n (%) 102 (30.2%) 92 (55.7%) 85 (58.9%) 279 (59.5%)
Living with the family n (%) 155 (97.1%) 140 (86.0%) 131 (88.8%) 426 (90.6%) .011b
Having a parent as MD n (%) 44 (28.2%) 56 (33.8%) 33 (22.9%) 133 (28.2%) .113b
> 2 years to enter medical school n (%) 31 (20.3%) 28 (16.9%) 21 (14.7%) 80 (17.1%) .516b

Mean years of studying to enter medical school – m (sd) 1.7 (1.0) 1.7 (0.8) 1.6 (0.9) 1.7 (0.9) .337a

Statistical analyses performed: a ANOVA; b Chi-square

Page 4 of 8
(page number not for citation purposes)
BMC Medical Education 2008, 8:60 http://www.biomedcentral.com/1472-6920/8/60

Table 2: Analysis by clusters analysis of depressive symptoms assessed by Beck Depression Inventory scores among medical students
(n = 481).

Periods Basic Intermediate internship Total ANOVA

Cluster m (sd) m (sd) m (sd) m (sd)

Affective 2.1 (2.7) 1.8 (2.0) 3.5 (2.4)a,b 2.4 (2.5) <.001
Cognitive 4.0 (3.9) 3.2 (3.4)c 4.7 (3.6) 3.9 (3.8) <.01
Somatic 2.8 (2.2)c 2.1 (2.2)d 3.8 (2.6)b 2.9 (2.4) <.001

BDI total 8.6 (7.9)c 7.0 (6.9) 11.7 (7.2)b 9.1 (7.6) <.001

significant post hoc Tukey comparing each year of medical course

a p < .001 in comparison to basic period medical school
b p < .001 in comparison to intermediate period medical school
c p < .01 in comparison to internship period medical school
d p < .05 in comparison to basic period medical school

toms in medical students. Linear regression for a total BDI quantitatively rather than qualitatively from major
revealed that there was a significantly higher risk for devel- depression [35]. The hypothesis was that the nature of the
opment of depressive symptoms in course periods (CI depressive experience would differ in intensity (quantita-
95%, .873:2.580, p < .001). Others significant risk factors tively), but not in kind (qualitatively), in individuals with
assessed by linear regression in the sample were female mild versus severe depression. In short, the cluster analy-
gender (CI 95%, -3.140:-.353, p < .05) and those not hav- sis was used to empirically sort the symptom profiles of
ing a parent who was a physician (CI 95%, -3.395:-.311, p the analogue participants of this study. The analysis by
< .05). Significant demographic characteristics assessed by clusters showed that the principal cluster responsible for
logistic regression for BDI > 16 in the sample were course BDI scores was the affective cluster, which includes a core
period (CI95%, 1.284:2.592; p < 0.01) and those not hav- evaluation of mood disorders, namely sadness, lack of
ing a parent who was a physician (CI95%, .168:.805, p < pleasure, crying, irritability, and losing interest in people.
.05) There were no other significant findings, including This aspect was significantly different among the medical
living alone (without family) (CI 95%, .567:5.013, p = education and training periods, with higher scores in the
.348) and having at least a two-year gap between the end internship period, followed by the basic course and lower
of high school and entering into medical school (CI 95%, scores in the intermediate course. The cognitive cluster,
.814:.3.059; p = .177). which included aspects of hopelessness, failure, criticism,
self-punishment, self-blame, indecision, guilt, self-
Discussion and conclusion appearance, and somatic worrying was also higher in the
This is the first study that directly evaluated, in a cross-sec- internship period, probably would be associated to nega-
tional design, the characteristics of depressive symptoms tive feelings, such as fear and insecurity [36], related to
by applying clusters. Higher total BDI scores during the entry into the internship period. Frequently pre-intern-
internship period (11. 7 ± 7. 2) were observed. On the ship students experience the fear of "knowing nothing"
other hand, the total BDI scores were lower in the inter- [37], and are insecure about the physical examination of
mediate period (7.0 ± 6.9). The means of somatic, cogni- other human beings [38,39]. Finally, the somatic cluster
tive, and affective clusters decreased during this part of the refers to insomnia [40], tiredness, loss of appetite [41],
course 1.5-fold and tiredness decreased by one-half. These loss of weight and sexual interest, therefore it is not sur-
increased rates of BDI scores during the internship [33,34] prising that in internship period the scores are signifi-
period of medical school are associated with a decrease in cantly higher reflecting the nights on call, without
student health, and probably refers to periods when pro- sleeping, devoid of friends and family support, facing the
fessors and educators should be aware about suicidal death of patients, with reduced time of pleasure activities
thoughts and risk. Future studies specifically addressing [42-44], and fear regarding the future and residency pro-
this aspect should be designed in order to investigate the gram examination [45].
suicide risk among medical students.
The relevance of analyzing the characteristics of depressive
The interpretation of both BDI score as means and stand- symptoms among medical students might help educators
ard deviation, and a second analysis using a cut-off for in dealing with different patterns of depression and devel-
depression in BDI > 16 was in agreement with the propo- oping specific target strategies [46]. The higher affective
sition of the experience of depression as a continuum, and cluster of depressive scores in the internship may affect
that sub-threshold or sub-syndromal depression differs since the beginning the feelings of pleasure and compas-

Page 5 of 8
(page number not for citation purposes)
BMC Medical Education 2008, 8:60 http://www.biomedcentral.com/1472-6920/8/60

sion of the physician-patient relationship [47,48]. The The findings of this study must be interpreted with cau-
higher cognitive cluster during the internship occurs at the tion due to its methodologic shortcomings, including the
same time of deceptions with the institution and health use of self-administered inventories and no structured
system, fears of the professional future, failures, perform- interviews for clinic diagnoses. However, it is unlikely that
ance, even dislike own appearance, this in general not our results of higher scores in the internship period were
good, due the lack of sleep and leisure and fewer physical limited only to the presence of non-specific depressive
and sexual activities. Taken all together, it might suggest symptoms due to the BDI that has been often been used
that the possible development of coping mechanisms dur- to separate depression\non-depression by applying a cut-
ing medical education and training is responsible for off at 17 points {Department of Psychiatry, 1998 #7901},
lower scores in the intermediate period. However, these therefore, the total of 13.3% with moderate/severe depres-
mechanisms are not necessarily functional [49,50], sive symptoms (BDI, 17–63) is still concerning. Even
healthy or effectives, perhaps influencing the future pro- though the inclusion of an exploratory analysis of the dif-
fessional [51]. Coping strategies developed in response to ferent subscales of BDI could be considered a little forced,
a stressful event can be subdivided into "functional" this was done in order to generate future hypotheses to
(problem solving, cognitive re-structuring, and seeking identify specific depressive symptoms in the different
social and emotional support) or "dysfunctional" (prob- periods of medical education and training, once previous
lem avoidance, social and emotional isolation, and guilt studies have assessed either the presence of depressive
and self-criticism) [52]. symptoms comparing only the total BDI scores. Better
knowledge regarding specific depressive symptoms could
The study also showed that female students were more assist medical educators in the development of suitable
susceptible towards developing depression symptoms strategies to deal with the particularities of depression in
than male students. Gender comparisons in the academic medical students.
profession, prior to, during [39] and after the medical
course of study sho ws that the gap between males and We should consider if the cohort characteristic of each
females in medicine is getting narrower. Even though aca- year or period might have influenced our results. We
demically there is no significant gender differences, sev- observed that the 2nd year students seems to be quite dif-
eral non-cognitive aspects often shows a different pattern ferent from the others years, in term of the lowest
among males and females. It is interesting to note that response rate (64%) and gender. This might have influ-
Hojat et al. (1999) studying a large sample of medical stu- enced the differences observed in the score. The absence
dents, indeed observed that male and female medical stu- of 20% of the medical students in class might limit the
dents have a different pattern of stress response [53]. generalization of our finding, However, it is feasible to
Further investigation on gender difference among medical think that absence might be also be related to the presence
students is required in order to develop a specific-gender of depressive disorder, particularly in the 2nd year. Several
program to prevent stress-related mental disorders. In the studies have related the presence of depressive symptoms
exploratory analysis of protective and risk factors for and absence at school or work [57]. The current study
developing depressive symptoms, it was found that stud- with 80% of all students of a medical school is based on
ying more than 2 years to get in medical school was a risk self-answer questionnaires that might be associated to a
factor for depression. In Brazil, there is a very special con- bias of minimizing or maximizing symptoms. However, it
dition the "vestibular" [27] examination has to be passed has the advantage of maintaining anonymities and
after high school in order to enter any Brazilian medical allowed for the study of a large sample of medical stu-
school. Even though the medical profession is highly dis- dents. Finally, even considering that medical students
tressful, there are many students competing for each med- possibly have more depression/anxiety symptoms than
ical school vacancy. Of interest to note that having a other students [58], it is necessary to consider aspects
parent as a physician seems to be a protective factor. It related to medical education and training, such as an
may occur that being a son/daughter of a physician is important risk factor to the development, sustaining and/
probably associated to lower idealism concerning the or worsening of emotional distress. However, our cross
medical course of study [50] and the consequent develop- sectional design prevented us from this particular analy-
ment of social or affective coping skills [54]. Further stud- sis.
ies that specifically address this aspect should be
conducted in order to investigate which particular aspects Finally, an additional limitation to this study is its cross-
of familiar support are possibly protecting the students. sectional design, where the different subgroups may differ
Depressive symptoms, off course, are not prevalent only in respects not directly attributable to the stage of educa-
in the medical students of this study, and as said Hickie tion. We have not measured intra-individual symptoms
and Davenport, it is necessary understanding depression changes during medical education and training using a
as a major condition in different cultural settings [55]. repetitive measure and a longitudinal design. Therefore,

Page 6 of 8
(page number not for citation purposes)
BMC Medical Education 2008, 8:60 http://www.biomedcentral.com/1472-6920/8/60

the presence of individual variability can not be excluded Acknowledgements

and, given the cross-sectional nature of the study, in We thank to the reviewers: Alan Apter, Charlotte E Rees and Kristi J Fer-
which it was not possible to define the temporal relation- guson. We believe that we have achieved a better quality manuscript after
ship between cause and consequence, we cannot state that contemplating the reviewers' suggestions.
the associations observed in the study are of a causal
nature However, our approach allowed us to assess the
References
1. Clark D, Zeldow PB: Vicissitudes of Depressed Mood during
presence of depressive symptoms in a large sample includ- Four Years of Medical School. JAMA 1988, 260:2521-2526.
ing up to 80% of all medical students regularly matricu- 2. Mosley THJ, Perrin SG, Neral SM, Dubbert PM, Grothues CA, Pinto
BM: Stress, coping, and well-being among third-year medical
lated at our medical school. Despite these limitations, this students. Acad Med 1994, 69:765-767.
study indeed contributes to the literature on medical edu- 3. Dyrbye LN, Thomas MR, Eacker A, Harper W, Massie FS Jr, Power
cation. DV, Huschka M, Novotny PJ, Sloan JA, Shanafelt TD: Race, ethnic-
ity, and medical student well-being in the United States. Arch
Intern Med 2007, 167:2103-2109.
In conclusion, our findings of a greater frequency and 4. Dahlin ME, Runeson B: Burnout and psychiatric morbidity
severity higher depressive scores among medical students among medical students entering clinical training: a three
year prospective questionnaire and interview-based study.
on internship period years and lower scores on intermedi- BMC Med Educ 2007, 7:6.
ate period, as well the better knowledge of the characteris- 5. Baldassin S: O desgaste no internato – o nascimento do coping
de um médico. In Saúde Mental do Médico e do Estudante de Medic-
tics of each cluster symptoms, are involved could help us ina Volume 6. 1st edition. Edited by: Guimaraes KBS. São Paulo: Casa
to develop programs (like mentoring and tutoring) specif- do Psicólogo; 2007:51-60.
ically designed for each level to prevent the development 6. Henderson P, Johnson MH: An innovative approach to develop-
ing the reflective skills of medical students. BMC Med Educ
of depression and its consequence, as well as, preparing 2002, 2:4.
our medical students and faculty members for their daily 7. Murdoch-Eaton DG, Levene MI: Formal appraisal of undergrad-
uate medical students: is it worth the effort? Med Teach 2004,
practice and teaching perceptions [59]. 26:28-32.
8. Facundes V, Ludermir A: Common mental disorders among
Competing interests health care students. Rev Bras Psiquiatr 2005, 27:194-200.
9. Kristal SL, Randall-Kristal KA, Thompson BM, Marx JA: 1998–1999
The authors declare that they have no competing interests. SAEM emergency medicine faculty salary and benefits sur-
vey. Acad Emerg Med 1999, 6:1261-1271.
10. Tyssen R, Vaglum P, Gronvold NT, Ekeberg O: Suicidal ideation
Authors' contributions among medical students and young physicians: a nationwide
Each author have made substantial contributions to con- and prospective study of prevalence and predictors. J Affect
ception, design, acquisition, analysis and interpretation of Disord 2001, 64:69-79.
11. Wallin U, Runeson B: Attitudes towards suicide and suicidal
data, have been involved in drafting the manuscript, revis- patients among medical students. Eur Psychiatry 2003,
ing it critically for important intellectual content and 18:329-333.
given final approval of the version to be published. 12. Helmers KF, Danoff D, Steinert Y, Leyton M, Young SN: Stress and
depressed mood in medical students, law students, and grad-
uate students at McGill University. Acad Med 1997, 72:708-714.
About the authors 13. Dyrbye LN, Thomas MR, Shanafelt TD: Systematic review of
depression anxiety, and other indicators of psychological dis-
SB, MD, MsC, is a Professor of Psychiatry and Medical Psy- tress among U.S. and Canadian medical students. Acad Med
chology Disciplines, Medical School, ABC Foundation, 2006, 81:354-373.
where he also coordinates the counseling and medical stu- 14. Andrade AG, Bassit AZ, Kerr-Corrêa F, Tonhon A, Boscovitz E,
Cabral M, Rassi R, Potério G, Marcondes E, Oliveira M, Duailibi K,
dent service. Fukushima JT: Fatores de risco associados ao uso de álcool e
drogas na vida, entre estudantes de medicina do Estado de
TCdTFA, MD, PhD is a Professor of Psychiatry and Medi- São Paulo. Rev ABPAPAL 1997, 19:117-126.
15. Houwink AP, Kurup AN, Kollars JP, Kral Kollars CA, Carmichael SW,
cal Psychology Disciplines, Medical School, ABC Founda- Pawlina W: Help of third-year medical students decreases
tion, where she coordinates the Affective Disorders first-year medical students' negative psychological reactions
on the first day of gross anatomy dissection. Clin Anat 2004,
Ambulatory and is also a researcher of the Psychiatry 17:328-333.
Department of the Medical Faculty, University of São 16. Rosal MC, Ockene IS, Ockene JK, Barrett SV, Ma Y, Hebert JR: A lon-
Paulo. gitudinal study of students' depression at one medical
school. Acad Med 1997, 72:542-546.
17. Shapiro SL, Shapiro DE, Schwartz GE: Stress management in
AGdA, MD, PhD is Chairman Professor of the Psychiatry medical education: a review of the literature. Acad Med 2000,
and Medical Psychology Disciplines, Medical School, ABC 75:748-759.
18. Rees CE, Bradley P, McLachlan JC: Exploring medical students'
Foundation, and Professor of the Psychiatry Department attitudes towards peer physical examination. Med Teach 2004,
of the Medical Faculty, University of São Paulo. 26:86-88.
19. Omigbodun OO, Odukogbe AT, Omigbodun AO, Yusuf OB, Bella
TT, Olayemi O: Stressors and psychological symptoms in stu-
LAN-M, MD, PhD is a Chairman Professor of Psychiatry dents of medicine and allied health professions in Nigeria.
Department at Federal University of São Paulo, where he Soc Psychiatry Psychiatr Epidemiol 2006, 41:415-421.
20. Baldwin JN, Scott DM, Agrawal S, Bartek JK, Davis-Hall RE, Reardon
is also a coordinator of the Medical Service Ambulatory TP, DeSimone EM 2nd: Assessment of alcohol and other drug
for Residents. use behaviors in health professions students. Subst Abus 2006,
27:27-37.

Page 7 of 8
(page number not for citation purposes)
BMC Medical Education 2008, 8:60 http://www.biomedcentral.com/1472-6920/8/60

21. Henning K, Ey S, Shaw D: Perfectionism, the imposter phenom- 44. Hirata FC, Lima MC, de Bruin VM, Nobrega PR, Wenceslau GP, de
enon and psychological adjustment in medical dental, nurs- Bruin PF: Depression in medical school: the influence of morn-
ing and pharmacy students. Med Educ 1998, 32:456-464. ingness-eveningness. Chronobiol int 2007, 24:939-946.
22. Chandra A, Minkovitz CS: Stigma starts early: gender differ- 45. Nogueira-Martins LA, Stella RC, Nogueira HE: A pioneering expe-
ences in teen willingness to use mental health services. J Ado- rience in Brazil: the creation of a center for assistance and
lesc Health 2006, 38:e751-758. research for medical residents (NAPREME) at the Escola
23. Skarsater I: Parents with first time major depression: percep- Paulista de Medicina, Federal University of São Paulo. Sao
tions of social support for themselves and their children. Paulo Med J 1997, 115(6):1570-1574.
Scand J Caring Sci 2006, 20:308-314. 46. Smith CK, Peterson DF, Degenhardt BF, Johnson JC: Depression
24. Brown S, Gunderman RB: Viewpoint: enhancing the profes- anxiety, and perceived hassles among entering medical stu-
sional fulfillment of physicians. Acad Med 2006, 81:577-582. dents. Psychol Health Med 2007, 12:31-39.
25. Hansen MN: [Social background in recruitment of medical 47. Thomas MR, Dyrbye LN, Huntington JL, Lawson KL, Novotny PJ,
students]. Tidsskr Nor Laegeforen 2005, 125:2213-2215. Sloan JA, Shanafelt TD: How Do Distress and Well-being Relate
26. Vasegh S, Mohammadi MR: Religiosity anxiety, and depression to Medical Student Empathy? A Multicenter Study. J Gen
among a sample of Iranian medical students. Int J Psychiatry Intern Med 2007, 22:177-183.
Med 2007, 37:213-227. 48. Yeh YC, Yen CF, Lai CS, Huang CH, Liu KM, Huang IT: Correlations
27. Millan ER, Azevedo RS, Rossi E, Marco LN: What is behind a stu- between academic achievement and anxiety and depression
dent's choice for becoming a doctor? Clinics 2005, 60:143-150. in medical students experiencing integrated curriculum
28. Rosenthal JM, Okie S: White coat, mood indigo – depression in reform. Kaohsiung J Med Sci 2007, 23:379-386.
medical school. N Engl J Med 2005, 353:1085-1088. 49. Blaya C, Dornelles M, Blaya R, Kipper L, Heldt E, Isolan L, Bond M,
29. Mino Y, Yasuda N, Tsuda T, Shimodera S: Effects of a one-hour Manfro GG: Do defense mechanisms vary according to the
educational program on medical students' attitudes to men- psychiatric disorder? Rev Bras Psiquiatr 2006, 28:179-183.
tal illness. Psychiatry Clin Neurosci 2001, 55:501-507. 50. Sreeramareddy CT, Shankar PR, Binu VS, Mukhopadhyay C, Ray B,
30. Merritt RK, Price JR, Mollison J, Geddes JR: A cluster randomized Menezes RG: Psychological morbidity, sources of stress and
controlled trial to assess the effectiveness of an intervention coping strategies among undergraduate medical students of
to educate students about depression. Psychol Med 2007, Nepal. BMC Med Educ 2007, 7:26.
37:363-372. 51. Woolf K, Cave J, McManus IC, Dacre JE: 'It gives you an under-
31. Gorenstein C, Andrade L: Validation of a Portuguese version of standing you can't get from any book'. The relationship
the Beck Depression Inventory and the State-Trait Anxiety between medical students' and doctors' personal illness
Inventory in Brazilian subjects. Braz J Med Biol Res 1996, experiences and their performance: a qualitative and quan-
29:453-457. titative study. BMC Med Educ 2007, 7:50.
32. Fleck M, Chaves M, Poirier-Littré M, Bourdel M, Loo H, Guelfi J: 52. Peterlini M, Tiberio IF, Saadeh A, Pereira JC, Martins MA: Anxiety
Depression in France and Brazil: factorial structure of the and depression in the first year of medical residency training.
17-item Hamilton depression scale in inpatients. J Nerv Ment Med Educ 2002, 36:66-72.
Dis 2004, 192:103-110. 53. Hojat M, Glaser K, Xu G, Veloski JJ, Christian EB: Gender compar-
33. Kjeldstadli K, Tyssen R, Finset A, Hem E, Gude T, Gronvold NT, Eke- isons of medical students' psychosocial profiles. Med Educ
berg O, Vaglum P: Life satisfaction and resilience in medical 1999, 33:342-349.
school – a six-year longitudinal, nationwide and comparative 54. Rees C, Sheard C: The relationship between medical students'
study. BMC Med Educ 2006, 6:48. attitudes towards communication skills learning and their
34. Alexandrino-Silva C, Pereira MLG, Bustamante C, Ferraz ACT, Bald- demographic and education-related characteristics. Medical
assin S, Andrade AG, Alves TCTF: Prevalencia de ideação suicida Education 2002, 36:1017-1027.
em estudantes da area da saude: um estudo de corte trans- 55. Hickie AMI, Davenport TA, Luscombe GM, Rong Y, Hickie ML, Bell
versal. Rev Bras Psiquiatr 2007, 29:S25-31. MI: The assessment of depression awareness and help-seek-
35. Cox BJ, Enns MW, Larsen DK: The continuity of depression ing behaviour: experiences with the International Depres-
symptoms: use of cluster analysis for profile identification in sion Literacy Survey. BMC Psychiatry 2007, 7:48.
patient and student samples. J Affect Disord 2001, 65:67-73. 56. Richter P, Werner J, Heerlein A, Kraus A, Sauer H: On the validity
36. Baldassin S, Andrade AG, Martins LC: Traços de ansiedade entre of the Beck Depression Inventory. A review. Psychopathology
estudantes de medicina. Arq Med ABC 2006, 31:27-31. 1998, 31:160-168.
37. Nogueira-Martins MC, Nogueira-Martins LA, Turato ER: Medical 57. D'Souza RM, Strazdins L, Broom DH, Rodgers B, Berry HL: Work
students' perceptions of their learning about the doctor- demands, job insecurity and sickness absence from work.
patient relationship: a qualitative study. Med Educ 2006, how productive is the new, flexible labour force? Aust N Z J
40:322-328. Public Health 2006, 30:205-212.
38. Sarikaya O, Civaner M, Kalaca S: The anxieties of medical stu- 58. Dyrbye LN, Thomas MR, Huntington JL, Lawson KL, Novotny PJ,
dents related to clinical training. Int J Clin Pract 2006, Sloan JA, Shanafelt TD: Personal life events and medical student
60:1414-1418. burnout: a multicenter study. Acad Med 2006, 81:374-384.
39. Chandavarkar U, Azzam A, Mathews CA: Anxiety symptoms and 59. Ahmady S, Changiz T, Masiello I, Brommels M: Organizational role
perceived performance in medical students. Depress Anxiety stress among medical school faculty members in Iran: deal-
2006, 24:103-111. ing with role conflict. BMC Med Educ 2007, 7:14.
40. Kawada T, Katsumata M, Suzuki H, Shimizu T: Actigraphic predic-
tors of the depressive state in students with no psychiatric
disorders. J Affect Disord 2007, 98:117-120. Pre-publication history
41. Crow S, Eisenberg ME, Story M, Neumark-Sztainer D: Suicidal The pre-publication history for this paper can be accessed
behavior in adolescents: Relationship to weight status, here:
weight control behaviors, and body dissatisfaction. Int J Eat
Disord 2008, 41:82-87.
42. Buela-Casal G, Miró E, Iáñez MI, Catena A: Relation between http://www.biomedcentral.com/1472-6920/8/60/prepub
habitual sleep duration and depressed mood state: Somatic
versus cognitive symptoms. Int J Clin Health Psychol 2007,
7:615-631.
43. Swami V, Chamorro-Premuzic T, Sinniah D, Maniam T, Kannan K, Sta-
nistreet D, Furnham A: General health mediates the relation-
ship between loneliness, life satisfaction and depression. A
study with Malaysian medical students. Soc Psychiatry Psychiatr
Epidemiol 2007, 42:161-166.

Page 8 of 8
(page number not for citation purposes)