Prevalence and Associated Factors of Depression in Medical Students in a Northern Thailand University: A Cross-Sectional Study

This study was conducted to investigate the prevalence and associated factors of depression in medical students. This cross-sectional study investigated the prevalence and associated factors of depression in medical students from May 2018 to April 2019. Depression was diagnosed using the nine-item Patient Health Questionnaire. We evaluated the following potential predictors: demographic data, stressors, psychiatric comorbidities, emotional intelligence (EI), and perceived social support. The association between potential factors and depression was analyzed using multiple logistic regression analysis. The prevalence of depression was 149 of 706 students with 12.5% suicidality. Second- and fourth-year medical students were high-risk groups. Risk factors identified were insufficient income, physical illness, and previous psychiatric illness. Depression in medical students likely coincides with anxiety, internet addiction, sleep problems, and loneliness. Highly associated stressors were personal relationships, physical health, mental health, difficulties in social relationships, satisfaction with grades, and boredom with medical education. Protective EI factors included emotional self-control, problem-solving abilities, inner peace, and life satisfaction. Up to 21.1% of medical students had depression. In this study, among multiple known risk factors of depression, we found that EI is the novel protective factor against depression among medical students. EI training might be protective intervention for medical students in the future.


Introduction
Depression can be characterized by persistent and recurrent sadness or lack of pleasure that eventually impairs a person's functional ability to work, learn, or cope with daily life. The World Health Organization reports that the prevalence of depression worldwide is currently 4.4%. However, the prevalence of depression in medical students is reported to be 3-10 times higher (11.5-48.2%) [1][2][3]. Among Thai medical students, the prevalence of depression is 27% [4][5][6][7][8]. Depression is the most concerning mental health problem in medical students because it impacts students' daily life, academic performance, and patient care [9,10]. Therefore, a comprehensive understanding of its impact and association is necessary to design preventive interventions. A study from Hawaii proposed the preventive intervention for depression, which consisted of faculty education, medical student counseling, and medical student education [11]. In Thailand, only screening and counseling is conducted to manage depression among medical students. Furthermore, no program for preventing depression has been established in Thailand. Exploring factors associated with depression among Thai medical students, especially protective factors, is beneficial for developing such a program.
Previous research has found an association between depression and relationship problems, lack of support systems, poor sleep quality, lack of motivation to study medicine, and pessimism [4,8]. Moreover, a Thai study in Vajira Hospital found an association between psychological abuse and depression [8].
Furthermore, many studies have shown that internet and game addiction have direct and indirect influences on anxiety, stress, and depression [12,13]. In addition, other psychiatric problems (anxiety, loneliness, and burnout) have been positively correlated with depression in many recent studies [14][15][16][17][18]. Conversely, emotional intelligence (EI) and perceived social support (PSS) are inversely correlated with depression and anxiety [19][20][21][22][23]. To our knowledge, these associated factors may also have correlations. Thus, adjusting for all factors that influence depression is necessary. Although many studies indicate associations between the said factors and depression, no study has suggested methods of adjusting all the potential factors of depression due to insufficient study size.
We hypothesized that a younger age (implying lower maturity) indicates worse mental health. Thus, an evaluation of depression throughout the years of study among Thai medical students is required. However, data on depression across the six years of medical school in Thailand are limited, especially in northern Thailand. Only one study was conducted across all year levels [8], and almost all studies in Thailand have been conducted among medical students in various levels of education [4][5][6][7]. Furthermore, studies of factors associated with depression in Thailand reported only problems but not protective factors.
For this reason, our primary objective was to investigate the prevalence of depression among northern Thai medical students. The secondary objective was to explore both potential risk and protective factors of depression across all year levels in medical school.

Study Design
This cross-sectional study was conducted at the Faculty of Medicine of Chiang Mai University in northern Thailand. The Research Ethics Committee of the Faculty of Medicine approved the study protocol (Reference Number 100/7003).

Research Setting
The Thai medical curriculum involves a six-year course and requires high school graduation for admission. The first-year curriculum consists of basic science lectures. Second-and third-year curricula involve a deeper study of medicine with lectures and problem-based learning of organ systems-based blocks. The last three years of medical school are the clinical training period.

Participants
We evaluated 1444 medical students of all year levels from May 2018 to April 2019. We contemporarily collect the data within a year across all year levels in medical school. Participants were asked to provide written consent and privately complete the anonymous self-reported questionnaires without receiving any compensation. Participants could withdraw at any time without giving a reason. Participants were limited to one response per person. The inclusion criteria of this study were current medical students, willing to participate, and having any device to conduct our online questionnaires. This study included both diagnosed and undiagnosed participants, providing a practical assessment of possible risk factors in a diverse population. We excluded students who did not provide consent or submitted incomplete questionnaires from further analysis.

General Demographic Data
The questionnaires consisted of items on sociodemographic and health-related characteristics, including sex, age, year of study [8], hometown [4,24,25], relationship status [26], financial status [27], accommodations [28], underlying diseases (including mental health difficulties and substance usage), and family psychiatric history. We also collected information on perceived causes of stress, such as education, finance, relationships, health, and others. The students were further interviewed about their social adjustment ability, academic satisfaction, and reasons for studying medicine. The revised Thai version of the Multi-Dimensional Scales of PSS (r-MSPSS) is a 12-item self-report measuring an individual's experience of being supported by family, friends, and special persons. The higher the score, the higher the PSS level (Cronbach's alpha of 0.91) [38].

Depression
The depression group was determined using the nine-item Patient Health Questionnaire (PHQ-9, Thai version), which evaluated the frequency of depressive symptoms occurring in the past 2 weeks. The recommended cutoff score is ≥9 for diagnosing major depression with a sensitivity of 0.84 and a specificity of 0.77 (compared with the Thai version of the Mini International Neuropsychiatric Interview) [39]. The severity of depression is further classified into three groups: mild (9-14 points), moderate (15-19 points), and severe (≥20 points) [40].

Study Size Estimation
In the previous study, the prevalence of depression among Thai medical students was 30.54% [8,12]. We confirmed the adequacy of the statistical power of the study size using W.G. Cochran's equation [41], obtaining an alpha of 0.05 and an error of 0.05%. Thus, 326 participants were required to obtain adequate statistical power.

Statistical Analysis
Categorical data were expressed as frequency and percentage. Normally distributed continuous data were expressed as mean ± standard deviation (SD). Non-normally distributed continuous data were expressed as median and interquartile range. Univariable logistic regression was used in the effect estimation of each predictor.
Regarding the exploratory model strategy, we included all known risk factors of depression in a multivariable analysis without pre-selection from the preceding univariable analysis. According to enough events per variable (EPVs), at least 10 EPVs were required for multivariable binary logistic regression [42]. However, the prevalence of depression was less than 10 times the number of the related factor; thus, including all factors into a single multivariable model is unacceptable. Therefore, we adjusted this effect using the confounder summary score concept [43].
First, we divided all potential risk factors into five domains (demographic data, stressors, psychiatric problems, EI domain, and social support). Each predictor domain was analyzed using multivariable binary logistic regression to estimate the probability of depression. Five models were generated, one from each domain. Each domain was influenced by the others. The predicted values from each model containing known potential factors from the first step were included in the final model as regressors. The investigators performed statistical analysis using the STATA program version 16 (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX, USA: StataCorp LLC.). The statistical significance was p < 0.05 (two-sided).

Results
The overall response rate was 48.9% (706 of 1444), including 77.7% (195 of 251) of first-year medical students, 45.7% (217 of 475) of late second-year students, and 40.9% (294 of 718) of third-to sixth-year students (clinical training period). The percentages of female medical students and of those who repeated a grade were similar between participants and total population (female: 48.7% of 48.9%; repeated a grade: 3.3% of 3.3%).
The mean age of the participants was 20.6 ± 2.0 years old, with no significant difference between the depression and non-depression groups. Among participants with depression, no significant difference in depression was found among those who received a higher income per month. However, a significant difference in depression was reported among those who received insufficient income. The depression group had significantly more medical comorbidities, psychiatric disorders, and family histories of psychiatric disorder. Although most students in the two groups drank alcohol less than once a week and did not exceed a standard measurement drink, more students in the depressed group reported using alcohol as self-medication to reduce stress and improve sleep. Only 4.7% of participants smoked (≤10 cigarettes per day), with no significant difference in smoking between the depression and non-depression groups. No significant differences were found in other demographic data (Table 1).

Depression
The prevalence of depression in this study was 21.1% (149 of 706). The most affected group were second-and fourth-year students. However, no significant differences were found among the year levels (p = 0.316). The rate of suicidality was 12.5% (88 of 706), with 62.5% (55 of 88) of students with suicidal ideation having some degree of depression.
Sixth-year medical students had the highest rate of suicidality (Table 2). Surprisingly, less than 20% (24 of 149) of participants in the depression group had ever made use of any mental health care service. About 70% (16 of 24) of them had been diagnosed with major depressive disorder (MDD); unfortunately, only 50% (7 of 16) of students with MDD continued treatment until the end of the study period.
All psychiatric problems and possible protective factors of interest were highly significant in univariable analysis (Table 3). Factors associated with depression from the five separate models using the confounder summary score model are shown in Table 2.
Three demographic factors, six stressors, four psychiatric problems, and four EI domains were associated with depression. Confounder summary score was developed as shown in Table 4.

Discussion
This study found that the prevalence of depression in northern Thailand medical students was 21.1%, of which 12.5% had suicidal ideation. Furthermore, nearly 60% of sixthyear students who were depressed had suicidal ideation. Psychological problems associated with depression were anxiety, sleep problems, internet addiction, and loneliness. A possible protective factor against depression was EI, especially with regards problem-solving, life satisfaction, peace, and emotional self-control. Among medical students with depression, the significant stressors were mental health, relationships with loved ones, physical health, dissatisfaction with their course, and difficulty in social relationships. However, their satisfaction with grades had a positive effect on reducing depression. Demographic data related to depression included a history of psychiatric disorders, medical comorbidities, and sufficient income.

Prevalence
Medical students are undisputedly at a high risk of depression [3]. Prevalence of depression among Thai medical students is lower than the global rate (16.0-39.1% vs. 11.5-48.2%) [6,[9][10][11][12], as confirmed in the present study. A higher prevalence of depression among second-and fourth-year medical students highlights the importance of the transitional periods in the course. The transition from learning basic science in the first year to learning medical concepts in the second year was frequently described as the most stressful transition due to changes in the learning environment, teaching styles, increased study burden, and competition among medical students. Moreover, the transition from the preclinical years to the first clinical year was the second most stressful transition. Many students reported experiencing a sense of lacking enough time and feelings of uselessness and inadequacy [44].
Our study emphasized the undertreatment of depression, which has remained unsolved for decades [28]. Medical students, whose teachers are doctors, nevertheless reported obstacles to accessing mental health care, usually due to a lack of time, stigmatization from themselves or others, and the fear of not achieving good grades and sabotaging their future career. More proactive solutions are required, which includes an increased role of the medical professors, not just the psychiatrists. In our opinion, medical professors have essential roles in solving these problems in many ways, for example:

1.
Providing time in the curriculum for students to take care of themselves; 2.
Identifying students highly prone to depression and providing sources of help; 3.
Taking a stand to reduce the stigmatization of depression; and 4.

Significant Psychiatric Problems
The significant association between anxiety and depression among medical students is not surprising, but the screening and management plan is. Anxiety co-occurs with depression frequently because of similar etiologies, such as genetic vulnerability, neuroticism, and elevated corticotropin-releasing factors [46,47]. In addition, anxiety increases the risk for depression due to increased impairment and debilitating factors [48]. Anxiety and depression, when combined, are more severe, more disabling, more resistant to treatment, and presents a greater risk of suicide than either disorder alone [49].
Robust research has shown sleep problems are a risk factor and symptom of depression [50]. Sleep problems involve a similar pathophysiology to depression, such as increased inflammatory cytokine expression, alterations in the level of monoamines (serotonin and norepinephrine), and interruption of the circadian rhythm. A genetic overlap between insomnia and depression has also been reported [51]. Moreover, the co-occurrence of sleep disturbance and depression decreases treatment response and increases depression severity, duration, and relapse rates [52,53].
We found that internet addiction was highly associated with depression. Depressed students tended to use the internet for social support, achievement, and pleasure of control, which can alleviate depressed feelings. However, at the same time, excessive internet use is a risk factor for depression because online communication displaces actual social communication, resulting in an ineffective ability to cope with difficult real-life situations [54].
Consistent with previous research [55], this study found an association between a high loneliness score with depression among medical students. Therefore, we hypothesized that we could reduce depression by reducing loneliness. To our knowledge, four primary intervention strategies for reducing loneliness have been identified: improving social skills, enhancing social support, increasing opportunities for social contact, and addressing maladaptive social cognition [56].

EI
We found that EI is a possible protective factor against depression, especially in the emotional self-control, problem-solving, peace, and life satisfaction domains.
Emotional self-control includes the ability to process negative thoughts and feelings. Inability to manage negative thoughts causes depression by overstimulating the hypothalamic pituitary adrenal axis and increasing cortisol production. Conversely, a sense of self-control raises self-esteem and decreases depression [57]. Problem-solving skills can buffer the negative correlation between stress and depression. Prior research has shown the effectiveness of problem-solving therapy for preventing and treating depression [58]. Peace, the Eastern concept of happiness, is a well-established protective factor against depression. In addition, showing gratitude can alleviate depressive symptoms [59]. Individual perception of life satisfaction depends not on how much they have but on how they feel. Ways to increase life satisfaction or decrease depression identified in previous studies include optimal self-esteem, good social interactions, and a peaceful environment [58].
Programs for improving EI improving may improve physician wellness and burnout problems [60]. However, no EI improvement programs have been included in Thailand's conventional medical program. Thus, our study might contribute to a national depression prevention program.

Stressors
Although medical students with depression in this study are twice as likely to be concerned about mental health than individuals without depression, the rate of treatment for depression was relatively low. This finding implies that access to mental health services might be caused not only by low awareness, but also by insufficient knowledge on depression. In addition, having poor mental health impacts their functioning, reducing self-esteem and developing depression. Physical health also affects depression because it affects students' ability to work or learn effectively [61].
Failure in or obstruction to psychosocial development (romantic relationship problems or difficulty in adjusting social relationships) can cause depression. The impact of age on an individual is discussed in Erik Erikson's psychosocial development transition periods of "identity vs. role confusion" and "intimacy vs. isolation." Intimacy not only means romantic relationships, but also friendships and other social relationships. Individuals need much more energy to adjust themselves and progress to the next development stage [62].
Academic boredom might be caused by feelings of anhedonia, leading to depression and, consequently, reduced learning performance and engagement [63]. Some studies found an association of boredom with inattention, anxiety, and depression [64]. Conversely, satisfaction with grades is, unsurprisingly, a possible protective factor against depression.

Demographic Data
As mentioned previously, having a psychiatric or medical illness increases the risk for depression. Medical students with depression showed a significant association between depression and perceived insufficiency of income, although they have a higher actual income per month. As also reported in a previous study, a student's perception of their financial status, rather than their actual financial status, was significantly associated with depression [65].

Limitations
Our study has some limitations. First, our findings could not explain the causality because of the cross-sectional design. Second, the actual prevalence of depression might be higher than reported because of selection bias. Medical students with severe depression might have been unable to complete the questionnaire because of poor concentration or lack of energy, and thus, might have been excluded from the analysis. However, our results might still be representative of the target population, specifically in northern Thailand, given no significant differences in the percentage of sex and medical students repeating a grade between participants and all medical students. Nevertheless, our findings have limited generalizability to other populations.

Conclusions
One-fifth of the medical students in this study had depression. Medical students at high risk for depression were those with medical comorbidities, any history of psychiatric disorders, and perceived insufficient income. Medical students facing problems in the following aspects should be monitored for depression: romantic relationships, physical health, mental health, difficulties in social relationships, satisfaction with grades, and boredom with education. Consultants should consider depression in medical students with the following psychiatric comorbidities: anxiety, internet addiction, poor sleep quality, and loneliness. Possible protective factors against depression were EI factors, such as emotional self-control, problem-solving, inner peace, and life satisfaction. Early detection is not enough, and continuous care is crucial. Integrating EI improvement programs might help mitigate depression and should be investigated further.
Appendix A   Table A1. The normal range of EI score in each subscale regarding age.