Depression was shown to have high prevalence rates among university students [1,2]. These rates were even higher when it came to medical students, where depression affected one-third [3] to one-half of this population [4]. Coping strategies, such as social support, were recommended for medical students to improve their quality of life [5]. Moreover, it was shown that social support per se may even be a potent protective factor for depression [6].

On average, there are about 864 million daily active users on Facebook, the most visited online social network (OSN) in the world [7]. Kross et al showed that Facebook negatively affected a person's well-being in two ways: affective well-being (ie, how he/she feels moment to moment) and cognitive well-being [8]. In addition, individuals who have used Facebook accounts longer and users who have a greater number of nonpersonal Facebook friends believe that others' lives are better and are even more pleasant than their own [9]; this and other Facebook-mediated comparisons were associated with greater depression [10]. Furthermore, Forest et al reported that people with low self-esteem will not take advantage of the easy means of disclosure on Facebook, despite their beliefs in such a possibility [11]. In addition, it was reported that the increased usage of Facebook photo applications is correlated with body image disturbance and weight dissatisfaction [12].

As a part of infodemiology [13], which is a new method to study health determinants and information on the Internet, Facebook is one possible screening method to study health-related issues. For instance, a previous study based on verbatim text analysis showed that displayers of depressive symptoms are more likely than nondisplayers to have a positive Patient Health Questionnaire-9 (PHQ-9) measure [14].

Furthermore, announcing suicide on Facebook could provide enough time for an intervention [15]. Also, Facebook may be a good means for individuals to disclose their depression, hence enabling others to constantly be updated [16]. Seemingly depressed participants in Social Network Service (SNS) had the opportunity to communicate with peers that they selected. Such behavior may have been promoted by anonymity, hence avoiding face-to-face contact [17].

Finally, depressed people had a declined usage of Facebook features such as location tagging, add friend, and the like feature [18]. However, we were concerned more with the increased usage of Facebook as a possible tool that students could resort to in fighting depression. From this perspective, a subjective report of how each feature is resorted to might be more significant than counting the usage of each. Hence, when it comes to depressed Facebook users, we questioned whether resorting to Facebook may be a means to relieve depressive symptoms.

Our primary objective was to study the prevalence of major depression in medical students attending the only state university in Lebanon and its correlation with these students resorting to Facebook.

Participants were medical students from the second up to the fifth year who attended the Faculty of Medicine at the Lebanese University. Sixth- and seventh-year students could not be included since they were in clinical rotations in several hospitals. Taking into consideration that all students at the Faculty of Medicine had taken an advanced English course and since lectures are delivered in English, we distributed the questionnaires in English. The survey extended from February 6-25, 2014. We visited each class twice on two consecutive days in order to guarantee maximum attendance.

The survey was totally anonymous. The ethics committee at our university required that we provide clear information to the participants prior to their written consent (see Multimedia Appendices 2 and ).

Statistical analysis was conducted using IBM SPSS Statistics for Windows version 16.0 (IBM Corp). Correlation analysis between depression and resorting to Facebook, along with sociodemographics, was performed using the chi-square test for which we calculated the P value and Cramer's V value, which is the appropriate measure of association. We performed t tests to show the predisposition to depression concerning gender. Furthermore, one-way analysis of variance (ANOVA) was conducted to determine the significance of the relationship between resorting to Facebook and depression, the variability of depression among the different academic years, as well as determining a significant difference between each FbRQ category. Moreover, multiple regression analysis was conducted to examine the relationship between different variables in the study for which we calculated the R²value, which reflects the degree of correlation, and the F statistic, which tells us if the model is a good fit. Statistical significance for hypothesis testing was set at 5% (P<.05).

A total of 365 of 480 students (76.0%) took the survey; the remaining students were absent on the days of our survey. A total of 25 students out of 365 (6.8%) were excluded from the study because some did not complete the three questionnaires; 9 of these students refused to participate. Hence, we included a total of 340 students in the study. The sociodemographic characteristics of the sample are summarized in Table 1.

On the PHQ-9 scale, for a cut-off of 10, the prevalence rate of depression was 34.4% (117/340). Severity was distributed as follows: out of 117 students, 5 (4.3%) had severe depression, 32 (27.4%) had moderate depression, and 80 (68.4%) had mild depression. Suicidal thoughts in depressed students—represented by Question 9 on the PHQ-9 scale—were attributed to 27.4% (32/117) of students. Table 2 shows the multiple regression analysis conducted between Question 2 of the PHQ-9 as a dependent variable and Questions 1, 3, 4, 6, 8, 9, and 10 of the PHQ-9 and Question 2 of the FbRQ as independent variables. The R²value reached 63.5%, the F statistic showed that the model was a good fit, and this regression was a good model (P<.001). As plotted in Figure 1 and as seen in Table 3 , after a one-way ANOVA, depression scores averaged 8.7, 9.0, 7.5, and 6.2, and depression reached 41.5%, 46.9%, 31.6%, and 21.4% in the second through the fifth year, respectively. The calculated value of F was 6.1 (P<.001). Comparing the different academic year classes, the statistical significance was owed only to the relationship between year 5 with respect to second- and third-year students (P<.05).

Among sociodemographics, the chi-square test showed a significant relationship between depression and gender only (P=.007), as shown in Table 4. In addition, Cramer's V value was significant (P=.007) and equaled .189. To determine which gender was more predisposed to depression, a t test was performed and showed that females had higher rates of depression than males with a factor of 1.8, as shown in Table 5.

Table 6 is a result of the one-way ANOVA showing that, out of 340 students, resorting to Facebook was high in 61 students (17.9%), moderate in 149 students (43.8%), and low in 130 students (38.2%). It was shown that there is a positive and significant relationship between the depression mean and the categories related to resorting to Facebook, whereby the depression mean was 6.9 in the low FbRQ group, 8.4 in the moderate group, and 9.0 in the high group. The F value was 5.843 (P=.003). Comparing the different FbRQ groups, the relationship was significant between the low and moderate groups (P=.03), and between the low and high groups (P=.01); however, it was not significant between the moderate and high groups.

There was no statistically significant relationship between the sociodemographic parameters studied and the FbRQ results as determined by the chi-square test. However, a regression analysis was performed locking the PHQ-9 score as a dependent variable and all the Facebook features as the independent ones. The R²equaled 25.3%, the model was a good fit, and the regression was a good model (P<.001). Table 7 shows the different ratios and the significance of each independent variable.

The current prevalence of depression in our sample (34.4%) is slightly higher than what is reported in the literature, which is around 30% [1,2]. Studies on Question 9 of the PHQ-9 showed the suicidal ideation in the studied sample to be lower than in previous reports (23%) [20]. Nevertheless, it is always considered high since a significant proportion of people reporting suicidal ideation accounted for half of fatal or nonfatal suicide attempts [20]. Multiple regression analyses showed the question of depressive mood (Question 2), locked as the dependent variable, to correlate highly with other PHQ-9 items; hence, reports of feeling down or depressed by the students was explained 63.5% of the time by all the independent variables studied. Among these variables, the problems of interest in doing things, fatigue, regarding one's self as a failure, and functionality were positively and significantly correlated and, thus, each 1% increase in these problems will increase feelings of depression by a factor of 0.19%, 0.15%, 0.1%, and 0.28%, respectively. On the other hand, a 1% increase in resorting to the check-in Facebook feature was associated with a 0.31% increase in feeling depressed, as shown in Table 2. Moreover, depression varied significantly among the different academic year classes and it peaked in the third year, which can be explained by the female predominance, then decreased gradually to a minimum PHQ-9 average of 6.2 in the fifth year. Female predominance was observed with a factor of 1.8, which is consistent with other reports [6,14,21]. From Cramer's V value of .189, we can conclude that this significant relationship was low to moderate. Finally, our results regarding family history as a predisposing factor is not consistent with the literature [6].

As a cross-sectional study, our results were confined to comparing, and not eliciting, a cause-and-effect relationship between depression improvement and resorting to Facebook. The standardized PHQ-9 questionnaire had not been validated in the Lebanese population during our survey. On the other hand, the FbRQ is a novel tool that we developed due to the lack of an equivalent standardized questionnaire; its point of weakness lies in not using a continuous variable scale.

Depression was shown to be highly prevalent among medical students of the Lebanese University. Moreover, depressed Facebook users were more likely than nondepressed users to resort to Facebook; the subjective report by the FbRQ allows us to hypothesize that Facebook may be a means to alleviate depressive symptoms. Specifically, among all the Facebook features, like, check-in, and add friend may be of therapeutic significance. These results should stimulate health care providers to question the high depression rate in this student population and to acknowledge the possible role of Facebook in a multidisciplinary treatment strategy to relieve depression.