INTRODUCTION
Sleep is a fundamental element crucial for overall well-being, cognitive functionality, and emotional regulation. However, the demanding nature of medical education often leads to sleep disturbance such as insufficient sleep duration, poor sleep quality, and irregular sleep schedules for medical students. Sleep disturbance criteria include sleep onset delay of at least 30 minutes, total sleep duration of less than or equal to 6.5 hours [1], or sleep efficiency of less or 85% [2]. A systematic review by Rao et al. [3] found that medical students have a higher prevalence of poor sleep quality compared to university students and older populations. The efficacy rate was observed to be 39.5% in the study conducted by Correa et al. [4] in Brazil in 2017. On the other hand, the study by Al-Khani et al. [5] in 2019 showed that this number was 63.2%. The sleep efficiency demonstrated by students in the health sciences group in Vietnam was also found to be within the mentioned range, as the studies of Quynh et al. [6] and Linh et al. [7] respectively showed the rates to be 59.1% and 49.4%.
Sleep disturbance can exacerbate the existing stressors and challenges faced by medical students, leading to a cascade of negative effects on their mental health, especially depression. Depression is characterized by feelings of sadness, irritability, and cognitive changes that impact a person’s functioning for at least 2 weeks. Sleep disturbance is an early sign that can help identify depression, along with depressed mood and loss of interest. According to a study by Al-Khani et al. [5] in Saudi Arabia, the rate of depression among students was 42%. Barman et al. [8] found the rate to be 24% in India. In Vietnam, the prevalence of depression among students was reported to range from 28.8% to 43.2% in studies conducted by various researchers [9-11].
These rates demonstrated an alarming situation since depression and sleep disturbance can lead to behavioral and cognitive disorders that might affect students’ mental health and academic performance. According to the research conducted by Hieu et al. [12] at Pham Ngoc Thach University of Medicine, academic and social pressures significantly influence the sleep quality of students. The research suggests that the higher the frequency of these pressures, the greater the likelihood of experiencing poor sleep quality. Similarly, a study carried out at the University of Medicine and Pharmacy in Ho Chi Minh City discovered that 6th-year students exhibit lower sleep quality compared to 4th-year students, which is attributed to exam stress, graduation concerns, and future job anxieties [6]. Notably, the School of Medicine at Vietnam National University has yet to investigate the sleep quality and mental well-being of its students, underscoring the urgent need for such research. Therefore, it is necessary to have a screening tool to help assess and detect sleep disturbance, especially for students at risk of depression, to help students become aware of their status and adjust their lifestyle accordingly.
Previous studies have translated the Pittsburgh Sleep Quality Index (PSQI) scale into Vietnamese for use on various populations. However, there is a lack of research evaluating the reliability and cut-off point of the scale specifically for health science students. Therefore, this study aims to assess the reliability and determine the cut-off point of the PSQI scale in evaluating sleep quality among health science students, particularly those who are at risk of depression. The sensitivity and specificity of the PSQI scale will be evaluated using the Beck Depression Inventory-II (BDI-II) scale as reference.
METHODS
Research Design
This study applies cross-sectional design. It was conducted at the School of Medicine, Vietnam National University Ho Chi Minh City (VNU-HCM) from September 9 to 30, 2021. The study’s inclusion criteria comprised the entire population of students enrolled at the School of Medicine, VNU-HCM, while the exclusion criteria were applied to students who did not provide informed consent. The final sample consisted of 906 individuals, yielding a response rate of 90%.
Research Instruments
The instrument used to collect data is a self-administered questionnaire. The questionnaire included a total of 34 questions and was completed within 15 minutes. The questions were separated into three parts, they are described as follows: 1) Four questions asked about background characteristics, which were age, gender, school year, and major. 2) The PSQI questionnaire translated into Vietnamese for this study was based on the research conducted by Ngoc et al. [13] in 2014. Nineteen questions of the PSQI scale were included to evaluate the 7 components of sleep, which are subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbance, use of sleep medication, and daytime dysfunction. Each component was scored from 0 to 3 and the total score ranged from 0 to 21. 3) Twenty-one questions of BDI-II scale were included to assess the most common manifestations of depression recognized by the American Psychiatric Association. The BDI-II consists of concise questions that comprehensively evaluate common depressive symptoms according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria. It takes approximately 5 minutes to complete the questionnaire and 1 minute for scoring [14]. The Vietnamese version is recommended for health science students in Vietnam [15]. Each question was scored from 0 to 3. The total score of BDI-II ranged from 0 to 63. Participants were made to choose the choice that best describes their current status, then the total score of the 21 answers was calculated. In this case, the participants with the higher total score were more likely to exhibit major depressive disorder. The total score was categorized as follows: no depression (score of ≤13), and depression (score of >13) [14].
Data Collecting Method
Data was collected using an online questionnaire via Google Forms. The method was employed in consideration of the complicated situation caused by the COVID-19 epidemic at the time of the study.
Data Processing and Analysis
Frequency and percentage were used to describe qualitative variables (gender, school year, and major). Mean and standard deviation (SD) were used to describe the quantitative variable (age). The Cronbach’s alpha coefficient was utilized to evaluate the internal reliability, and item-rest correlation index was used to assess the correlation between components of each toolkit. The results of receiver operating characteristic (ROC) curve analysis and Youden index were employed to determine the cut-off point allowing for the optimal sensitivity and specificity in the purpose of screening students with sleep disturbance. In this study, sleep disturbance was characterized as either taking more than 30 minutes to fall asleep [16,17], sleeping for 6.5 hours or less in total [1], or having a sleep efficiency of 85% or lower [16].
RESULTS
General Characteristics of the Study Sample
The proportion of male and female participants was almost equal, with 43.71% male and 56.29% female. The largest group of participants were 4th-year students, accounting for 23.51% of the total, while the smallest group were 6th-year students, who made up only 6.73%. The remaining participants were distributed among the other years, with 18.10%, 19.65%, 18.65%, and 13.36% for 1st year, 2nd year, 3rd year, and 5th year, respectively. In terms of the distribution of participants by their majors, the faculty of medicine had the highest proportion of students (62.8%), followed by pharmacy (27.37%), and the faculty of dentistry had the lowest number of participants (9.82%). The PSQI scale’s mean score was 4.67 ± 2.99, in which the average score of its 7 components ranged from 0.07 to 1.06 (SD = 0.35–1.03). The BDI-II scale had a mean score of 4.69 and an SD of 3.00 (Table 1).
Reliability of the Vietnamese Version of the PSQI Scale
The Vietnamese version of the PSQI had a Cronbach’s alpha coefficient of 0.68, indicating acceptable internal consistency. Additionally, the correlation of each item of the scale ranged from 0.21 to 0.60, suggesting moderate to strong associations with the overall construct (Table 2).
The Cut-Off Points of the Vietnamese Version of the PSQI Scale
The area under the ROC curve was 0.8498, which showed that the Vietnamese version of the PSQI scale was effective and valuable in assessing of the manifestation of sleeping disorders in students (Fig. 1).
Analysis of the ROC curve and Youden’s index indicated the cut-off point of ≥5 resulting in the best sensitivity and specificity in the assessment of sleep disorder in students (Table 3).
Correlation between the PSQI Scale and the BDI-II Scale
There were mild correlations between the PSQI scale, and the BDI-II scale in both the mean total scores and the scores of each scale’s components, with statistical significance (p < 0.01). The correlation coefficients (r) for these values are presented in Table 4.
The result indicated that students without depression had better sleep quality compared to those with the condition (the BDI-II scale was used in the assessment of depression). The number of participants found to be at risk of depression (BDI-II > 13 points) was 289, and their mean PSQI total score was 7.03 ± 3.02. This value was significantly higher compared to the average mean score of the non-depressed student (BDI-II ≤ 13 points), which consisted of 617 participants with an average PSQI total score of 3.59 ± 2.28 (Table 5).
DISCUSSION
The study’s results showed that the Vietnamese version of the PSQI scale had a Cronbach’s alpha of 0.68 (≥0.6), which was consistent with the value of the other versions of the scale in the other studies, such as the study of Suleiman and Yates [18] in Arabic (the obtained Cronbach’s alpha of the study was 0.65). This study found acceptable internal reliability of the Vietnamese version of PSQI in students. Correlations between the 7 PSQI components ranged from 0.21–0.60. Only the sleep medication component had an item-rest correlation <0.3. However, no modifications were made given the PSQI is a standardized international instrument with established scoring. Omitting sections could alter sleep disturbance classification. The study aimed to assess the PSQI’s reliability in Vietnamese students, not modify content. Thus, the full validated Vietnamese version of PSQI was retained to evaluate sleep quality in this population.
The ROC curve analysis revealed that a cut-off point ≥5 on the PSQI scale exhibited a sensitivity rate of 78.08% and a specificity rate of 75.60%. This cut-off point was found to be the optimal threshold to distinguish students with and without sleep disturbance for our study. The obtained value is similar to the cut-off point found in the other translated versions of the scales ranging from ≥5 to ≥6.5. Specifically, the cut-off point of ≥5 of the Vietnamese version exhibited lower sensitivity and specificity values compared to the cut-off point of ≥5 in the original English version from the study of Buysse et al. [19] (sensitivity of 89.6% and specificity of 86.5%), the cut-off point ≥5.5 of the Japanese version in the study by Doi et al. [20] (sensitivity of 80% and specificity of 86.6%), and the cut-off point of ≥6 of the English version in the study by Manzar et al. [21] conducted in India (sensitivity of 75% and specificity of 88.9%). However, our study achieved higher sensitivity and specificity rates compared to the cut-off point of >6.5 in the Indonesian version in Setyowati and Chung’s study (sensitivity of 66% and specificity of 74%) [22]. Overall, these results illustrate the effectiveness of the PSQI scale in the assessment of health science students’ sleep quality.
Furthermore, our study is the first to evaluate the correlation between the results of the PSQI and BDI-II scales to investigate the effectiveness measuring sleep quality among health science students in Vietnam. The total PSQI score in the non-depressed group of students was significantly lower compared to that of students with depression and poor sleep quality, consistent with previous research [22-24]. Moreover, all 7 PSQI component scores significantly correlated with total BDI-II score (r ranging from 0.004 to 0.06 with p < 0.05), supporting our hypothesis of an association between depression symptoms and impaired sleep quality. These findings advocate the PSQI as a valuable instrument to evaluate and identify early signs of sleep disturbances in Vietnamese health science students at depression risk, enabling timely preventative interventions to assist these students. In contrast to the previous studies [22], our study showed that the total PSQI score was statistically different between male and female students, which indicated a factor that can be explored further in future studies.
One limitation of this study is that data on emotional and sleep disturbance levels were collected via self-report through an online survey. As such, these results reflect subjective perceptions of disturbance levels rather than objective measurements. Additionally, the sample recruited was limited to medical students at one university in Vietnam. The homogeneity of this sample limits generalizability and precludes comparisons between health sciences students at different academic institutions. Future studies would benefit from using standardized instruments to obtain more objective assessments of emotional and sleep disturbances, as well as recruiting a larger and more diverse sample from multiple universities to improve generalizability of findings.
In conclusion, the results of this study show that the PSQI scale demonstrated high reliability in the assessment of sleep quality in health science students. A cut-off point of ≥5 indicates students with poor sleep quality. The PSQI scale as such can be employed to help detect early signs of sleep disturbance among students that are at risk of depression.
