1. Introduction
Adequate sleep duration has a critical role in promoting optimal physical health, immune function, mental health, and cognition [
1]. According to consensus recommendations of the National Sleep Foundation and the American Academy of Sleep Medicine and Sleep Research Society guidelines, young adults should sleep 7 to 9 hours per night on a regular basis for optimal sleep health [
2,
3]. Sleep duration shorter than the recommended is generally associated with adverse health-related outcomes, including poor attention, depression, obesity, and cardiovascular disease [
4].
University students are recognized as a population group, particularly affected by short sleep duration and sleep disturbances [
5,
6]. Indeed, a young adult's life is going through numerous transitions during their time in university in which students have reduced parental support, increased stress from academic loads and lifestyle, and irregularities in the sleep-wake cycle, all resulting in shortened and delayed sleep [
7]. These sleep disturbances are especially worrisome as they have an adverse effect on mental and physical health and on cognitive skills, which are vital for students' daily functioning and academic success [
8,
9,
10]. Importantly, sleep quality in this population has been identified as the strongest predictor of well-being compared with physical activity, depression, and use of tobacco [
11]. Attending university is also characterized by fatigue, which is often overlooked and contributes also to poor sleep [
12].
Recent studies, representing different socio-cultural regions mainly form US and China, have shown that sleep disturbances and dissatisfaction are particularly prevalent among university students, affecting 30 to 70% of this population [
5,
14,
15,
16,
17,
18,
19,
20,
21,
22,
23]. During times of theorized greater stress, such as exams periods, students seem to demonstrate even worse sleep quality and less sleep than recommended [
24]. However, these findings may not accurately represent sleep disorders rates among university students attending universities in Europe, which has distinct features associated to living arrangements, educational expenses, the application process, and facilities [
25].
In Greece, research on sleep in university students remain scarce [
26] and is mainly derived from studies during the COVID-19 pandemic [
26,
27,
28]. Little is also known about how students’ sleep patterns change before and during the academic exam period [
29,
30]. More specifically, there are limited data indicating impaired sleep quality and quantity during exam periods, that might lead to impaired daytime functioning [
29,
30]. Therefore, the aim of our study was to assess changes in Greek university students’ sleep quality and fatigue before (low stress) and during an academic exam (high stress) period and to identify possible associated factors.
2. Materials and Methods
2.1. Study Setting and Participants.
A Web-based survey was conducted by university students of 20 different Tertiary Institutions (Medical/Health, Physics, Educational Sciences, Technical, Social sciences, Economic etc) in Crete, Greece, across 2 periods, at the beginning of the semester and on examination period in the academic year of 2018-2019 (before COVID-19 lockdown). The process of identifying and recruiting student participants involved two phases. The initial phase encompassed enlisting university professors and conveying the objectives of the current study on social media platforms. Subsequently, with the consent of their professors, university’s public e-mail board and social media platforms were used to send online anonymous survey links to students.
The students were asked to answer questions about demographics, sleep habits, exercise habits, caffeine, tobacco and alcohol use, hours of technology use (cell phones, tablets, laptop computers), obstructive sleep apnea (OSA) and insomnia symptoms, excessive daytime sleepiness, subjective sleep quality (using the Pittsburgh Sleep Quality Index - PSQI) and fatigue (Fatigue severity scale – FSS).
All procedures were approved by the University of Crete Research Ethics Committee and all participants gave their informed consent to participate prior to both survey administrations, using a digital form.
2.2. Study Tools and Outcomes
PSQI
A self-reported assessment of sleep was determined using the PSQI questionnaire, which is a standard instrument that has been validated as differentiating poor from good sleep. The higher the score, the greater the negative impact on sleep quality. A global score of 6 or higher indicates poor sleep [
31].
FSS
In this scale, individuals rate their agreement (range, 1–7) with 9 statements concerning the severity, frequency, and impact of fatigue on daily life (physical functioning, exercise and work, and family or social life). A total score of less than 36 is considered normal. A score above that limit (maximum score 63) is suggestive of a significant negative impact of fatigue on daily life activities [
32].
Outcomes
The primary outcome of the study was to compare the absolute change in students’ global PSQI and FSS scores between a time of low stress (at the beginning of the semester) and a time of perceived high stress (academic exam period). For secondary outcomes, we identified potential factors associated with these changes in sleep quality and fatigue among these students.
2.3. Statistical Analysis
Results are presented as mean ± standard deviation (SD) for continuous variables if normally distributed and as median (25th-75th percentile) if not. Qualitative variables are presented as absolute number (percentage). To compare changes from the beginning of semester to exam period, the paired samples t-test (for normally distributed data) and the Wilcoxon Signed Rank test (for non-normally distributed data) were used. Changes of continuous variables from baseline to follow-up were defined as baseline minus follow-up values. Factors associated with poor sleep quality and fatigue were analyzed with bivariate logistic regression after adjustment for various potential explanatory variables, including age, gender, BMI, smoking status, presence of chronic disease, use of alcohol, caffeine, physical exercise, work, and use of technology. Multivariate linear regression analysis was used to examine any association of the previous potential confounders with changes in questionnaires scores (PSQI and FSS) at exam period. We checked multicollinearity among the predictors using collinearity statistics to ensure that collinearity between predictor variables was in the acceptable range as indicated by the tolerance value variance inflation factor. Results were considered significant when p values were < 0.05. Data were analyzed using SPSS software (version 25, SPSS Inc, Chicago, IL).
3. Results
3.1. Study population
A total of 940 university students completed the questionnaire, of whom 60% were females (
Table 1). Ages ranged from 17 to 48 years with a mean age of 21 years. Most of the participants were from medical/health (36%) and Physics (18%) universities. The majority of the participants were in the fourth year of education (50%) followed by the second (18%) and third (17%) years.
Further sample characteristics including exercise habits, caffeine, tobacco and alcohol use, and hours of technology use are presented in
Table 2.
3.2. Sleep patterns
Sleep patterns of the surveyed students are depicted in
Table 3, where most students (75%) report late bedtimes. No significant difference was noted in the number of students’ reported hours of sleep among all academic years. Only 29 (3%) of the students reported frequently taking prescription medicine for insomnia.
According to the PSQI results, 554 (59%) out of 940 university students were classified as poor sleepers, a rate that remained similar across academic years (p=0.299). On average, the PSQI global score of our sample was 6.1 ± 1.8 (Range: 1 to 14), which is above the cutoff for good sleepers (≤5), indicating that sleep quality was impaired. Impaired sleep quality was independently associated with female gender (OR = 1.524, 95% CI 1.086-2.138; p=0.015), younger age (OR = 0.922, 95% CI 0.866-0.981; p=0.011), high alcohol consumption (OR = 2.095, 95% CI 1.023-4.290; p=0.043), lack of physical activity (OR = 0.566, 95% CI 0.338-0.946; p=0.030), presence of a chronic disease (OR = 2.695, 95% CI 1.542-4.711; p<0.001), depressive symptoms (OR = 3.232, 95% CI 2.311-4.519; p<0.001), sleepiness (OR = 9.893, 95% CI 5.163-18.956; p<0.001), and fatigue (FSS≥36) (OR = 1.550, 95% CI 1.094-2.195; p=0.014).
3.3. Daytime functioning
On average, the FSS total score of our sample was 32.7 ± 11.4 (Range: 9 to 62) and 365 out of 940 university students (39%) reported suffering from fatigue. The prevalence of fatigue remained similar across academic years (p=0.754). Fatigue was independently associated with lack of physical activity (OR = 0.299, 95% CI 0.187-0.478; p<0.001), presence of a chronic disease (OR = 1.619, 95% CI 1.080-2.426; p=0.020), depressive (OR = 3.498, 95% CI 2.646-4.625; p<0.001), and OSA symptoms (OR = 1.990, 95% CI 1.154-3.432; p=0.013), sleepiness (OR = 1.911, 95% CI 1.333-2.739; p<0.001), and impaired sleep quality (PSQI≥6) (OR = 1.682, 95% CI 1.135-2.495; p=0.010).
Regarding the frequency of excessive daytime sleepiness (2 times or more per week), younger age (OR = 0.906, 95% CI 0.824-0.996; p=0.040), presence of a chronic disease (OR = 2.071, 95% CI 1.145-3.747; p=0.016), depressive symptoms (OR = 1.907, 95% CI 1.205-3.019; p=0.006), presence of OSA symptoms (OR = 2.299, 95% CI 1.224-4.316; p=0.010) and impaired sleep quality (OR = 14.565, 95% CI 6.166-34.042; p<0.001) all significantly associated with more frequent reported sleepiness among students.
3.4. Changes in sleep quality and fatigue during the exam period
During the exam period average sleep duration was significantly reduced (6.9 vs 7.4, p<0.001), 134 out of 242 (55%) smokers increased smoking, 489 out of 940 (52%) students increased coffee/energy drink consumption, 514 out of 940 (55%) decreased alcohol consumption and 474 (50%) decreased exercise. Insomnia symptoms, daytime sleepiness, and depressive symptoms were also more frequently reported in the exam period (
Table 3).
PSQI global score was significantly elevated at exam period compared to pre-exam period (8.9 vs 6.1, p<0.001). Notably, all sub-scales of the PSQI contributed to the decline in sleep quality at Exam Period (all p<0.05). The increase of PSQI score (2.9 ± 1.6) which was similar in all years of education (p=0.109) was independently associated with age (β=0.111, p=0.011), presence of chronic disease (β=0.914, p=0.006), worsening of FSS (β=0.048, p<0.001) depressive symptoms (β=0.459, p=0.001) and sleepiness (β=0.601, p<0.001). Furthermore, the prevalence of poor sleep quality (PSQI global score>5) was also higher (98% vs 59%) in this period.
FSS score was significantly elevated at exam period compared to pre-exam period (36.9 vs 32.7, p<0.001). Notably, all sub-scales of the FSS in the exam period were significantly elevated compared to pre-exam period (all p<0.001). This increase of FSS score (3.2 ± 6.4) was independently associated with female gender (β=1.658, p<0.001), younger age, (β=0.198, p=0.010), increase in smoking (β=1.7, p=0.029), coffee/energy drinks consumption (β=1.988, p<0.001), decreased levels of physical exercise (β=1.660, p<0.001), depressive symptoms (β=2.526, p<0.001), sleepiness (β=1867, p<0.001) and worsening of PSQI (β=0.754, p<0.001). The prevalence of fatigue (FSS>36) was also higher (50% vs 39%, p<0.001) in this period.
4. Discussion
The results of our study, obtained from students of different Tertiary Institutions, showed that sleep quality and fatigue are frequent at the beginning of the semester and deteriorate during academic exam periods. Furthermore, this study provides a wide understanding of the factors that influence these observations during the demanding exam period. These factors include younger age, female gender, presence of chronic disease, decreased levels of physical exercise, depressive symptoms, increase in smoking and coffee/energy drinks consumption.
University students are recognized as a population group particularly affected by sleep problems and fatigue. In the present study, at the beginning of semester when stress would theoretically be low, about 60% of participants reported poor sleep quality with an average of 7.4 hours of sleep per night, indicating also sleep deprivation among students. Increased autonomy, irregularities in the sleep-wake cycle, resulting from academic and social pressure at an age when the circadian rhythm is delayed and the central nervous system is still maturing, along with lifestyle and physical inactivity, increased caffeine and alcohol consumption may explain the poor sleep quality in this population [
33,
34,
35,
36]. Insomnia symptoms were also frequently reported, with similar [
37] or higher rates compared to previous studies [
38]. Potential factors that could contribute to poor sleep quality in our study were identified, such as being female, younger age, consuming high amounts of alcohol, not engaging in physical activity, and having a chronic disease, consistent with previous research [
39,
40,
41,
42,
43,
44]. Importantly, presence of depressive symptoms was associated with an approximately threefold increase in the risk of reporting poor sleep. This association is particularly worrisome for students' mental health, as sleep disturbances have been identified as a risk factor not only for depression but also for other psychiatric disorders [
45,
46].
Academic responsibilities during university may also lead to fatigue, with concerning prevalence rates reported [
47]. In our study we found a fatigue prevalence of 39% in our university students at the beginning of semester, which is lower compared to a recent study (59.5%) [
48] but higher compared to previous studies (16.7% and 13.7%) [
49,
50]. However, as a considerable proportion of our sample is from demanding disciplines like Medical/Health universities, a substantial prevalence of fatigue is anticipated. The level of fatigue was associated with the presence of chronic disease, depressive and OSA symptoms. At the same time, an inverse relationship was noted between fatigue and physical activity, in line with previous studies [
48,
52]. Consistent with previous research [
48,
52], poor sleep quality was also associated with fatigue, which could potentially have a more pronounced effect on university students’ academic performance and mental health.
During periods of high stress, such as academic exam periods, university students are particularly affected by lower sleep duration and poor sleep quality [
29,
30]. In our study, shorter night sleeps on exams were reported compared with students’ typical routines and PSQI scores were significantly worse, with a prevalence of poor sleep quality increasing from 59 to 98%. Although there are several research in the literature on the topic of sleep quality in university students, few studies have examined changes in university students’ sleep quality and fatigue across multiple time points (before and during an academic exam) [
24,
29,
30,
53]. Based on recent research university students exhibited negative outcomes, like deteriorating sleep quality [
24,
29,
30], daytime dysfunction [
29] and declining academic performance [
30] during high stress periods; however, the number of participants in previous studies was low (31 to 252 participants) [
24,
29,
30]. In support of this, students who adhered to a regular sleep-wake routine in the 2 weeks preceding their end-of-semester exam tended to perform better academically [
53]. Notably, in our cohort we highlighted a significant relationship between worsening of sleep quality and age, chronic disease and depressive symptoms.
Further, we found that students who reported worsening sleep quality also had a higher risk of fatigue, the prevalence of which was also significantly elevated during exam period, from 39% to 50%. Female gender, age, depressive symptoms, increase in smoking, coffee/energy drinks consumption and lower physical activity levels were significant predictors of fatigue during exam periods. However, although level of fatigue has been found previously to be associated with female gender, caffeine intake, physical activity, and sleep duration [
48,
54,
55], these relationships in university students during exam periods are not documented.
Considering that sleep disruption and fatigue potentially affect university students’ academic performance, getting sufficient quality sleep, especially before exams, may be associated with better academic performance and lower odds of course failures [
56]. Importantly, in a recent study, evaluating cognitive performance in medical students with visual and auditory evoked potentials, sleep deprivation during previous 2-3 nights before exam session and psychosomatic fatigue were found to be closely related to cognitive abilities, which in turn adversely affected academic achievements. [
57]. To support this, it has been shown that university students who reported sleeping less than seven hours per night had a higher risk of exhaustion and low professional efficacy [
30,
58]. It is also worth noting that our results are in line with previous studies regarding significant association of poor sleep with excessive daytime sleepiness and support evidence for subsequent associated poor academic achievements [
56,
59]. Therefore, universities should attempt to raise students’ awareness of the possible effects of their lifestyle and sleeping habits on their academic performance. Such guidance along with recommendations for establishing good sleep hygiene would be valuable, particularly for first-year students and may enable students to improve sleep habits, not only in the university years but also in their later professional careers.
Our study has some limitations that deserve comments. First, the cross-sectional design of the current study precludes causal interpretations of the results. Second, as this was a self-administered survey, it was prone to recall bias. Future studies including objective methods such as actigraphy or qualitative methods using semi-structured interviews and sleep diaries could obtain information that may have been overlooked due to the use of self-reported measures in the current study. Third, sleep quality and fatigue evaluation were taken from students studying in different institutions; therefore, factors such as academic demand and level of difficulty may potentially affect our results. Lastly, taking into account that the participants included lived in Crete (south of Greece), sleep quality and fatigue levels may be different in other regions of Greece. Consequently, the generalization of our conclusions to all Greek university students should be made with caution.