Changes in sleep and the prevalence of probable insomnia in undergraduate university students over the course of the COVID-19 pandemic: findings from the U-Flourish cohort study
Department of Public Health Sciences, Queen's University, Canada
Department of Public Health Sciences, Queen's University, Canada; and Department of Health Sciences, Brock University, Canada
Dalla Lana School of Public Health, University of Toronto, Canada; and Li Ka Shing Knowledge Institute, St Michael's Hospital, Toronto, Canada
Big Health Inc., San Francisco, USA; and Sir Jules Thorn Sleep and Circadian Neuroscience Institute (SCNi), Nuffield Department of Clinical Neurosciences, University of Oxford, UK
Department of Biology, Queen's University, Canada
Department of Psychiatry, Queen's University, Canada; and Department of Psychiatry, University of Oxford, UK
Abstract
Background
Sleep problems associated with poor mental health and academic outcomes may have been exacerbated by the COVID-19 pandemic.
Aims
To describe sleep in undergraduate students during the COVID-19 pandemic.
Method
This longitudinal analysis included data from 9523 students over 4 years (2018–2022), associated with different pandemic phases. Students completed a biannual survey assessing risk factors, mental health symptoms and lifestyle, using validated measures. Sleep was assessed with the Sleep Condition Indicator (SCI-8). Propensity weights and multivariable log-binomial regressions were used to compare sleep in four successive first-year cohorts. Linear mixed-effects models were used to examine changes in sleep over academic semesters and years.
Results
There was an overall decrease in average SCI-8 scores, indicating worsening sleep across academic years (average change −0.42 per year; P-trend < 0.001), and an increase in probable insomnia at university entry (range 18.1–29.7%; P-trend < 0.001) before and up to the peak of the pandemic. Sleep improved somewhat in autumn 2021, when restrictions loosened. Students commonly reported daytime sleep problems, including mood, energy, relationships (36–48%) and concentration, productivity, and daytime sleepiness (54–66%). There was a consistent pattern of worsening sleep over the academic year. Probable insomnia was associated with increased cannabis use and passive screen time, and reduced recreation and exercise.
Conclusions
Sleep difficulties are common and persistent in students, were amplified by the pandemic and worsen over the academic year. Given the importance of sleep for well-being and academic success, a preventive focus on sleep hygiene, healthy lifestyle and low-intensity sleep interventions seems justified.
Untitled section
Keywords: Sleep–wake disorders, epidemiology, university students, COVID-19 restrictions, health promotion and early intervention
Article notes
Untitled section
Received 2022 Dec 20; Revised 2023 Aug 16; Accepted 2023 Sep 30; Collection date 2023 Nov.
Background
In the transition to university, students assume increased responsibility for managing their schedules, lifestyle and sleep habits. Heightened stress related to academic pressures, making new friends, living away from home and increased substance use can affect sleep patterns and contribute to poor sleep quality in post-secondary students.1 In 2018, approximately 18% of first-year undergraduates at Queen's University, Canada (12% of men and 20% of women) reported sleep difficulties indicative of probable insomnia at school entry.2 Insomnia in university students is associated with reduced academic performance,3,4 and common mental health concerns such as anxiety and depression.5,6
The COVID-19 pandemic had a particularly disruptive and generally negative effect on university student well-being and mental health.7,8 Shifts to remote learning, changes in the frequency and mode of examinations, increased uncertainty about current and future prospects, and social isolation from peers have been associated with increased levels of stress and anxiety.7,8 During the peak of pandemic-related restrictions, remote learning and asynchronous lectures put more responsibility on students to regulate their schedules and study habits, not uncommonly under non-ideal learning conditions. Therefore, studying under pandemic conditions may have further increased the risk of sleep difficulties in university students.
A large international survey study including over 25 000 adults across 14 countries reported a 10% or more increase in sleep difficulties (e.g. onset and maintenance problems) and related daytime problems (i.e. fatigue and daytime sleepiness) during the COVID-19 pandemic from May to August 2020. Social restrictions and financial concerns were significantly associated with reported sleep difficulties. Although sleep worsening was reported across all countries, poor sleep quality was particularly high in this study in Canada and the UK (four-fold and three-fold increases, respectively). It was speculated that this may reflect a pandemic stage effect, in that more sleep problems were reported in countries where the pandemic and related public health restrictions were escalating compared with those where it was more stable.9
Present study
The U-Flourish Student Well-Being Survey study launched in 2018 at Queen's University in Canada.2 This longitudinal study engaged robust cohorts of first-year undergraduate students to complete biannual electronic surveys at entry and completion of each academic year. Therefore, we had a unique opportunity to describe changes in sleep over varying stages of the COVID-19 pandemic from the autumn of 2018 (pre-pandemic) to the spring of 2022 (return to in-person classes and relaxed pandemic restrictions). We hypothesised that sleep difficulties would be common at entry to university, increase over the academic year and be most pronounced during the peak of the pandemic (associated with remote learning and the most severe pandemic-related restrictions). Our specific study objectives were to (a) describe sleep in successive student cohorts at entry to university over different stages of the pandemic, (b) examine changes in student sleep over the academic year and different stages of the pandemic, and (c) explore associations between modifiable risk factors and screening positive for probable insomnia. To our knowledge, this is the first large-scale study to describe sleep patterns in undergraduate university students, including documentation of trends in sleep over the course of the COVID-19 pandemic.
Method
Data source
Data from this study come from four consecutive cohorts of the longitudinal U-Flourish Student Well-being and Academic Success Survey study described in detail elsewhere.2,10 Since the autumn of 2018, all first-year undergraduates entering Queen's University have been invited to complete online surveys at school entry (mid-September) and the end of the academic year (mid-March), before to winter examinations. Follow-up surveys were repeated at the same time points each successive year. The survey collects data on demographics, mental health risk factors, current lifestyle and indicators of mental health and well-being, using validated measures. Participation incentives and a student-led engagement campaign that included in-class presentations, social media posts and the use of working booths at university events and in campus residence buildings, were used to encourage student participation. Before the COVID-19 pandemic, the baseline response rates for cohort 1 (n = 3029) and cohort 2 (n = 2949) were 58%. During the pandemic, when in-person engagement activities were not possible, the response rate dropped to 23.3% (cohort 3, n = 1472). In cohort 4, students started the year in-person and the response rate was 35% (n = 1991). The response rates on the first spring follow-up survey by cohort were 64.3%, 37.4%, 33.2% and 41.1%, respectively (Supplementary Fig. 1 available at https://doi.org/10.1192/bjo.2023.597). All students provided informed consent to access and complete the survey.10
Study variables
Year of study
To describe sleep in students beginning university over the course of the pandemic, we compared the sleep of four successive cohorts of first-year students at school entry (autumn 2018 to autumn 2021). To examine changes in sleep over the academic year and throughout the pandemic, we combined cohorts, using all available data to compare sleep across academic semesters and years. Each autumn term and academic year corresponded to a unique stage of the pandemic, as described below.
Pre-pandemic conditions (cohort 1, 2018–2019)
Students entered university in the autumn of 2018 and completed their first academic year in the spring of 2019, under normal pre-pandemic conditions.
Transitional conditions (cohort 2, 2019–2020)
Students entered university in the autumn of 2019, under normal pre-pandemic conditions. The World Health Organization declared COVID-19 a pandemic on 11 March 2020, and Queen's University suspended undergraduate classes on 16 March before moving online on 23 March. The U-Flourish Spring survey was sent out on the first day courses moved online, and was therefore completed during a time of great uncertainty when the full impact of the pandemic was not yet realised.
Peak pandemic conditions (cohort 3, 2020–2021)
Students started their undergraduate studies online. In the autumn of 2020, Queen's University campus was closed and the 2020–2021 academic programming was completed fully online, under pandemic-related public health restrictions such as social distancing and business and recreational facility closures. During this time, Queen's University housing residences operated at half capacity.
Hybrid pandemic conditions (cohort 4, 2021–2022)
Students started the 2021–2022 academic year on campus with the following pandemic restrictions in place: mandatory masking, self-screening, and vaccination and social distancing safety protocols. The year was a hybrid of in-person and online learning, as the university reacted to the evolving pandemic situation and responded to government recommendations. It started in person, but December examinations were moved online in response to the Omicron variant, and courses remained online until the end of February. Residences were operating at close to full capacity (93%), offering only single-room accommodations.
Demographic characteristics
Age in years, international status and gender identity were self-reported. Students identified their ethnicity and reported personal and family history (first-degree relatives) of a diagnosed mental disorder from standard lists. The highest level of education completed by either parent was recorded. As possible risk factors for sleep disorders, experiences of childhood physical and/or sexual abuse, and being bullied or teased by peers were measured with items from the Childhood Experience of Care and Abuse questionnaire.11 Symptoms of anxiety and depression at university entry were measured with the Generalised Anxiety Disorder 7 scale12 and the Patient Health Questionnaire 9,13 respectively. On both scales, a score of 10 or more indicates a positive result for clinically significant symptoms.
Sleep variables
The eight-item Sleep Condition Indicator (SCI-8) was used to assess sleep and identify screen positives for probable insomnia. Higher total scores (range 0–32) indicate better sleep, and a screening cut-off of ≤16 is used to identify students with probable insomnia.14,15 To describe the nature of sleep difficulties and provide context for the total SCI-8 score, individual items of the SCI-8 were dichotomised by applying cut-offs that were informed by the scoring guidelines for the scale.14 A more strict cut-off was used for three of the eight items based on the distribution of responses in our sample.
Lifestyle variables
As possible correlates of sleep disorders, frequency of binge drinking, defined as consuming five or more drinks on one occasion, and cannabis use during the past month were selected from a set of options provided. Exercise frequency in the past month was measured with the item ‘How often have you worked out, gone to the gym, or attended a fitness class?’. Students also reported how often in the past month they set aside time for self-care or recreational activities (i.e. to relax, recharge, do something you enjoy, etc.), and how much time they usually spent per day engaged in passive screen time unrelated to their studies (i.e. watching television or videos or using an application or game).
Analysis
We used SAS version 9.4 for Windows software to conduct the analyses for describing sleep in successive cohorts over different stages of the pandemic and exploring associations between risk factors and positive insomnia screening. We used R version 4.1.3 for Windows software (R Foundation, Vienna, Austria; see https://cran.r-project.org/bin/windows/base/) to conduct the analysis for changes in student sleep over the academic year and during the pandemic.
To account for differential response rates across the baseline surveys, an inverse probability weighting approach was used.16 Student cohorts 1, 3 and 4 were weighted to cohort 2 (the largest sample), using the following risk factors for sleep problems: age, gender, international status, personal and family history of a mental disorder, history of childhood physical or sexual abuse, history of childhood bullying and parental education. Sleep difficulties and probable insomnia at school entry were described by cohort, using weighted proportions and 95% confidence intervals. We similarly described positive screening for probable insomnia in male and female students (there was insufficient power to stratify by non-binary gender identity).
Multivariable log-binomial regression models were used to examine associations between cohort membership and the likelihood of sleep difficulties at school entry. To capture changes in risk over time, separate models were run comparing each successive cohort to the previous one. These associations were not significantly different in male and female students (all P-interaction > 0.10), so only results from the full sample are presented. The proportion of students missing data on key covariates and/or the SCI-8 at baseline ranged from 12.6% (cohort 2) to 24.7% (cohort 4).
Using longitudinal data from all cohorts, changes in average sleep quality (SCI-8) were examined across semesters and stages of the pandemic. First, a minimally adjusted linear mixed-effects model (LMM) was fitted, accounting for semester (survey) to assess potential non-linear patterns of mean sleep quality scores over time (Supplementary Fig. 2). Next, a full LMM was fitted, adjusting for demographic characteristics, cohort membership and categorical semester, as there was visual evidence of non-linearity. This model included interaction terms between cohort and semester, allowing each cohort to have their own mean longitudinal trajectory, as well as random intercepts and slopes for semester number to account for lack of independence between within-participant measurements. Average sleep scores were estimated for semesters and academic years with the fully adjusted LMM, and contrasts were performed to compare mean sleep scores in the spring and autumn semester of each academic year, and successive academic years representing different stages of the pandemic.17 Ten multivariate multiple imputed data-sets were used for missing participant-specific covariate data,18 and inverse probability weighting was used to account for potential bias resulting from loss to follow-up.19 Results were pooled with Rubin's rules.18
Bivariate and multivariable log-binomial regression models were used to examine the association between demographic and theoretically important risk factors measured at the beginning of the academic year and probable insomnia at completion of the year. This analysis was conducted with data from all four cohorts collected during the 2021–2022 academic year, which most closely represented the current situation. The Markov chain Monte Carlo method of multiple imputation was used to impute missing data with effect estimates generated based on ten imputed data-sets, using all available data (N = 3434).20 The imputation model included demographic characteristics, theoretically important risk factors and clinically significant insomnia symptoms at baseline and follow-up.
Ethics
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. All procedures involving human participants were approved by the Queen's University and Affiliated Teaching Hospitals Research Ethics Board (approval number HSREB PSIY-609–18).
Results
The four incoming cohorts of students surveyed in successive autumn terms are described in Table 1. The majority of students in cohort 2 (the reference cohort) self-identified as female (68%) and of White ethnicity (67%). Approximately 9% were international students, and the most common programmes of study were arts, humanities and social sciences (34%), life and physical sciences (28%), and engineering and applied science (17%). Nearly a third of students indicated a lifetime history of a mental disorder, and almost half (48%) reported a family history of a mental disorder.
| Cohort 1 autumn 2018, pre-pandemic | Cohort 2 autumn 2019, transitional pandemic | Cohort 3 autumn 2020, peak pandemic | Cohort 4 autumn 2021, hybrid pandemic | |||||
|---|---|---|---|---|---|---|---|---|
| n | (%) | n | (%) | n | (%) | n | (%) | |
| Overall | 2501 | (100) | 2575 | (100) | 1201 | (100) | 1500 | (100) |
| Age, years, mean (s.d.)a | 18.2 | (2.0) | 18.0 | (1.1) | 18.7 | (2.5) | 18.7 | (2.1) |
| Gendera | ||||||||
| Female | 1701 | (68.0) | 1754 | (68.1) | 808 | (67.3) | 1092 | (72.8) |
| Male | 782 | (31.3) | 800 | (31.1) | 377 | (31.4) | 378 | (25.2) |
| Non-binary | 18 | (0.7) | 21 | (0.8) | 16 | (1.3) | 30 | (2.0) |
| Ethnicity | ||||||||
| White | 1684 | (67.4) | 1710 | (66.5) | 726 | (60.5) | 928 | (61.9) |
| Asian | 490 | (19.6) | 434 | (16.9) | 271 | (22.6) | 297 | (19.8) |
| Black | 35 | (1.4) | 37 | (1.4) | 22 | (1.8) | 26 | (1.7) |
| Indigenous | 4 | (0.2) | 6 | (0.2) | 5 | (0.4) | 6 | (0.4) |
| Other | 35 | (1.4) | 75 | (2.9) | 43 | (3.6) | 65 | (4.3) |
| Multiple | 251 | (10.0) | 311 | (12.1) | 134 | (11.2) | 177 | (11.8) |
| Missing | 2 | 2 | 0 | 1 | ||||
| Programme of study | ||||||||
| Arts, humanities and social sciences | 866 | (34.6) | 885 | (34.4) | 451 | (37.6) | 530 | (35.4) |
| Life and physical sciences | 722 | (28.9) | 729 | (28.3) | 247 | (20.6) | 387 | (25.8) |
| Engineering and applied science | 391 | (15.6) | 438 | (17.0) | 199 | (16.6) | 158 | (10.5) |
| Business | 280 | (11.2) | 249 | (9.7) | 77 | (6.4) | 103 | (6.9) |
| Computing | 86 | (3.4) | 54 | (3.4) | 43 | (3.6) | 37 | (2.5) |
| Health sciences | / | / | 86 | (3.3) | 77 | (6.4) | 126 | (8.4) |
| Other Programmesb | 156 | (6.3) | 134 | (3.9) | 107 | (8.9) | 158 | (10.6) |
| International student, yesa | 236 | (9.4) | 226 | (8.8) | 113 | (9.4) | 78 | (5.2) |
| Childhood physical or sexual abuse, yesa | 559 | (22.4) | 473 | (18.4) | 303 | (25.2) | 467 | (31.1) |
| Childhood bullying, yesa | 540 | (21.6) | 511 | (19.8) | 247 | (20.6) | 407 | (27.1) |
| Lifetime history of mental disorder, yesa | 679 | (27.2) | 787 | (30.6) | 393 | (32.7) | 548 | (36.5) |
| Screen positive for anxiety, yesc | 816 | (32.8) | 794 | (31.1) | 474 | (39.5) | 675 | (45.1) |
| Screen positive for depression, yesd | 694 | (27.9) | 766 | (29.9) | 448 | (37.3) | 635 | (42.4) |
| Family history of mental disorder, yesa | 1076 | (43.0) | 1227 | (47.7) | 635 | (52.9) | 823 | (54.9) |
| Parental education, highest completeda | ||||||||
| Professional school or doctorate | 595 | (23.8) | 617 | (24.0) | 250 | (20.8) | 263 | (17.5) |
| Master's degree | 599 | (24.0) | 572 | (22.2) | 237 | (19.7) | 297 | (19.8) |
| Bachelor's or trades/apprenticeship | 977 | (39.1) | 1115 | (43.3) | 546 | (45.5) | 687 | (45.8) |
| Completed high School or less | 330 | (13.2) | 271 | (10.5) | 168 | (14.0) | 253 | (16.9) |
Description of sleep at university entry over the pandemic
Sleep difficulties in the four successive first-year student cohorts are described in Table 2. The proportion of students screening positive for probable insomnia increased from 18% in cohort 1 (autumn 2018) to 30% in cohort 3 (autumn 2020) (P-trend < 0.001), followed by 28% in cohort 4 (autumn 2021). Approximately a fifth of students reported poor or very poor sleep quality, which worsened during the peak of the pandemic (autumn 2020). The extent to which sleep reportedly affected mood, energy and relationships varied over the cohorts and ranged from 36 to 48%, whereas about half to two-thirds of students felt their sleep affected their concentration, productivity or ability to stay awake, and 20% indicated that sleep troubled them in general and for longer than a year. Positive screening rates for probable insomnia were higher in women than men, but the pattern of change across time was similar in the two groups (Fig. 1).
| Cohort 1 autumn 2018, pre-pandemic | Cohort 2 autumn 2019, transitional pandemic | Cohort 3 autumn 2020, peak pandemic | Cohort 4 autumn 2021, hybrid pandemic | |
|---|---|---|---|---|
| Time it takes to fall asleep | ||||
| 0–15 min | 34.6 (32.7–36.4) | 33.4 (31.5–35.2) | 27.5 (24.9–30.1) | 30.9 (28.4–33.4) |
| 16–30 min | 35.4 (33.5–37.3) | 34.6 (32.8–36.5) | 31.8 (29.1–34.5) | 34.7 (32.2–37.2) |
| 31–45 mins | 18.1 (16.5–19.6) | 16.8 (15.3–18.3) | 19.9 (17.6–22.2) | 17.6 (15.6–19.6) |
| >45 mina | 12.0 (10.7–13.3) | 15.2 (13.8–16.6) | 20.8 (18.4–23.1) | 16.8 (14.9–18.8) |
| If waking during the night, awake for… | ||||
| 0–15 min | 66.9 (65.1–68.8) | 64.3 (62.4–66.1) | 62.4 (59.5–65.2) | 63.3 (60.8–65.8) |
| 16–30 | 18.1 (16.5–19.6) | 18.8 (17.3–20.4) | 16.2 (14.1–18.4) | 18.4 (16.3–20.4) |
| 31–45 min | 8.8 (7.7–9.9) | 9.3 (8.1–10.4) | 9.8 (8.1–11.5) | 9.5 (7.9–11.0) |
| >45 minsa | 6.2 (5.3–7.2) | 7.6 (6.6–8.6) | 11.6 (9.8–13.4) | 8.9 (7.4–10.3) |
| Problems with sleep, nights per week | ||||
| 0–2 | 68.9 (67.2–70.8) | 66.5 (64.7–68.4) | 63.4 (60.6–66.2) | 68.7 (66.3–71.1) |
| 3–4a | 21.8 (20.1–23.4) | 21.7 (20.1–23.4) | 22.0 (19.6–24.4) | 19.9 (17.8–21.9) |
| 5–7 | 9.3 (8.1–10.4) | 11.7 (10.5–13.0) | 14.6 (12.5–16.6) | 11.4 (9.8–13.1) |
| Sleep quality | ||||
| Very good or good | 41.0 (39.1–43.0) | 40.1 (38.2–42.0) | 37.3 (34.5–40.1) | 37.2 (34.6–39.8) |
| Average | 41.8 (39.9–43.8) | 41.2 (39.3–43.1) | 41.4 (38.5–44.2) | 44.3 (41.6–46.9) |
| Poor or very poora | 17.1 (15.6–18.6) | 18.7 (17.2–20.2) | 21.4 (19.0–23.7) | 18.5 (16.5–20.6) |
| Duration of sleep problem | ||||
| Do not have a problem or <1 month | 60.2 (58.3–62.1) | 54.2 (52.2–56.1) | 52.9 (50.0–55.8) | 55.3 (52.7–57.9) |
| 1–2 months | 11.5 (10.3–12.8) | 14.3 (12.9–15.7) | 11.8 (9.9–13.7) | 12.1 (10.4–13.9) |
| 3–12 monthsa | 8.5 (7.4–9.6) | 8.0 (6.9–9.0) | 11.7 (9.8–13.5) | 8.4 (6.9–9.9) |
| >1 year | 19.7 (18.2–21.3) | 23.5 (21.9–25.2) | 23.6 (21.2–26.1) | 24.2 (22.0–26.4) |
| To what extent has your sleep… | ||||
| Affected your mood, energy or relationships | ||||
| A little or not at all | 64.0 (62.1–65.9) | 51.9 (49.9–53.8) | 58.2 (55.4–61.1) | 55.3 (52.7–58.0) |
| Somewhata | 21.4 (19.7–23.0) | 25.7 (24.0–27.4) | 23.0 (20.5–25.4) | 25.7 (23.4–28.0) |
| Much or very much | 14.7 (13.3–16.1) | 22.4 (20.8–24.0) | 18.8 (16.6–21.0) | 19.0 (16.9–21.0) |
| Affected your concentration, productivity or ability to stay awake | ||||
| A little or not at all | 46.4 (44.5–48.4) | 34.4 (32.5–36.2) | 41.3 (38.4–44.1) | 35.5 (33.0–38.1) |
| Somewhata | 24.2 (22.5–25.9) | 26.9 (25.2–28.6) | 24.7 (22.2–27.2) | 25.3 (23.0–27.6) |
| Much or very much | 29.4 (27.6–31.2) | 38.8 (36.9–40.7) | 34.0 (31.3–36.8) | 39.2 (36.6–41.8) |
| Troubled you in general | ||||
| A little or not at all | 63.8 (61.9–65.7) | 51.8 (49.8–53.7) | 56.8 (53.9–59.6) | 52.8 (50.2–55.5) |
| Somewhata | 20.0 (18.4–21.6) | 25.1 (23.4–26.8) | 23.3 (20.8–25.7) | 25.0 (22.7–27.3) |
| Much or very much | 16.2 (14.7–17.6) | 23.1 (21.5–24.8) | 20.0 (17.7–22.3) | 22.2 (20.0–24.3) |
| Symptoms of insomnia (SCI-8 ≤ 16) | 18.1 (16.6–19.6) | 24.1 (22.4–25.7) | 29.7 (27.1–32.4) | 27.9 (25.6–30.3) |
The multivariable regression results provide further evidence that sleep difficulties and rates of probable insomnia reported at school entry increased before the pandemic (i.e. were more common in autumn 2019 compared with autumn 2018), and then increased further during the peak of the pandemic (autumn 2020) (Supplementary Table 1). Sleep improved somewhat as the university shifted back to in-person teaching (autumn 2021). Students starting in the autumn of 2019 were 31% more likely to screen positive for probable insomnia than those starting the previous year. Similarly, students entering university in the autumn of 2020 (peak pandemic) were 23% more likely to report significant symptoms than those in the autumn 2019 (pre-pandemic) cohort. In contrast, students entering university in the autumn of 2021 were 7% less likely to report significant symptoms than the previous 2020 cohort.
Changes in sleep over the academic year and the pandemic
The average total SCI-8 score (range 0–32) is plotted in Fig. 2 and described by semester and academic year, adjusting for key covariates and cohort membership in Supplementary Table 2. Sleep was better on average at the start of the academic year (autumn) compared with the end of the academic year (spring) (total SCI-8 scores of 21.1 (95% CI 20.3–21.8) and 20.0 (95% CI 19.2–20.8), respectively). On average, sleep scores decreased, indicating worse sleep, from the 2018–2019 pre-pandemic year to the 2020–2021 peak pandemic year (from 21.5 (95% CI 20.8–22.3) to 20.2 (95% CI 19.4–21.0); average change −0.42 per year; P-trend < 0.001). Estimated differences in sleep scores across different semesters and academic years accounting for differences associated with other key covariates are presented in Table 3. The average decrease in sleep scores (indicating worsening sleep) from the autumn to spring semesters was −1.07 (95% CI −1.21 to −0.92). Although statistically significant, this effect is considered fairly small given the pooled s.d. of 8.39.21 The greatest decrease in sleep scores over the academic year was seen during the 2018–2019 pre-pandemic year (−2.06; 95% CI −2.41 to −1.71), followed by the 2020–2021 peak pandemic year (−1.39; 95% CI −1.65 to −1.12). Estimated differences in average sleep between academic years provided evidence of a successive worsening of sleep reaching the lowest point during the peak of the pandemic (2020–2021), with some improvement when classes resumed in person and restrictions eased (2021–2022).
| Time period contrast | Change in sleep (95% CI) | |
|---|---|---|
| Autumn to spring semester (average change) | −1.07 | (−1.21 to −0.92) |
| Autumn 2018 to spring 2019, pre-pandemic | −2.06 | (−2.41 to −1.71) |
| Autumn 2019 to spring 2020, transitional | −0.26 | (−0.53 to 0.02) |
| Autumn 2020 to spring 2021, peak pandemic | −1.39 | (−1.65 to −1.12) |
| Autumn 2021 to spring 2022, hybrid pandemic | −0.98 | (−1.22 to −0.74) |
| Academic year to academic year | ||
| 2018–2019 to 2019–2020, pre-pandemic to transitional | −0.61 | (−0.88 to −0.35) |
| 2019–2020 to 2020–2021, transitional to peak pandemic | −0.73 | (−0.96 to −0.50) |
| 2020–2021 to 2021–2022, peak pandemic to hybrid pandemic | 0.10 | (−0.11 to 0.30) |
Demographic and lifestyle factors associated with probable insomnia
Evidence from the most recent 2021–2022 academic year (across cohorts) supported that being younger and self-identifying as other than male were associated with screening positive for probable insomnia at the end of the year (spring semester) (Table 4). Further, more frequent cannabis use, spending more time engaged in passive screen time and spending less time on exercise or recreational activities at the start of the academic year were associated with screening positive for probable insomnia at the end of the year. Further, screening positive for insomnia at school entry was significantly associated with screening positive at the end of the academic year (relative risk 1.82; 95% CI 1.74–1.91), and once this was controlled for, effects with other lifestyle variables became more modest.
| Bivariate | Adjusted model 1a | Adjusted model 2b | ||||
|---|---|---|---|---|---|---|
| Relative risk | (95% CI) | Relative risk | (95% CI) | Relative risk | (95% CI) | |
| Age, years | ||||||
| Younger than 18 | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| 18 to 19 | 0.95 | (0.89–1.02) | 0.95 | (0.88–1.01) | 0.97 | (0.91–1.03) |
| 20 or older | 0.96 | (0.91–1.02) | 0.98 | (0.91–1.05) | 0.97 | (0.91–1.03) |
| Gender | ||||||
| Male | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| Female | 1.07 | (1.01–1.14) | 1.07 | (1.01–1.14) | 1.02 | (0.96–1.08) |
| Non-binary | 1.36 | (1.17–1.58) | 1.31 | (1.14–1.51) | 1.20 | (1.07–1.35) |
| Student status | ||||||
| Domestic | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| International | 0.98 | (0.88–1.08) | 0.99 | (0.89–1.10) | 1.03 | (0.93–1.14) |
| Cohort (year of study) | ||||||
| Cohort 1 (year 4) | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| Cohort 2 (year 3) | 1.08 | (0.98–1.18) | 1.08 | (0.98–1.18) | 1.04 | (0.96–1.12) |
| Cohort 3 (year 2) | 1.03 | (0.95–1.12) | 1.06 | (0.97–1.16) | 1.00 | (0.92–1.09) |
| Cohort 4 (year 1) | 1.05 | (0.98–1.12) | 1.06 | (0.97–1.17) | 1.01 | (0.94–1.09) |
| Binge drinking | ||||||
| Never | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| Monthly or less often | 1.01 | (0.95–1.07) | 1.00 | (0.94–1.07) | 1.00 | (0.94–1.07) |
| Weekly or more often | 1.02 | (0.94–1.10) | 0.99 | (0.91–1.08) | 0.98 | (0.91–1.06) |
| Cannabis use | ||||||
| Never | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| 3 days a month or less | 1.08 | (1.03–1.13) | 1.09 | (1.04–1.14) | 1.06 | (1.01–1.10) |
| 1–2 days a week or more | 1.14 | (1.06–1.22) | 1.15 | (1.07–1.24) | 1.07 | (1.00–1.15) |
| Exercise, past month | ||||||
| Never | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| Once a week or less | 0.98 | (0.94–1.04) | 1.00 | (0.95–1.05) | 1.02 | (0.98–1.06) |
| 2–3 times a week or more | 0.88 | (0.83–0.93) | 0.93 | (0.88–0.98) | 0.96 | (0.92–1.00) |
| Time for self-care and recreation | ||||||
| Less than weekly/never | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| Once a week | 0.86 | (0.80–0.93) | 0.87 | (0.81–0.94) | 0.93 | (0.86–1.01) |
| 2 or more times a week | 0.82 | (0.76–0.89) | 0.84 | (0.78–0.90) | 0.94 | (0.88–1.01) |
| Passive screen time, h/day | ||||||
| 3 h or less | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| 4–6 h/day | 1.04 | (0.98–1.10) | 1.04 | (0.98–1.11) | 1.01 | (0.95–1.07) |
| >6 h/day | 1.13 | (1.04–1.23) | 1.12 | (1.03–1.22) | 1.04 | (0.96–1.12) |
| Screen positive at baseline (SCI-8 ≤ 16) | ||||||
| No | 1.00 | Reference | / | / | 1.00 | Reference |
| Yes | 1.87 | (1.77–1.96) | / | / | 1.82 | (1.74–1.91) |
Discussion
In this large ongoing study of university students, sleep difficulties and clinically significant symptoms of insomnia were commonly reported by undergraduate students at entry to university. About a fifth of students reported a sleep problem lasting over a year, and screening positive for insomnia at the beginning of the academic year was strongly associated with screening positive at the end of the year, indicating the persistence of sleep difficulties. Sleep difficulties were reportedly associated with symptoms of cognitive impairment and negatively affected mood, energy and relationships. Probable insomnia was associated with several lifestyle factors, including increased cannabis use and passive screen time, and reduced time spent in exercise, recreation and self-care. With respect to the COVID-19 pandemic, sleep difficulties such as time to fall asleep, poor sleep quality and symptoms of probable insomnia showed a progressive increase from before the pandemic (autumn 2018) to the peak of the pandemic (autumn 2020), when learning was fully remote and pandemic-related restrictions most severe. Sleep improved somewhat when in-person classes resumed and restrictions were eased, but remained below pre-pandemic levels.
Study findings build on the evidence base establishing that sleep problems are common in university students, with an estimated prevalence of insomnia of 18.5% (95% CI 11.2–28.8%) before the pandemic.22 Our estimates fall within this range, and are therefore consistent with previous literature. Although sleep problems were most common at the peak of the pandemic, this was an emergent problem that existed before the pandemic, consistent with findings of others,23,24 with increases in technology use (smart phones and social media) implicated in the aetiology. Late-night technology use has been associated with night-time awakenings, difficulties getting back to sleep and insomnia.25
Based on the U-Flourish survey data, we have demonstrated that psychological risk factors at entry to university (i.e. low self-esteem, external locus of control and negative self-evaluation) are associated with insomnia and anxiety and depressive symptoms at the end of first year. Our observed pattern of an increased prevalence of probable insomnia and sleep difficulties over the pandemic aligns with a recent meta-analysis that reported a pooled prevalence of sleep problems in the general population of 32.3% (95% CI 25.3–40.2%) during the COVID-19 pandemic.26 In studies that used the Pittsburgh Sleep Quality Index, this prevalence was estimated at 37.9% (95% CI 25.2–52.4%), and younger age was associated with a higher prevalence of reported sleep problems.
In our analysis, there was a worsening of sleep difficulties over time until the peak of the pandemic. Such trends were associated with moving exclusively to an online learning format, along with other public health restrictions that curtailed living and work situations, social interactions, and sport and recreational pursuits. Pandemic conditions and associated limitations were associated with significant student concern around finances, studies, and the health and wellness of family members and friends.27 University students, especially women, reported sleeping more often as a means of coping with the stress of the COVID-19 pandemic.28 A majority of university students in the USA (85%) reported changing their sleep habits during the pandemic.29 In French university students, inactivity and idleness and academic worries during the pandemic were strongly associated with sleep problems.30 In university students in the USA, overall sleep quality did not change with the onset of the pandemic, but they reported greater sleep latency (time to fall asleep) and sleep medication use, and poorer sleep efficiency, associated with the COVID-19 pandemic.31 Bedtime and wake time were reported to be significantly later, and sleep duration was significantly longer.31 In addition to increased stress and worry, and greater inactivity, the lack of routine, use of asynchronous lectures and time changes for students studying from overseas or in other parts of the country may have contributed to the pandemic-associated rise in sleep problems.
We observed a pattern of worsening sleep from the beginning to end of each academic year across all cohorts, the actual magnitude of which varied from year to year and by phase of the pandemic. This finding aligns with a reported increase in screen-positive rates of anxiety and depressive symptoms over the academic year in undergraduates.4 A contributing factor underlying these trends may be increasing levels of stress as academic pressures and responsibilities mount, and final examinations draw closer.32 Examination-related anxiety can manifest in many ways, including disrupted sleep and increased fatigue.33 Seasonal effects may also play a role, affecting sleep duration, bedtime and wake time.34 Specifically, spring has been associated with negative effects on sleep duration, later bedtimes and earlier waking times, possibly related to day length and temperature.
In this study, cannabis use was modestly associated with a higher risk of probable insomnia. An extensive literature supports that substance misuse negatively affects the mental health of young people,35 and is associated with a number of negative academic, cognitive and personal outcomes.36 In particular, substance use may impede cognitive development, leading to lower academic achievement and performance. Over the course of the COVID-19 pandemic, our data show a decline of alcohol binging in university students, likely reflecting less opportunity for social gatherings. Yet, the pandemic seems to have had a relatively neutral or negative effect on student cannabis use.7 The pandemic and associated restrictions severely curtailed recreation and exercise opportunities, while promoting time spent alone and passively on screens and social media.37 Striking a healthy study–life balance through investing time in recreational activities and exercise, and reducing passive screen time, has proven benefits for student well-being, sense of connectedness and mental health.38,39
Strengths and limitations of this study warrant comment. This study addressed a health priority by using a robust cohort study design. The U-Flourish survey protocol relies on well-validated measures applied in four successively recruited samples of first-year students. Response rates were high compared with other health surveys among college and university students,40 especially before the pandemic. Statistical approaches were advanced and adjusted for potential confounding, differential response rates, missing data and attrition. Limitations included an over-representation of women, who report more sleep problems on average than men, leading to the potential for inflated prevalence estimates. Although steps were taken analytically to address potential bias, waning response rates over different stages of the pandemic and points of follow-up may have led to selection bias owing to attrition. Incomplete control for confounders may also be possible. Finally, experiences of Queen's University students may not be generalisable to other student populations, both within and outside of Canada.
Study findings are informative for prevention and the planning of clinical services. First, they highlight that sleep hygiene remains an important focus for health promotion efforts as students transition to university. Second, given that lifestyle affects sleep, student health promotion and mental health literacy efforts should address the negative effects of recreational drug use, passive screen time and reduced exercise on mental health, sleep and academic performance. Third, timely recognition of clinically significant and persistent sleep problems is an important consideration for early intervention efforts, and may unmask associated common mental health concerns such as anxiety and depression. Finally, there appears to be a persistent temporal pattern that predated the COVID-19 pandemic and was exacerbated by it, suggesting that sleep difficulties are becoming endemic in university student populations. The origin, consequences and potential treatment of sleep problems remain a clear priority.
Supporting information
Acknowledgements
We thank the students who participated in the U-Flourish research study and the student volunteers who helped with student engagement, contributing to the high participation rates.
Supplementary material
Supplementary material is available online at https://doi.org/10.1192/bjo.2023.597
Data availability
Access to the data and analytic code that support the findings of this study will be considered upon reasonable request to the principal investigator, A.D.
Funding
This work was supported by the Canadian Institutes of Health Research (grant numbers CIHR PJT 152976 and TG1 165597), the Rossy Family Foundation and the Mach Gaensslen Foundation.
Declaration of interest
C.B.M. is employed by Big Health Inc. (the developer of Sleepio), and is salaried by the company. Big Health Inc. holds the copyright for the use of the Sleep Condition Indicator. All other authors have no declarations of interest.
References
Untitled section
References
- 1.Taylor DJ, Bramoweth AD. Patterns and consequences of inadequate sleep in college students: substance use and motor vehicle accidents. J Adolesc Health 2011; 46(6): 610–2.
- 2.King N, Pickett W, McNevin SH, Bowie CR, Rivera D, Keown-Stoneman C, et al. Mental health need of students at entry to university: baseline findings from the U-flourish student well-being and academic success study. Early Interv Psychiatry 2021; 15(2): 286–95.
- 3.Duffy A, Keown-Stoneman C, Goodday S, Horrocks J, Lowe M, King N, et al. Predictors of mental health and academic outcomes in first-year university students: identifying prevention and early-intervention targets. BJPsych Open 2020; 6(3): e46.
- 4.Lindsay JAB, McGowan NM, King N, Rivera D, Li M, Byun J, et al. Psychological predictors of insomnia, anxiety and depression in university students: potential prevention targets. BJPsych Open 2022; 8(3): e86.
- 5.Johansson M, Jansson-Fröjmark M, Norell-Clarke A, Linton SJ. Changes in insomnia as a risk factor for the incidence and persistence of anxiety and depression: a longitudinal community study. Sleep Sci Prac 2021; 5: 5.
- 6.Baglioni C, Battagliese G, Feige B, Spiegelhalder K, Nissen C, Voderholzer U, et al. Insomnia as a predictor of depression: a meta-analytic evaluation of longitudinal epidemiological studies. J Affec Disord 2011; 135(1–3): 10–9.
- 7.King N, Pickett W, Rivera D, Byun J, Li M, Cunningham S, et al. The impact of the COVID-19 pandemic on the mental health of first-year undergraduate students studying at a major Canadian university: a successive cohort study. Can J Psychiatry 2023; 68(7): 499–509.
- 8.Elharake JA, Akbar F, Malik AA, Gilliam W, Omer SB. Mental health impact of COVID-19 among children and college students: a systematic review. Child Psychiatry Hum Dev 2023; 54(3): 913–25.
- 9.Partinen M, Holzinger B, Morin CM, Espie C, Chung F, Penzel T, et al. Sleep and daytime problems during the COVID-19 pandemic and effects of coronavirus infection, confinement and financial suffering: a multinational survey using a harmonised questionnaire. BMJ Open 2021; 11(12): e050672.
- 10.Goodday SM, Rivera D, Foran H, King N, Milanovic M, Keown-Stoneman CD, et al. U-Flourish university students well-being and academic success longitudinal study: a study protocol. BMJ Open 2019; 9(8): e029854.
- 11.Bifulco A, Brown GW, Harris TO. Childhood experience of care and abuse (CECA): a retrospective interview measure. J Child Psychol Psychiatry 1994; 35(8): 1419–35.
- 12.Spitzer RL, Kroenke K, Williams JB, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006; 166(10): 1092–7.
- 13.Kroenke K, Spitzer R, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001; 16(9): 606–13.
- 14.Espie CA, Kyle SD, Hames P, Gardani M, Fleming L, Cape J. The sleep condition indicator: a clinical screening tool to evaluate insomnia disorder. BMJ Open 2014; 4(3): e004183.
- 15.Bayard S, Lebrun C, Maudarbocus KH, Schellaert V, Joffre A, Ferrante E, et al. Validation of a French version of the sleep condition indicator: a clinical screening tool for insomnia disorder according to DSM-5 criteria. J Sleep Res 2017; 26(6): 702–8.
- 16.Karabon P. Applying Propensity Score Methods to Complex Survey Data using PROC PSMATCH. Oakland University William Beaumont School of Medicine, 2019. (https://www.sas.com/content/dam/SAS/support/en/sas-global-forum-proceedings/2019/3634-2019.pdf).
- 17.Seaman SR, White IR. Review of inverse probability weighting for dealing with missing data. Stat Methods Med Res 2013; 22(3): 278–95.
- 18.Lenth RV. emmeans: Estimated Marginal Means, aka Least-Squares Means. R Package Version 1.7.2. Comprehensive R Archive Network, 2022. (https://CRAN.R-project.org/package=emmeans).
- 19.van Buuren S. Flexible Imputation of Missing Data. 2nd Ed. Chapman and Hall/CRC, 2018.
- 20.Yuan Y. Multiple Imputation for Missing Data: Concepts and new Development. SAS Institute, 2010. (http://www.math.montana.edu/~jimrc/classes/stat506/notes/multipleimputation-SAS.pdf).
- 21.Lakens D. Calculating and reporting effect sizes to facilitate cumulative science: a practical primer for t-tests and ANOVAs. Front Psychol 2013; 4: 863.
- 22.Jiang XL, Zheng XY, Yang J, Ye CP, Chen YY, Zhang ZG, et al. A systematic review of studies on the prevalence of insomnia in university students. Public Health 2015; 129(12): 1579–84.
- 23.Patte KA, Qian W, Leatherdale ST. Sleep duration trends and trajectories among youth in the COMPASS study. Sleep Health 2017; 3(5): 309–16.
- 24.Chattu VK, Manzar MD, Kumary S, Burman D, Spence DW, Pandi-Perumal SR. The global problem of insufficient sleep and its serious public health implications. Healthcare (Basel) 2018; 7(1): 1.
- 25.Rosen L, Carrier LM, Miller A, Rokkum J, Ruiz A. Sleeping with technology: cognitive, affective, and technology usage predictors of sleep problems among college students. Sleep Health 2016; 2(1): 49–56.
- 26.Jahrami H, BaHammam AS, Bragazzi NL, Saif Z, Faris M, Vitiello MV. Sleep problems during the COVID-19 pandemic by population: a systematic review and meta-analysis. J Clin Sleep Med 2021; 17(2): 299–313.
- 27.Appleby JA, King N, Saunders KE, Bast A, Rivera D, Byun J, et al. Impact of the COVID-19 pandemic on the experience and mental health of university students studying in Canada and the UK: a cross-sectional study. BMJ Open 2022; 12(1): e050187.
- 28.Prowse R, Sherratt F, Abizaid A, Gabrys RL, Hellemans K, Patterson ZR, et al. Coping with the COVID-19 pandemic: examining gender differences in stress and mental health among university students. Front Psychiatry 2021; 12: 650759.
- 29.Wang X, Hegde S, Son C, Keller B, Smith A, Sasangohar F. Investigating mental health of US college students during the COVID-19 pandemic: cross-sectional survey study. J Med Internet Res 2020; 22(9): e22817.
- 30.Truchot D, Andela M, Takhiart H. Stressors met by quarantined French students during the COVID-19 pandemic. Their links with depression and sleep disorders. J Affect Disord 2021; 294: 54–9.
- 31.Benham G. Stress and sleep in college students prior to and during the COVID-19 pandemic. Stress Health 2021; 37(3): 504–15.
- 32.Sreeramareddy CT, Shankar PR, Binu VS, Mukhopadhyay C, Ray B, Menezes RG. Psychological morbidity, sources of stress and coping strategies among undergraduate medical students of Nepal. BMC Med Educ 2007; 7: 26.
- 33.Rizvi AH, Awaiz M, Ghanghro Z, Jafferi MA, Aziz S. Pre-examination stress in second year medical students in a government college. J Ayub Med Coll Abbottabad 2010; 22(2): 152–5.
- 34.Mattingly SM, Grover T, Martinez GJ, Aledavood T, Robles-Granda P, Nies K, et al. The effects of seasons and weather on sleep patterns measured through longitudinal multimodal sensing. NPJ Digit Med 2021; 4(1): 76.
- 35.Esmaeelzadeh S, Moraros J, Thorpe L, Bird Y. Examining the association and directionality between mental health disorders and substance Use among adolescents and young adults in the U.S. and Canada-a systematic review and meta-analysis. J Clin Med 2018; 7(12): 543.
- 36.King KM, Meehan BT, Trim RS, Chassin L. Substance use and academic outcomes: synthesizing findings and future directions. Addiction 2006; 101(12): 1688–9.
- 37.Lesser IA, Nienhuis CP. The impact of COVID-19 on physical activity behavior and well-being of Canadians. Int J Environ Res Public Health 2020; 17(11): 3899.
- 38.Herbert C, Meixner F, Wiebking C, Gilg V. Regular physical activity, short-term exercise, mental health, and well-being among university students: the results of an online and a laboratory study. Front Psychol 2020; 11: 509.
- 39.Tang S, Werner-Seidler A, Torok M, Mackinnon AJ, Christensen H. The relationship between screen time and mental health in young people: a systematic review of longitudinal studies. Clin Psychol Rev 2021; 86: 102021.
- 40.American College Health Association. American College Health Association-National College Health Assessment II: Canadian Reference Group Data Report Spring 2016. American College Health Association, 2016. (https://www.acha.org/documents/ncha/NCHA-II%20SPRING%202016%20CANADIAN%20REFERENCE%20GROUP%20EXECUTIVE%20SUMMARY.pdf).
Associated Data
Supplementary Materials
Data Availability Statement
Access to the data and analytic code that support the findings of this study will be considered upon reasonable request to the principal investigator, A.D.