Abstract
Sleep is fundamental to health. The aim of this study was to analyse and determine factors predicting sleep quality during and after national lockdowns due to severe acute respiratory syndrome coronavirus 2 (COVID-19) in the UK. A longitudinal online survey-based study (SleepQuest) involving UK adults was administered in Spring 2020, Winter 2020, and Winter 2022 including questionnaires probing sleep quality, depression, anxiety, beliefs about sleep, demographics, COVID-19 status, and exercise. The primary outcome was sleep quality (Pittsburgh Sleep Quality Index). A linear mixed-effects model evaluated factors associated with baseline and longitudinal sleep quality. Complete data were provided by 3306 participants in Spring 2020, 2196 participants in Winter 2020, and 1193 in Winter 2022. Participants were mostly female (73.8%), white (97.4%), and aged over 50 years (81.0%). On average, participants reported poor sleep quality in Spring 2020 (mean [SD] Pittsburgh Sleep Quality Index score = 6.59 [3.6]) and Winter 2020 (mean [SD] Pittsburgh Sleep Quality Index score = 6.44 [3.6]), with improved but still poor sleep quality in Winter 2022 (mean [SD] Pittsburgh Sleep Quality Index score = 6.17 [3.5]). Improved sleep quality was driven by better subjective sleep and reduced daytime dysfunction and sleep latency. Being female, older, having caring responsibilities, working nightshifts, and reporting higher levels of depression, anxiety, and unhelpful beliefs about sleep were associated with worse baseline PSQI scores. Better sleep quality was associated with more days exercising per week at baseline. Interventions focusing on improving mental health, exercise, and attitudes towards sleep, particularly in at-risk groups, may improve sleep-related outcomes in future pandemics.
Original language | English |
---|---|
Article number | e14205 |
Journal | Journal of Sleep Research |
Volume | 33 |
Issue number | 6 |
Early online date | 23 Apr 2024 |
DOIs | |
Publication status | Published - 31 Dec 2024 |
Data Availability Statement
The data that support the findings of this study are openly available in Dementias Platform UK Data Portal at https://doi.org/10.48532/043000.Funding
This work was funded by a Above & Beyond grant (ABL-2019-20-01) and by Alzheimer's Research UK (ARUK-NC2019-BB). AW was funded by BRACE. JB receives funding from Alzheimer's Research UK (supported by the Margaret Jost Fellowship and the Don Thoburn Memorial Scholarship) and the David Telling Charitable Trust. EC has received funding from BRACE and ARUK (Bristol & Bath Network). VG has received funding from Above & Beyond and NIHR Bristol Biomedical Research Centre. SleepQuest was a longitudinal web-based survey hosted on the REDCap system at the University of Bristol, UK, open to all UK adults over the age of 18 years. Data were collected at two timepoints during periods of UK national lockdown (T1: 29/04/2020\u201313/05/2020; T2: 05/11/2020\u201302/12/2020), and a third timepoint that followed relaxation of all lockdown rules (T3: 05/12/2022\u201319/12/2022). Participants were provided with information online before consenting to take part. The questionnaire took approximately 20 min to complete at each timepoint. To prioritise recruitment of older adults, participants were primarily recruited via the Join Dementia Research (JDR) database managed by the National Institute for Health Research in collaboration with the charitable organisations Alzheimer's Society, Alzheimer's Research UK, and Alzheimer's Scotland. The register comprises over 45,000 people who have agreed to be contacted about research opportunities. While JDR targets recruitment of people with dementia and caregivers, this is not mandatory, and many volunteers are healthy older adults without diagnoses or caring responsibilities. Email invitations were sent to JDR participants summarising the SleepQuest study with an opportunity to enrol via a website link. Participants were invited for follow-up via direct email. To encourage recruitment of younger and middle-aged adults to explore the association between age and sleep quality, the study was also advertised via Twitter and Facebook, through press releases by the University of Bristol public relations office, and through personal social media accounts. For full details of participant selection, see Figure\u00A01. Ethical approval for the study was through the Faculty of Health Sciences Research Ethics Committee, University of Bristol (FREC ref: 103244). SleepQuest was a longitudinal web-based survey hosted on the REDCap system at the University of Bristol, UK, open to all UK adults over the age of 18 years. Data were collected at two timepoints during periods of UK national lockdown (T1: 29/04/2020\u201313/05/2020; T2: 05/11/2020\u201302/12/2020), and a third timepoint that followed relaxation of all lockdown rules (T3: 05/12/2022\u201319/12/2022). Participants were provided with information online before consenting to take part. The questionnaire took approximately 20 min to complete at each timepoint. To prioritise recruitment of older adults, participants were primarily recruited via the Join Dementia Research (JDR) database managed by the National Institute for Health Research in collaboration with the charitable organisations Alzheimer's Society, Alzheimer's Research UK, and Alzheimer's Scotland. The register comprises over 45,000 people who have agreed to be contacted about research opportunities. While JDR targets recruitment of people with dementia and caregivers, this is not mandatory, and many volunteers are healthy older adults without diagnoses or caring responsibilities. Email invitations were sent to JDR participants summarising the SleepQuest study with an opportunity to enrol via a website link. Participants were invited for follow-up via direct email. To encourage recruitment of younger and middle-aged adults to explore the association between age and sleep quality, the study was also advertised via Twitter and Facebook, through press releases by the University of Bristol public relations office, and through personal social media accounts. For full details of participant selection, see Figure\u00A01. Ethical approval for the study was through the Faculty of Health Sciences Research Ethics Committee, University of Bristol (FREC ref: 103244). General sociodemographic information was collected including age, gender, ethnicity, education, employment (including night-shift work), changes to employment due to COVID-19, number of dependents, caring responsibilities, smoking and alcohol consumption, medical comorbidities (including presence of sleep disorders or dementia), and presence of confirmed or suspected COVID-19 infection. The Pittsburgh Sleep Quality Index (PSQI) is a self-report, validated questionnaire designed to assess multiple aspects of sleep over the previous month (Buysse et al.,\u00A01989). A total of 19 individual items generate seven component scores whose summation provides a global sleep quality score. Component scores cover subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction. Scoring is based on a four-point scale ranging from 0 to 3, with increased values representing more compromised sleep. A global score of 5 or greater indicates a \u201Cpoor\u201D sleeper (Buysse et al.,\u00A01989). The PSQI has good internal consistency and reliability (Cronbach's \u03B1 = 0.83) for its seven components (Buysse et al.,\u00A01989). The PHQ-8 (Kroenke et al.,\u00A02009) is an eight-item self-report measure assessing depressive symptoms. It removes the ninth item from the more widely used PHQ-9\u00A0(\u201Cthoughts that you would be better off dead or of hurting yourself in some way\u201D), as recommended in self-administered internet surveys where further probing about positive responses to item 9 are not possible (Kroenke & Spitzer,\u00A02002). Scores range from 0 to 24, with scores equal or greater than 10 indicating depression may be present. It has been shown to shown to be a valid diagnostic and severity measure in large clinical studies (Kroenke & Spitzer,\u00A02002). The GAD-7 is a validated self-report questionnaire that assesses generalised anxiety using seven items with four-point Likert scales. Total scores range from 0 to 21, with clinical cut-offs of 5, 10 and 15 for mild, moderate and severe anxiety, respectively (Spitzer et al.,\u00A02006). The DBAS-16 is a 16-item self-report measure assessing beliefs and attitudes towards sleep using a Likert-type scale ranging from 0 to 10 for items such as \u201CI am worried that I may lose control over my abilities to sleep\u201D. Higher scores represent potentially unhelpful attitudes around sleep. It has been shown to be a reliable and valid measure of sleep-related cognition in clinical samples (Morin et al.,\u00A02007). We captured the number of days in the previous week that any time was spent outside and the mean duration of these occasions. The number of days in the preceding week in which the participant had engaged in moderate to vigorous exercise (defined as activity that raises heart rate, increases breathing rate and makes them feel warmer) for \u2265 15 min was also captured. R version 4.2.2 (2022-10-31) \u2013 \u201CInnocent and Trusting\u201D and RStudio (v2023.03.1 + 446) statistical software was used for data cleaning and analysis. Analysis of each timepoint included only participants supplying a full dataset for that timepoint. Post-hoc sensitivity analyses were conducted to check whether findings were replicable when only including participants who completed all three timepoints. Descriptive statistics were used to characterise participants' demographics, health status, depression and anxiety, attitudes towards sleep, and time spent outside and exercising. Unpaired t-tests assessed population-level self-reported change in sleep quality, depression, anxiety and dysfunctional sleep beliefs as reflected by total PSQI, PHQ-8, GAD-7 and DBAS-16 scores, respectively, at timepoints T1 versus T2, T2 versus T3, and T1 versus T3. Population-level change in each of the seven PSQI sub-components was also compared using the Wilcoxon Rank Sum Test. A significance level of 0.05 was utilised. Factors predicting change in sleep were assessed using linear mixed-model analysis (LMM) using the lme function in R. PSQI total scores at timepoints T1, T2 and T3 were dependent variables for analysis. An intraclass correlation coefficient of 0.762 calculated from a grand mean (null) model suggested significant intra-individual clustering, thus validating an LMM approach. Timepoint was assigned as 0, 1 and 5, respectively, for timepoints T1, T2 and T3, reflecting the proportional interval between data collection periods. Log-likelihood values derived from unconditional models with time as a fixed effect (random intercept) and time as a random effect (random slope and intercept models) were compared by ANOVA with the latter providing improved model fit (p < 0.001) and therefore utilised for analysis. The final conditional model incorporated continuous covariates including age at baseline, PHQ-8, GAD-7 and DBAS-16 total scores, days of exercise and daylight exposure per week, and dummy variables to incorporate gender, COVID-19 infection within 6 months, nightshift work, presence of cognitive impairment/dementia, presence of children in the household, outdoor employment, and carer status. Covariates were selected by systematic examination of each collected data variable for its potential (hypothetically or evidentially) to influence the dependent variable (total PSQI). The final model specification therefore utilised participant identity as a random effect with interactions between covariates and timepoint included as a fixed effect of the form: Total PSQI Score~Timepoint+Covariates+Timepoint\u00D7Covariates+TimepointParticipant. The final adjusted regression model was checked for multicollinearity (Variance Inflation Factors < 10), normality of errors, homoskedasticity and linearity.
Funders | Funder number |
---|---|
Alzheimer's Society | |
Margaret Jost Fellowship | |
Alzheimer's Scotland | |
NIHR Bristol Biomedical Research Centre | |
National Institute for Health and Care Research | |
David Telling Charitable Trust | |
BRACE | |
Anova Health Institute | |
Bristol & Bath Network | |
RStudio | GAD-7, DBAS-16, v2023.03.1 + 446, PHQ-8 |
University of Bristol | 29/04/2020–13/05/2020, 05/11/2020–02/12/2020, 05/12/2022–19/12/2022 |
University of Bristol | |
Flavor Research and Education Center, Ohio State University | 103244, 2022-10-31 |
Flavor Research and Education Center, Ohio State University | |
Alzheimer's Research UK | ARUK-NC2019-BB |
Alzheimer's Research UK |
Keywords
- coronavirus
- COVID-19
- mental health
- older adults
- pandemic
- sleep
ASJC Scopus subject areas
- Cognitive Neuroscience
- Behavioral Neuroscience