key: cord-0718296-brmjviqb
authors: Korman, Maria; Tkachev, Vadim; Reis, Cátia; Komada, Yoko; Kitamura, Shingo; Gubin, Denis; Kumar, Vinod; Roenneberg, Till
title: Outdoor daylight exposure and longer sleep promote wellbeing under COVID‐19 mandated restrictions
date: 2021-09-21
journal: J Sleep Res
DOI: 10.1111/jsr.13471
sha: e9611dae3a4fee4df0c61b8df003357fcb644e22
doc_id: 718296
cord_uid: brmjviqb

Light is an important regulator of daily human physiology in providing time‐of‐day information for the circadian clock to stay synchronised with the 24‐hr day. The coronavirus disease 2019 (COVID‐19) pandemic led to social restrictions in many countries to prevent virus spreading, restrictions that dramatically altered daily routines and limited outdoor daylight exposure. We previously reported that sleep duration increased, social jetlag decreased, and mid‐sleep times delayed during social restrictions (Global Chrono Corona Survey, N = 7,517). In the present study, we investigated in the same dataset changes in wellbeing and their link to outdoor daylight exposure, and sleep–wake behaviour. In social restrictions, median values of sleep quality, quality of life, physical activity and productivity deteriorated, while screen time increased, and outdoor daylight exposure was reduced by ~58%. Yet, many survey participants also reported no changes or even improvements. Larger reductions in outdoor daylight exposure were linked to deteriorations in wellbeing and delayed mid‐sleep times. Notably, sleep duration was not associated with outdoor daylight exposure loss. Longer sleep and decreased alarm‐clock use dose‐dependently correlated with changes in sleep quality and quality of life. Regression analysis for each wellbeing aspect showed that a model with six predictors including both levels and their deltas of outdoor daylight exposure, sleep duration and mid‐sleep timing explained 5%–10% of the variance in changes of wellbeing scores (except for productivity). As exposure to daylight may extenuate the negative effects of social restriction and prevent sleep disruption, public strategies during pandemics should actively foster spending more daytime outdoors.

An adaptation to Earth's rotation is essential for survival, and has fostered the evolution of endogenous, circadian clocks, that synchronize (entrain) to cyclic environmental cues (zeitgebers) (Aschoff & Pohl, 1978) . Light-dark cycles (LD) are the dominant zeitgeber for the human clock as shown in laboratory and real-life studies (Pilz et al., 2018; Stothard et al., 2017; Wright et al., 2013) . Individuals entrain differently to LD cycles, earlier or later, depending on the clock's characteristics and zeitgeber strength. Circadian clocks adapt a stable relationship to the zeitgeber ("phase of entrainment" or chronotype), which range from extremely early ("larks") to extremely late ("owls"), with the majority of individuals ("doves") falling in between . Chronotype depends on genes, age and sex (Roenneberg et al., 2004; 2019), on geographic location (Leocadio-Miguel et al., 2017) and

zeitgeber strength (i.e. the maximum and minimum of the LD cycle) (Pilz et al., 2018) .

In industrialised societies, people predominantly live inside and artificially illuminate the night, which weakens zeitgeber strength, thereby delaying chronotype in most people. Early schedules expose especially late chronotypes to a mismatch between circadian and social time (Wittmann et al., 2006) , which is quantified as difference between mid-sleep times on work and work-free days. This social jetlag (SJL) has been linked to health-risk behaviours and diseases (Mota et al., 2019; Wittmann et al., 2010) .

Besides acting as a zeitgeber for the circadian clock, light promotes alertness, mood, vitality, cognitive function, and social interactions (Gaggioni et al., 2014; Partonen & Lonnqvist, 2000) .

Morning light can compensate for cognitive deficits, e.g. attentiondeficit hyperactivity disorder in adults (Korman, Palm, et al., 2020; Rybak et al., 2006) . Light therapy is widely used to treat depression and mood disorders, e.g. seasonal affective disorder (SAD) (Sit et al., 2018; Wirz-Justice et al., 2005) . Light therapy is thought to activate dopaminergic (Kim et al., 2017) , adrenergic (Bowrey et al., 2017) , and serotonergic (Li, 2018) pathways that are directly linked to affect, emotion, mood, and melatonin production. The effects of light depend on time-of-day, duration, intensity, and its wavelength (Marshall, 2016; Wirz-Justice et al., 2005) . Integration of light over the day is also important (Leocadio-Miguel et al., 2017) . Notably, artificial light is mostly orders of magnitude lower compared to outdoor daylight. Light-at-night, especially blue light, alerts, suppresses melatonin levels (usually rising after dusk), and delays the circadian clock (Duffy & Czeisler, 2009) . Light-at-night gains relevance with increased blue-light bulbs and screens and is considered harmful to health (Marshall, 2016) .

(COVID-19) pandemic were often associated with robust changes in outdoor daylight exposure (OLE), daily behaviour, and sleep.

Although sleep worsened in many individuals, changes were positive at the population level: sleep duration increased and SJL decreased significantly (Gao & Scullin, 2020; Korman, Tkachev, et al., 2020; Leone et al., 2020; Wright et al., 2020) . The more under-slept and misaligned individuals were before social restrictions, the more they increased sleep duration or decreased SJL during social restrictions (Korman, Tkachev, et al., 2020) . Sleep quality was not uniformly affected (Gao & Scullin, 2020; Kocevska et al., 2020; Leone et al., 2020) . Notably, sleep quality improved in individuals who suffered from clinical insomnia prior to the pandemic (Kocevska et al., 2020) , while it did not change in a USA sample (Gao & Scullin, 2020) . These positive changes may reflect decreased social time pressure due to home offices, elimination of commutes or more relaxed work schedules (Korman, Tkachev, et al., 2020) and contrast reports of the pandemic's negative impact on psychological and psychiatric wellbeing (Ozamiz-Etxebarria et al., 2020) , challenging mental health services worldwide .

It was proposed that managing sleep, along with stress, anxiety, and symptoms of depression is important during social restrictions (Altena et al., 2020) . In the present study, we examined associations between changes in OLE, sleep-wake behaviour, and various aspects of wellbeing (sleep quality, quality of life, physical activity, screen time and productivity) in the same participants as in (Korman, Tkachev, et al., 2020) . We expected the changes in wellbeing parameters to be cross-correlated. We also hypothesised that the magnitude of the decrease in OLE during social restrictions would be associated with negative changes in wellbeing, particularly, in sleep quality, physical activity, and quality-of-life aspects, known to be changes in sleep quality and quality of life. Regression analysis for each wellbeing aspect showed that a model with six predictors including both levels and their deltas of outdoor daylight exposure, sleep duration and mid-sleep timing explained 5%-10% of the variance in changes of wellbeing scores (except for productivity). As exposure to daylight may extenuate the negative effects of social restriction and prevent sleep disruption, public strategies during pandemics should actively foster spending more daytime outdoors.

circadian rhythms, light-dark cycle, resilience, screen time, sleep-wake behaviour related to OLE. In addition, we expected that changes in sleep duration and timing would be associated with those in wellbeing and OLE during social restrictions.

The internet-based Global Chrono Corona Survey (GCCS) was approved by the Ariel University Human Research Ethics Committee of the Faculty of Health Sciences (AU-HEA-MK-20200629). Survey participants provided electronic consent to participate in the study.

We collected data using the SoGoSurvey platform (Herndon, VA, USA), a cloud-based platform for creation and distribution of multilingual surveys. The GCCS was translated into 10 languages (English, German, Hebrew, Arabic, Hindi, Japanese, Italian, Portuguese, Russian, and Spanish) by an international network of colleagues (see Acknowledgements). Recruitment methods included digital advertisements at universities, academic and non-academic social networks, and email-based approaches. Participation in the survey was anonymous.

The GCCS contained 40-54 items (for specifics of their implementation, see Korman, Tkachev, et al., 2020) For individual deltas (ΔSD, ΔMST, ΔSJL), we subtracted the respective preSR values from those inSR.

Additionally, participants indicated their average OLE separately for workdays and work-free days as falling into one of the following categories: <30 min, 30-60 min, 1-2 hr, 2-3 hr, 3-4 hr, 4-5 hr, 5-6 hr, 6-7 hr, >7 hr. Categorical answers were transformed to numerical values using the mid-point of the category interval: 15 min, 45 min, …, 390 min, and 450 min (for >7 hr), respectively. Individual mean OLE over the week was calculated as a weighted average of OLE on workdays (5 days assumed) and work-free days.

The GCCS also queried subjective wellbeing parameters using 5item Likert scales: Sleep-Quality (got worse-got better), Quality-of-Life (got worse-got better), Physical-Activity (increased-decreased), Screen-Time (increased-decreased) and Productivity (got worsegot better), inSR compared to preSR. Participants' responses were coded as: very negative (−2), negative (−1), no change (0) 

Demographic data were published previously (Korman, Tkachev, et al., 2020) . In total, 11,431 respondents (aged ≥18 years) from 40 countries completed the GCCS during the first wave of SR (April 4 to May 16, 2020). The highest response rates (>200 respondents/ country) were from Portugal, Italy, USA, UK, Germany, Israel, India, Russia, Japan, and Brazil. Exclusion criteria were a COVID-19 diagnosis (1.1%), shift/night workers (16.3%), extreme sleep durations (<3 hr and >14 hr; 3.1%) and missing/invalid data (8.3%). To correct for the over representation of young (aged 18-22 years) participants from Russia compared with the other two leading countries (Japan and India), we randomly excluded 656 participants from Russia (5.7%). To this end, a uniform random selection procedure (rand function in Excel) was applied to a subgroup of Russian participants in the age group 18-22 years (N original = 1,006, N final = 350). The final sample included 7,517 participants (68.2% female), all under SR on the day of response. On average, participants had been under SR for 32.7 ± 9.1 days (range 10-59 days), presumably allowing full adjustment to new schedules. In all, 80% of respondents worked or studied both preSR and inSR. InSR, 66% worked from home (preSR, 11%).

Material (Table S1 ).

Data were pre-processed as published (Korman, Tkachev, et al., 2020) . We used non-parametric data analyses due to the non-normal distribution or homoscedastic nature of the behavioural data. Onesample Wilcoxon tests (separate for each wellbeing parameter) were used to assess significant in Δscores and the Kruskal-Wallis H test 

Reported impairments outnumbered reported improvements inSR (Figure 1a-e); medians differed significantly from "no change" for all queried aspects of wellbeing (separate one-sample Wilcoxon tests, see Table S2 ). Changes in most wellbeing aspects (Δscores) were mostly negative: Quality-of-Life (49.6%), Physical-Activity (51.0%), Productivity (66.8%) and Screen-Free-Time (74.3%). Notably, more participants reported no change in Sleep-Quality (42.8%) than worsening (34.2%) or improving (23.0%).

Women had more negative ΔSleep-Quality, ΔQuality-of-Life, and

ΔScreen-Time scores than men (Mann-Whitney U tests between sex groups: p < 0.001, r g = 0.06; p = 0.033, r g = 0.03; p < 0.001, r g = 0.06, respectively), but the effect sizes of the differences between sexes were negligible (r g < 0.1). No sex differences were found for ΔPhysical-Activity and ΔProductivity. The Δscores were age independent, with the only exclusion for Screen-Time: older participants reported smaller increase in inSR (Spearman's ρ = 0.222,

All Δscores cross-correlated ( Figure 1f ); the strongest links (ρ > 0.3) were found between the ΔSleep-Quality and ΔQualityof-Life, and between ΔPhysical-Activity and ΔQuality-of-Life ( Figure 1f ).

The median weekly OLE was reduced from 1 hr 47 min (interquartile Table S3 ). The ΔOLE was −72 ± 112 min.

The ΔOLE correlated with age, with larger losses in young people (Spearman's ρ = 0.15, p < 0.001), but there were no differences between the sexes.

As previously reported (Korman, Tkachev, et al., 2020) Kruskal-Wallis tests were followed by pairwise Dunn test comparisons between subgroups (results were Bonferroni corrected). The same four aspects also showed significant differences between the negative-positive and negative-no change subgroups. Only for ΔQuality-of-Life was the difference between the no change-positive pair significant.

Negative ΔOLE was also associated with both later MST inSR (ρ = 0.23) and larger ΔMST (ρ = 0.16). The ΔOLE was independent of SD inSR, ΔSD, SJL inSR, and ΔSJL. None of the predictors accounted for >1% change in the variance in ΔProductivity.

To assess the impact of alarm clock use inSR on wellbeing Δscores, we selected a group of participants who worked/studied both preSR and inSR, used an alarm clock on workdays preSR, and worked/studied from home inSR. This group (N = 4,135) was then subdivided into those who stopped using an alarm clock inSR (Alarm/NoAlarm; N = 1,539

[37%]) and those who continued to use alarm clock inSR (Alarm/ Alarm; N = 2,596 [63%]). On average, the Alarm/NoAlarm group had higher ΔSleep-Quality and ΔQuality-of-Life scores (Mann-Whitney tests; Z = 3.53, p < 0.001, and Z = 3.04, p < 0.001, respectively) but lower ΔProductivity scores (Z = −5.06, p < 0.001) compared to the Alarm/Alarm group. There were no significant differences between the groups in ΔPhysical-Activity and ΔScreen-Time scores. The two groups were similar in age and sex composition (Table S5 ).

As part of preventing infections with COVID-19, governments around the world imposed drastic restrictions on their citizens' freedom to move. These social restrictions represented a global experiment that changed OLE, social time pressures, and many aspects of daily routines. Our previously published findings of the GCCS study (Korman, Tkachev, et al., 2020) showed that participants slept longer and later inSR with a concomitant decrease in SJL. In the present study, we show the importance of changes in OLE and sleep-wake behaviour linked to changes in wellbeing during the period of social restrictions. Our most important findings are summarised in Social restrictions impaired all aspects of wellbeing, with sleep quality, quality of life, physical activity, and productivity deteriorating and screen time increasing in their medians. Yet, many GCCS participants also reported no changes or even improvements.

Notably, more participants reported no changes in Sleep-Quality (43%) than deteriorations or improvements (34% and 23%, respectively). This is consistent with previous reports of large scale (Florea et al., 2021; Gao & Scullin, 2020; Kocevska et al., 2020; Leone et al., 2020) and a recent meta-analysis that found that sleep problems of people from the general population during the COVID-19 pandemic affected ~32% (Jahrami et al., 2021) .

Thus, longer sleep and less SJL (as reported for the same sample by Korman, Tkachev, et al., 2020) seem not directly linked to sleep quality. However, analyses of individuals show that those reporting deteriorations in Sleep-Quality and Quality-of-Life also reported smaller gains in sleep duration and used alarm clocks more often. Although causalities in this association remain untested, it is plausible that social restrictions affect quality of life through stress mechanisms (Gao & Scullin, 2020; Ozamiz-Etxebarria et al., 2020) , thereby preventing longer sleep despite relaxed social time pressure. Altogether, relief from social time pressure during social restrictions allows both longer sleep (Korman, Tkachev, et al., 2020) and waking without an alarm clock, thereby improving sleep quality and quality of life (both scores were higher in the subgroup that stopped using an alarm clock inSR).

Notably, deteriorations in Sleep-Quality, Quality-of-Life, Physical-Activity and Screen-Time during the pandemic were associated with higher losses in weekly OLE. As decreased weekly OLE is predominantly caused by the social restrictions rather than merely associated with them, it is fair to presume a causal positive influence of OLE on many aspects of wellbeing. A combination of decreased OLE and increased Screen-Time has predictably powerful effects on circadian timing. They combine decrease in zeitgeber strength and more light after sunset, and both these effects individually delay the circadian phase (MST inSR) in most individuals (Moderie et al., 2017 ).

An important methodological limitation of the present study in this respect is that it is unknown during what time of day the changes in screen time took place. An increase in the Screen-Time/OLE ratio has been suggested to exacerbate myopia during the recent pandemic (Wong et al., 2021) .

The division of participants into subgroups reflecting their wellbeing changes (negative, no change, positive) strongly indicates that the loss of OLE during the pandemic actually mediates TA B L E 1 Multiple linear regressions for six predictors of changes in wellbeing categories. (Green background, predictors that were responsible for >1% change in the variance in the wellbeing change scores) , 1996) .

It is important that the present data were collected during the first wave of COVID-19 in all participating countries and a small number of participants who had COVID-19 during the data collection period were excluded from the analysis (see Methods). Therefore, the present study reflects the impact of social restrictions on wellbeing and daily behaviour rather than the consequences of viral infections.

Since then, millions have contracted COVID-19 and many continue to suffer from its long-term effects that frequently include sleep problems (Jahrami et al., 2021) . Our present study has several limitations, including possible selection bias, absence of data about existing medical conditions, medication use and sleep/circadian disorders (described in the first publication of the GCCS study by Korman, Tkachev, et al., 2020) . Nonetheless, the large sample size, ethnic and PreSR-inSR directions of change in OLE, SD and MST parameters: participants were exposed to less OLE and slept longer and later inSR. (b) Four aspects of wellbeing (Sleep-Quality, Quality-of-Life, Screen-Time, and Physical-Activity) that significantly correlated with changes in OLE, SD and MST, by wellbeing Δscore subgroups: negative change (−), no change (=), positive change (+); numbers, % of total. The "staircases" show which subgroups within each wellbeing aspect were significantly different from each other in terms of respective changes in OLE (yellow), SD (blue) and MST (green). inSR, inSocialRestriction; preSR, preSocialRestriction not only a factor of resilience during the pandemic, but also a probable remediating factor. In summary, strategies to improve wellbeing under social restrictions and to accelerate COVID-19 recovery should actively foster spending more daytime outdoors and keeping good sleep hygiene.

All authors declare no competing interest.

MK and TR designed research; MK, VT and TR performed research;

MK, TR, CR, YK, SK, DG, and VK contributed translations of the GCCS to different languages and advertised the study in their countries; MK, VT and TR analysed data; MK, VT, CR, YK, SK, DG, VK and TR wrote the paper.

We included all the data needed for the evaluation of the conclu- 

Dealing with sleep problems during home confinement due to the COVID-19 outbreak: Practical recommendations from a task force of the European CBT-I Academy

Phase relations between a circadian rhythm and its zeitgeber within the range of entrainment

Sunlight exposure increased Covid-19 recovery rates: A study in the central pandemic area of Indonesia. The Science of the Total Environment

New directions for the treatment of depression: Targeting the photic regulation of arousal and mood (PRAM) pathway

Evening use of light-emitting eReaders negatively affects sleep, circadian timing, and next-morning alertness

Sleep and Resilience during the COVID-19 Pandemic

Effect of light on human circadian physiology

Prevalence of depression symptoms in US adults before and during the COVID-19 pandemic

Trajectories of anxiety and depressive symptoms during enforced isolation due to COVID-19 in England: A longitudinal observational study

Sleep during COVID-19 lockdown: A cross-cultural study investigating job system relevance

Sleep health early in the coronavirus disease 2019 (COVID-19) outbreak in the United States: Integrating longitudinal, cross-sectional, and retrospective recall data

Sleep problems during the COVID-19 pandemic by population: A systematic review and meta-analysis

Implications of circadian rhythm in dopamine and mood regulation

Sleep quality during the COVID-19 pandemic: not one size fits all

ADHD 24/7: Circadian clock genes, chronotherapy and sleep/wake cycle insufficiencies in ADHD

COVID-19-mandated social restrictions unveil the impact of social time pressure on sleep and body clock

Impact of daily high dose oral vitamin D therapy on the inflammatory markers in patients with COVID 19 disease

Latitudinal cline of chronotype

Effects of lockdown on human sleep and chronotype during the COVID-19 pandemic

The antidepressant effect of light therapy from retinal projections

Light in man's environment

Circadian phase, dynamics of subjective sleepiness and sensitivity to blue light in young adults complaining of a delayed sleep schedule

Association between social jetlag food consumption and meal times in patients with obesity-related chronic diseases

Psychological symptoms during the two stages of lockdown in response to the COVID-19 outbreak: An investigation in a sample of Citizens in Northern Spain

Bright light improves vitality and alleviates distress in healthy people

Sleep and light exposure across different levels of urbanisation in Brazilian communities

Simulated sunlight rapidly inactivates SARS-CoV-2 on surfaces

A marker for the end of adolescence

Chronotype and social jetlag: A (self-) critical review

Daylight saving time and artificial time zones -A battle between biological and social times

An open trial of light therapy in adult attention-deficit/ hyperactivity disorder

Adjunctive bright light therapy for bipolar depression: A randomized double-blind placebo-controlled trial

Circadian entrainment to the natural light-dark cycle across seasons and the weekend

The impact of the early phase of the COVID-19 pandemic on mental-health services in Europe

Chronotherapeutics (light and wake therapy) in affective disorders

Natural' light treatment of seasonal affective disorder

The relevance of daylight for humans

Social jetlag: Misalignment of biological and social time

Decreased psychological well-being in late 'chronotypes' is mediated by smoking and alcohol consumption. Substance Use and Misuse

Digital screen time during COVID-19 pandemic: Risk for a further myopia boom?

Sleep in university students prior to and during COVID-19 stay-at-home orders

Entrainment of the human circadian clock to the natural light-dark cycle