Associations between screen use, outdoor time/daylight exposure and sleep changes during the first COVID‐19 lockdown in French children from the ELFE and EPIPAGE2 birth cohorts

Abstract Aims To investigate associations between outdoor and screen time and changes in sleep patterns in children from two nationwide birth‐cohorts in the SAPRIS project. Methods During the first French COVID‐19 pandemic lockdown, volunteer parents of children enrolled in the ELFE and EPIPAGE2 birth‐cohorts completed online questions about their child's outdoor time, screen time, and changes in sleep duration and quality compared with the pre‐lockdown situation. In 5700 children (aged 8–9 years, 52% boys) with available data, we assessed associations between outdoor time, screen time, and sleep changes using multinomial logistic regression models adjusted for confounders. Results Children spent on average 3 h08 outdoors and 4 h34 using screens/day (3 h27 for leisure, 1 h07 for class‐work). Sleep duration increased in 36% of children and decreased in 13.4%; sleep difficulties appeared/increased in 22.5% and decreased/disappeared/remained stable in 18.3%. After adjustment, increased screen time, especially for leisure, was associated with increased and decreased sleep duration (OR(95%CI) = 1.03(1.00–1.06) and OR = 1.06(1.02–1.10), respectively). No association was observed between outdoor time and sleep changes after adjustment. Conclusions Our study adds evidence for the association between high leisure‐time screen time and shorter sleep time. It supports current screen guidelines for children, especially during leisure time and for those whose sleep duration is short.


| INTRODUC TI ON
Sleep is essential for human life. Circadian sleep-wake rhythmicity is associated with the regulation of many functions, including metabolic, hormonal, cardiorespiratory, immune, neurological, cognitive, and psychological functions. [1][2][3][4] It is largely regulated by light, via the stimulation of melanopsin ganglion cells and the suprachiasmatic nucleus of the hypothalamus, resulting in the inhibition of melatonin secretion and activation of the wakefulness systems. 5 These effects depend on (i) the light spectrum (wavelengths), with a maximum effect around 480 nm (blue light from LED bulbs, computer screens or smartphones); (ii) the light intensity; (iii) the exposure duration, and (iv) the exposure period, with a maximum effect at bedtime and at dawn. 6,7 Thus, inappropriate light exposure (less exposure to outdoor light or overexposure to blue light, especially in the evening) could lead to disturbances of the circadian rhythm and therefore of sleep. [8][9][10] Sleep disturbances in children are associated with current and subsequent health, behavioral, and cognitive difficulties. [11][12][13][14] On December 31, 2019, the first cases of COVID-19 were reported in Wuhan, China. A few months later, this disease became a pandemic, with 87,137 cases listed worldwide as of March 1, 2020. 15 This unprecedented situation led to lockdown measures -sometimes successive -in many countries around the world. In France, the first strict general lockdown, including school closures, was declared on March 16, 2020 and ended on May 11, 2020. This was followed by a period of restrictions, still including school closures, travel restrictions and a ban on gatherings of more than 10 people until June 2. These periods changed daily behaviors and activities, which had effects on exposure to outdoor light and screens and on sleep patterns.
Several studies reported an increase in both qualitative and quantitative sleep difficulties during lockdown in adults. 16,17 In preschoolers, evidence is less consistent. Some studies report shorter sleep duration and lower quality, and others the inverse. [18][19][20] In school-age children and adolescents, studies show a reduced sleep quality, with an alteration in circadian synchronization, and an increased sleep duration. 21,22 In France, sleep durations are longer than in other countries with similar socioeconomic status, 23,24 so the impact of lockdown on sleep durations and difficulties may be different. In addition, very few studies have examined the associations between screen use, outdoor time/daylight exposure, and sleep in children.
In this study, we aimed first at describing school-aged children's behaviors, their changes during the lockdown, particularly regarding (i) changes in sleep duration and sleep difficulties, (ii) exposure to screen use (overall time and by purpose of use), and (iii) outdoor time/daylight exposure; and second at analyzing the associations between screen use, outdoor time/daylight exposure, and sleep changes during the first French lockdown.

| Study participants
Parents and children included were those recruited in 2011 from the ELFE 25 and EPIPAGE-2 26 birth cohorts and who agreed to participate in the SAPRIS project ("SAnté, Perception, pratiques, Relations et Inégalités Sociales en population générale pendant la crise COVID-19"), during the first lockdown in spring 2020. 27 Briefly, ELFE is a nationwide, multidisciplinary, birth cohort study, which included

| Data
Data were collected in the SAPRIS project from parents through two internet questionnaires. The first questionnaire was available from April 16 to May 4, 2020 (T1) and the second one from May 5 to May 31, 2020 (T2). Information on sleep and light exposure was asked only once at either T1 or T2. We considered the responses collected at T1 and those at T2 when missing at T1.

| Sleep and light exposure
Collected sleep data focused on change in both sleep duration and sleep difficulties during the lockdown compared to before. The questions asked to the parents were: "Would you say his or her sleep duration has changed since the lockdown began?" (No, it is as usual; yes, it has increased; yes, it has decreased), and "Would you say that since the lockdown began, your child has had difficulty sleeping (such as difficulty falling asleep, waking up at night frequently or too soon with no opportunity to fall back asleep)?" (That have appeared; that have increased; that have decreased; that have disappeared, that have remained stable; no difficulty). We grouped the categories "that have appeared" and "that have increased" in one category "Appeared/increased" and the categories "that have decreased," "that have disappeared," and "that have remained stable" (1.4% of the children) into one category "Stable/improved".
Outdoor time/daylight exposure was estimated from two questions about the last 7 days, "How much time per day did your child spend playing sports or walking outside home?" and "How much

| Confounding factors
Confounding factors were selected based on both the literature and a directed acyclic graph (DAG). 28 The considered socioeconomic factors were: maternal education (bachelor's degree, bachelor's degree +2 years, > bachelor's degree +2 years), household socioprofessional category (upper and middle management; employed or self-employed; blue-collar and inactive), family structure (2 parents, 1 parent or alternating custody), single child (no, yes), and maternal age; more specifically during this lockdown period: household financial status (affluent, income constant; affluent, income decreasing; modest, income constant; modest, income decreasing), work status within the couple (neither works; one teleworks, one does not; at least one works outside the home; both are teleworking), dwelling with garden or yard (no, yes), living area (rural, urban), dwelling occupancy index calculated as the ratio of the number of rooms to the number of persons living in the dwelling. Children' factors were gender, birth term, presence of a chronic or developmental pathology (no, yes), time of physical activity indoors during the survey (hour/day), overall behavior assessed by the Strengths and Difficulties Questionnaire (SDQ), and the presence of at least one of the following symptoms in the last 15 days: fever, unusual fatigue, muscle aches/pains, breathing difficulties/unusual shortness of breath, runny nose, pharyngitis, sore throat, febrile conjunctivitis, trouble smelling or tasting, nausea/vomiting, diarrhea, chest pain/oppression, appearance of persistent patchy frostbite on feet, hands, or face. The cohort (ELFE, EPIPAGE 2) and the wave of the questionnaire (T1, T2) were also taken into account.

| Statistical analysis
Analyses were performed using SAS® version 9. Multinomial logistic regression models, with either stable sleep time or no sleep difficulty as reference categories, were used to study the associations between the duration of exposure to light, whether outdoor or via screens, and the different sleep changes, without (M1) and with (M2) adjustment for family, child, and study characteristics. Interactions between outdoor light and screen exposure times were tested by adding multiplicative terms into the models.

| RE SULTS
Compared to the children aged 8-9 years old followed in the SAPRIS project and excluded from our analysis (n = 493), the included children (n = 5700) were more often from the ELFE cohort, from households with both parents having high socioeconomic level, working during the lockdown at home or outside and living in large dwelling (occupancy index >1). (Table 1).
The median and mean outdoor time/daylight exposure time were 2 h03 (Q1-Q3: 1 h26-3 h53) and 3 h08 (95%CI: 2 h58-3 h17) per day, respectively, at age 8-9 years old. Median and mean screen time (any kind and any use) were 3 h43 (Q1-Q3: 2 h25-5 h55) and  In France, during the first lockdown in March-May 2020, going outside homes was allowed for only 1 h per day to run essential errands or take out pets. However, we reported a mean outdoor TA B L E 2 Raw and adjusted associations between both outdoor time/daylight exposure and screen use (total, for leisure and schoolwork), and sleep changes during the first lockdown (N = 5700). All durations are in hours/day. The mean and median screen time in the studied French children (4 h34 h and 3 h43 per day, respectively) are lower than those reported in the USA, Italian, or worldwide school-aged children.
Indeed, North-American children used screen for a mean duration of 3 h40 during lockdown, 31 Italian school-aged children (6-17 years, mean 13 years) for 7 h39, 33 and 69% of those Italian children aged 12-13 years were exposed to screens for more than 3 h per day. 32 This prevalence raised 83% of children from different continents (Americas, Middle East, and Europe) aged 6 to 12-13 years were exposed to screens for more than 3 h per day. 29 The difference might be explained by the large opportunity for the studied children to access outdoor activities within a private or collective yard or garden.
This information is not reported in the other studies. One important limitation of our study is that the screen use questionnaire was not validated and suffered from measurement bias, with a substantial number of invalid screen time (i.e., daily screen time greater than 24 h per day) that we excluded from our analyses.  29 Those studies also reported association between increased screen time and decreased sleep time during lockdown. 29,30 We report here that each hour of leisure screen time was associated with increased odds of decreased sleep duration compared to before the lockdown. Screen for leisure is also highly consumed in the evening, 40,41 since up to 75% of children aged 6-17 years have screens in their bedroom. 42 It is now recognized that exposure to blue light from screens in the evening blocks the secretion of melatonin, a hormone that promotes sleep. 6,43 In addition, excessive screen time, especially before bedtime, has been shown to promote a phase delay of the circadian timing system (later bedtime and later rising) and difficulties in falling asleep. 8,[43][44][45] In our study, we did not have information on the timing of screen use and therefore cannot confirm that these non-visual effects of light were involved in the decreased sleep duration observed, however, they cannot be excluded. Especially because we showed that screen use for school-work, usually done during day- In conclusion, this study shows that screen time, especially for leisure, was associated with sleep duration modifications during lockdown. It supports current screen guidelines for children, especially for leisure and for those whose have sleep troubles.

AUTH O R CO NTR I B UTI O N S
Alex Wilfried Kamga Fogno performed the analyses and wrote the first draft of the manuscript. Alexandra Rouquette designed the data collection, Claude Gronfier designed the study, Jonathan Y. Bernard and Sabine Plancoulaine designed the study and supervised the analyses. All co-authors were involved in both interpretation and critical revision of the manuscript. All co-authors approved the final version.

ACK N OWLED G M ENTS
The authors gratefully acknowledge all the volunteers in the ELFE and EPIPAGE2 cohorts.
We thank the staff of the SAPRIS study group who worked with dedication and commitment to collect and manage the data used for this study and to ensure continuous communication with the cohort participants.
The SAPRIS study group: Nathalie Bajos (co-principal investi- The data underlying this article cannot be shared publicly due to the privacy of individuals that participated in the study. The data will be shared on reasonable request to the cohorts' scientific committees.

CO N FLI C T O F I NTER E S T S TATEM ENT
No interest to disclose. The authors have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data underlying this article cannot be shared publicly due to the privacy of individuals that participated in the study. The data will be shared on reasonable request to the cohorts' scientific committees.