Digital media, book reading, and aspects of sleep and sleep-related fears in preschoolers: the Ulm SPATZ Health Study

High levels of digital media use with screen-based devices has been found to threaten several aspects of child and adolescent health. However, the rapid change in digital media use behavior in our society over the past decade has rendered much of the existing evidence outdated. The present work reports data ascertained from 2016 to 2019 on preschoolers’ digital media use and book reading among 4‑ to 6‑year-old children from the longitudinal Ulm SPATZ Health Study. These data are supplemented by exploring the association of child digital media use and book reading with aspects of children’s sleep and sleep-related fears. After data management, 581 children participated at the age of 4 years. At the ages of 5 and 6 years, data were available for 508 and 426 participants, respectively. Median age of the mothers was 33 years (range 21–54) and 70% had graduated from high school. Most time spent using digital media was represented by passive activities in front of a screen. Time spent using digital media increased with children’s age. An increased risk of the fear of sleeping alone (RR: 1.35, [95% CI: 1.07; 1.70]) and fear of the dark (1.47 [95% CI 1.16; 1.87]) was documented for exposure to “all media” higher than the median as compared to lower levels. Exposure to “books” above the median and compared to lower levels was tentatively associated with a relevant 18% reduction in the risk of both fear of sleeping alone and fear of the dark (0.82 [95% CI 0.65; 1.03] and 0.82 [95% CI 0.64; 1.05], respectively). Our results suggest that reducing time spent using digital media and potentially substituting this time by reading books, which seem to be a safe alternative, is a valid solution to reduce some common aspects of poor sleep in preschoolers.

High levels of digital media use with screen-based devices have been found to threaten several aspects of child and adolescent health [29,32]. In particular, media use has been identified as a potential determinant of circadian clock disruption. On one hand, a negative physiological effect of media use on sleeping has been hypothesized to act through multiple mechanisms such as melatonin suppression [34], and on the other hand, the media content and the timing were demonstrated to influence sleep in children and adolescents [16]. Several systematic reviews of varying quality are available on the effects of screen time on domains of sleep health such as sleep duration, prolonged bedtime, sleep onset problems, night wakings, or sleep problems [3,7,12,14]. The majority of published evidence is inferred from studies including school-aged or older children and adolescents [23], but two recent systematic reviews document the consistent detrimental effect of screen time for sleep health in preschoolers as well [23,35]. However, the overall quality of evidence in one of the two systematic reviews judged by part of the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) scale was low [23]. Longitudinal evidence on determinants of children's sleep behavior is emerging and has been summarized by a further systematic review [5]. Here, the authors found moderate evidence for screentime asa determinantofchild sleep duration [5]. Furthermore, two systematic reviews have set out to summarize the effect of interventions to control screen time on children's sleep [6,26]. While the studies suggest that small improvements in screen time and sleep duration can be achieved in children [26], the number of studies is limited and the evidence may be blurred by the presence of co-interventions, low methodological quality, and heterogeneity [26]. Indeed, evidence to guide policy on safe screen time exposure among children and adolescents is limited [32].
Still, guidelines on media use, also with regard to child sleep health, have been issued in the past years [1,18,19,30]. For Germany [18], as for other countries [1,19], the guideline does not give an explicit, age-adapted time limit spent with digital media per day that should not be exceeded. Other guidelines have stated at most 0.5 to 1 h per day for 2-to 5-year-olds [9,13,30]. The advent and spread of mobile devices, such as tablets and smartphones, has spurred some interest into the effects of the type of end device [29,33]. Another im-portant dimension of digital media use could be active (e.g., gaming) vs. passive (e.g., TV viewing) use [1]. However, the rapid change of digital media use behavior in our society over the past decade has rendered a lot of the existing evidence outdated. A scoping review on 5-to 18-year-olds including 130 population-based surveillance studies showed that on average about 50% of the participants exceed 2 h per day of screen time [33]. This proportion is mirrored in contemporary data on German 10to 18-year-olds [2]. The German Health Interview and Examination Survey for Children and Adolescents (KiGGS) has so far only published data on digital media use of 11-to 17-year-olds ascertained in the years 2009-2012 [25]. Data from the 2017 Spanish National Health Survey indicate that 67% of 3-to 5-yearolds spend ≥1 h per day on daily leisure screen time and that 36% even exceed 2 h per day [11]. Contemporary data on digital media use in German preschoolers are lacking, as they are in many other countries. Moreover, most existing data do not inform about the timing of media use across the day, although morning or evening use could have distinct effects on a child's sleep.
The present study seeks to close this gap by reporting descriptive data ascertained from 2016 to 2019 on preschooler's digital media use and book reading among 4-to 6-year-old children from the longitudinal Ulm SPATZ Health Study. These data are supplemented by exploring the association of child digital media use and book reading with aspects of sleep and sleep-related fears.

The Ulm SPATZ Health Study
The Ulm SPATZHealthStudyis a prospective birth cohort study based on 1006 children consecutively recruited after their birth between April 2012 and May 2013 at the University Medical Center of Ulm, Southern Germany. Baseline data were collected during the days following delivery. Subsequently, every year at the child's birthday, a battery of self-administered questionnaires is filled by participating families. These questionnaires concern a number of factors affecting children's and parents' health, in partic-ular diagnoses, psychological health and wellbeing, lifestyle, sleeping habits, and socioeconomic characteristics. The Ulm SPATZ Health Study has a prospective design with repeated measures aimed at investigating the wellbeing of the parents and their child. The present work includes data from study waves at the children's ages of 4, 5, and 6 years. Participation is voluntary and informed consent was obtained from all mothers of the participating children, their partners, and also of their siblings if older than 18 years. The Ulm SPATZ Health Study was approved by the ethics board of Ulm University (no. 311/11).

Children's sleep-related fears assessment
The German version of the Children's Sleep Habits Questionnaire (CSHQ) was used to collect parent-reported child sleep problems over a typical week [27]. For the present analyses, three items of the sleep-related fears domain and one question regarding waking up at night were selected according to previous evidence relating sleep to media use in SPATZ children [17]: "Needs parents in room to sleep, " "Afraid of sleeping in the dark, " "Afraid of sleeping alone, " and "Wakes up once during the night. " The items were scored on a three-level Likert scale, as "rarely, " "sometimes, " and "usually. " The CSHQ is a tool validated for the German population [31].

Digital media use and book reading
Parents reported the time their child spent using digital media and reading books separately for weekdays and weekends and for morning (06:00-12:00), afternoon (12:00-18:00), and evening (18:00-06:00). The four items (watching TV/DVD, video gaming, other computer/internet use, reading books) were coded as: never; up to 1 h/day; 1 to Abstract Background. High levels of digital media use with screen-based devices has been found to threaten several aspects of child and adolescent health. However, the rapid change in digital media use behavior in our society over the past decade has rendered much of the existing evidence outdated. Materials and methods. The present work reports data ascertained from 2016 to 2019 on preschoolers' digital media use and book reading among 4-to 6-year-old children from the longitudinal Ulm SPATZ Health Study. These data are supplemented by exploring the association of child digital media use and book reading with aspects of children's sleep and sleep-related fears.
Results. After data management, 581 children participated at the age of 4 years. At the ages of 5 and 6 years, data were available for 508 and 426 participants, respectively. Median age of the mothers was 33 years (range 21-54) and 70% had graduated from high school. Most time spent using digital media was represented by passive activities in front of a screen. Time spent using digital media increased with children's age. An increased risk of the fear of sleeping alone (RR: 1.35, [95% CI: 1.07; 1.70]) and fear of the dark (1.47 [95% CI 1.16; 1.87]) was documented for exposure to "all media" higher than the median as compared to lower levels. Exposure to "books" above the median and compared to lower levels was tentatively associated with a relevant 18% reduction in the risk of both fear of sleeping alone and fear of the dark (0.82 [95% CI 0.65; 1.03] and 0.82 [95% CI 0.64; 1.05], respectively). Conclusion. Our results suggest that reducing time spent using digital media and potentially substituting this time by reading books, which seem to be a safe alternative, is a valid solution to reduce some common aspects of poor sleep in preschoolers.

Keywords
Circadian clocks · Questionnaires · Sex · Sleep initiation and maintenance disorders · Fear

Schlüsselwörter
Zirkadiane Uhren · Fragebögen · Geschlecht · Ein-und Durchschlafstörungen · Angst <2 h/day, 2 to <3 h/day, 3 to <4 h/day, and ≥4 h/day. Values of time spent watching TV or DVD or using the computer or internet passively were aggregated and labelled as "TV/screen" whereas active video gaming was considered separately and labelled "gaming. " Both together were labelled and analyzed as "all digital media. " The categories were irrespective of the end device, i.e., the use of, e.g., a tablet, smartphone, or monitor/TV screen. Furthermore, the time spent on self or parent-assisted book reading was analyzed separately and labelled as "books. " Supplementary Table 1 reports the items investigating digital media use and book reading.

Statistical methods
All descriptive evaluations were stratified by sex and age. Digital media use and book reading were also described separately for weekdays and weekends. Values for time spent with "TV/screen, " "gaming, " "all digital media, " and "books" were reported by averages and the upper 95th centile. Independent-sample Mann-Whitney U test was used to compare the sexes at different timepoints. For this comparison, a type-I error rate of 5% was considered as statistically significant. The percentages of time spent on different activities by time of the day were reported for weekdays and weekends using pie charts. Responses to the four selected items of the CSHQ were reported by age and sex using bar charts. A correspondence analysis was applied to investigate the joint association of "books" and "all digital media" on the aforementioned sleeping features and sex. To this end, we first defined an indicator matrix of the median time spent with "books" and "all digital media" during weekdays and at weekends. This indicator matrix was transformed into a Burt matrix and then analyzed using a multiple correspondence analysis. Afterwards, children's sleeping features, aggregating "sometimes" and "usually" levels into "yes" and labelling "rarely" as "no, " and sex were passively plotted over the dimension defined by the correspondence analysis. Finally, a generalizing estimating equation (GEE) based on a binomial model was used to compute relative risks for all four sleep problems, comparing levels of "books" and "all digital media" averaged over weekdays and weekends. These models were adjusted for sex and subsequently stratified by sex. The sleep problems outcome of the logistic model was again coded according to the "sometimes" and "usually" levels vs. the level "rarely. " All statistical evaluations were performed using R version 3.6 (R: A language and environment for statistical computing; R Foundation for Statistical Computing, Vienna, Austria; https://www.Rproject.org/). The packages GEE and FactoMineR were used to perform GEE fitting and correspondence analyses, respectively.

Results
After exclusion of missing records and inconsistencies, 581 children (291 [50%] boys) participated at the age of 4 years. At the ages of 5 and 6 years, data were available for 508 participants (259 [51%] boys) and 426 participants (211 [50%] boys), respectively (a flowchart of participant selection is reported in Supplementary  Fig. 1). The median age of the mothers was 33 years (range 21-54 years), approximately 70% had graduated from high school, and 11% of the mothers had an independent or high-profile/leadership job (. Fig. 1).
Most time spent with digital media use was represented by time spent with passive activities in front of a screen. We also observed that time spent with digital media use increased with child age. There were no substantial differences between boys and girls at the ages of 4 and 5 years. In contrast, we found a statistically significant difference at the age of 6 years with girls spending 30 min/day more compared to boys on a weekday and 75 min/day more compared to boys on weekends. Moreover, we observed a larger use of digital media at weekends compared to weekdays in both boys and girls, with a difference of 20, 24, and 54 min at the ages of 4, 5, and 6 years, respectively. In boys, these differences were 20, 25, and 31 min, whereas the values were higher in girls, especially at age  We also detected that a comparable amount of time was spent with "books" when compared to "all digital media" at ages 4 and 5 years, both for boys and girls and during weekdays and at weekends. This shifted at age 6 years, when more time at the weekend was spent with "all digital media" compared to "books, " especially in girls. The time spent on different media activities between 4 and 6 years of age is reported in . Fig. 1 for boys and girls separately. Investigating the timing across the day, time spent passively with "TV/screen" and on "book reading" had a similar pattern, with less time spent on these activities in the morning compared to the evening (Supplementary Fig. 2).
On the contrary, more time was spent with "gaming" during the afternoon. No particular patterns were observed at different ages, or by sex, or during weekdays and at weekends, apart from a higher use of "all digital media" and "books" in the mornings of weekends (. Fig. 2, Supplementary Fig. 3).
Concerning sleep problems, a similar pattern of frequency of fears and tendency to wake up at night was found for boys and girls at 4 years of age. In particular, at 4 years of age, 40% of boys and 38% of girls usually need their parents to fall asleep. About 10 and 13% of boys and girls usually feared the dark or feared sleeping alone, respectively. Finally, 26% of the boys and 24% of the girls usually woke up at night. The need for parents in the room to fall asleep decreased with age in both boys and girls. At 6 years of age, 24% of the boys and 23% of the girls needed their parents to fall asleep. The frequency of waking up at night also decreased over time, down to 16% of the boys and 10% of the girls usually waking up at 6 years of age. In contrast, fears increased with age, with 18% of the boys and 20% of the girls usually having a fear of the dark at the age of 6 years. The distribution of items regarding sleep-related fears is reported in . Fig. 3.
The correspondence analyses of "books" and "all digital media" resulted in two well-defined dimensions. The first dimension separates high exposure to "all digital media" and low exposure to "books" from low exposure to "all digital media" and high exposure to "books. " This first dimension results in an explained variance of 53.5%, 50.2%, and 50.0% at the ages of 4, 5, and 6 years, respectively. The second dimension separates low exposure from high exposures to both "all digital media" and "books, " explaining the remaining 46.5%, 49.8%, and 50% of variance at the age of 4, 5, and 6 years, respectively. The three sleep-related fears (fear of the dark, fear of sleeping alone, and need of parent to fall asleep) cluster together at 4 years of age. They were related to higher exposure to "all digital media" and were apparently more common in girls. At the ages of 5 and 6 years, these relations are maintained for the fear of sleeping alone and fear of the dark. On the contrary, the need of parents to fall asleep appears to be a more independent factor at the ages of 5 and 6 years. The correspondence analysis also showed that waking up during night was more Notably, this sleep problem appears to be related to higher exposure to "books" at age 4, while at 5 and 6 years of age, it is related to lower exposure to "books. " Results of correspondence analysis of digital media use and book reading in association with sleep-related fears are reported in . Fig. 4.
Finally, the results from the GEE analyses adjusted for sex documented a statistically significantly increased risk of the fear of sleeping alone (relative risk [

Discussion
In the present work including n = 581 children aged 4 to 6 years recruited in the context of the Ulm SPATZ Health Study and located in the south of Germany, it was shown that more time is spent on digital media use and book reading at weekends compared to weekdays, which was also stated by previous studies in part including adolescents [15,22]. Moreover, we found that girls have higher exposure to digital media use but also to book reading compared to boys of the same age. Of note, total time spent with digital media exceeds recommendations for this preschool age group but is moderate compared to studies conducted in other societies such as the US [28].
Furthermore, in our study, sleep-related fears are different in boys and girls and this difference increased with age. Moreover, waking up at night was more common in boys than in girls of the same age, with an increase in this difference between 4 and 5 years of age. Furthermore, sleep-related fears were more common in girls than in boys of the same age. Also, we found that the need of parents to fall asleep is more frequent at 4 years of age, while its prevalence decreases with age irrespective of sex. However, to need a parent to fall asleep was stated as a criterion concerning childhood insomnia in the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5). Therefore, one may assume that about 25% of the children suffered from insomnia symptoms and might benefit from professional help.
Very importantly, we showed the relations of digital media use and book reading with sleep-related fears and waking up at night. In particular, we found that excessive digital media use may result in a higher risk of the fear of sleeping alone and fear of the dark, which was also stated by parents as an important reason for impaired sleep of young children [4]. On the contrary, we found a non-significant but suggestive association between book reading and a decreased risk of the fear of sleeping alone and fear of the dark. This might be in line with the finding that playing with activating toys is very sleep disturbing, alongside anxiety and fears [4]. Notably, these associations were observed in mutually adjusted models showing their independence. There are numerous possible explanations for why digital media use may cause sleep-related fears, as shown in previous studies [20]. Excessive digital media use was suggested to be related to fear and anxiety through mechanisms mediated by a possible lack of emotion-regulation skills [21]. Also, digital media use without the presence of a parent could come with a lack of interaction and effective guidance, which, in turn, could lead to a lack of confidence or self-regard, predisposing to sleep-related fears [8]. In contrast, book reading, especially if assisted by the parents, could increase the child's self-confidence, leading to reduced sleep-related fears.
The current work has numerous strengths. Firstly, it employed a rigorous approachbased ona prospective designin which data regarding sleep features were collected using a questionnaire validated for the German population. Secondly, the work was based on preschoolers, Hier steht eine Anzeige. K a population of great epidemiological interest for which there is still a lack of scientific evidence regarding sleeprelated fears in relation to digital media use. Thirdly, data on digital media use and book reading were explored, the latter seldom ascertained in studies on digital media use and sleep-related fears.
The current research could be improved by extension to other sleep disorder symptoms or disorders such as sleep onset delay or parasomnias such as nightmares. Moreover, the current work might be affected by measurement error and information bias, given the use of self-reported outcomes. The use of objectively assessed sleep outcomes would reinforce and improve the quality of the evidence provided; however, sleep-related fears are difficult if not impossible to ascertain objectively. Furthermore, the observational nature ofthe studyand the cross-sectional assessment of results at different ages potentially permits reverse causation, which could, for instance, drive the association of book reading with waking up at night in 4-year-old boys. Finally, we only in-vestigated sex-specific effects and did not explore relations with other potentially important factors such as socioeconomic status of the family. Therefore, residual confounding could have affected our results. However, in the present work we did not aim to accurately estimate causal associations; we merely sought to describe patterns of digital media use and book reading in contemporary data and to explore their potential association with common sleep-related problems in preschoolers.
Further studies, particularly with an experimental approach [24], considering objective measures of sleep such as actigraphy or polysomnography and possibly taking into account socioeconomic status [10] and other possible confounders are necessary to uncover the causal nature of the impact of digital media use and book reading on children's sleep. Until these are available, the current results suggest that reducing time spent using digital media-and potentially substituting this time by reading books, which seem to be a safe alternative-is a valid solution to reduce some common aspects of poor sleep in preschoolers.

Conclusion
Book reading should be promoted to entertain children by specific public health programs and campaigns in schools. On the other hand, scientific research should focus on valid experimental studies based on objectively assessed methods.