Adolescents’ interactive electronic device use, sleep and mental health: a systematic review of prospective studies

Summary Optimal sleep, both in terms of duration and quality, is important for adolescent health. However, young people's sleeping habits have worsened over recent years. Access to and use of interactive electronic devices (e.g., smartphones, tablets, portable gaming devices) and social media have become deep‐rooted elements of adolescents’ lives and are associated with poor sleep. Additionally, there is evidence of increases in poor mental health and well‐being disorders in adolescents; further linked to poor sleep. This review aimed to summarise the longitudinal and experimental evidence of the impact of device use on adolescents’ sleep and subsequent mental health. Nine electronic bibliographical databases were searched for this narrative systematic review in October 2022. Of 5779 identified unique records, 28 studies were selected for inclusion. A total of 26 studies examined the direct link between device use and sleep outcomes, and four reported the indirect link between device use and mental health, with sleep as a mediator. The methodological quality of the studies was generally poor. Results demonstrated that adverse implications of device use (i.e., overuse, problematic use, telepressure, and cyber‐victimisation) impacted sleep quality and duration; however, relationships with other types of device use were unclear. A small but consistent body of evidence showed sleep mediates the relationship between device use and mental health and well‐being in adolescents. Increasing our understanding of the complexities of device use, sleep, and mental health in adolescents are important contributions to the development of future interventions and guidelines to prevent or increase resilience to cyber‐bullying and ensure adequate sleep.


| Sleep problems in adolescents
Adequate sleep quantity and quality is imperative to the health and well-being of young people (Dewald et al., 2010). Sleeping difficulties are common in adolescence (Inchley et al., 2020), and the biological and physiological changes occurring at this age can have an impact on sleep (Galván, 2020). The recommended sleep duration for adolescents, aged 13-18 years, is 8-10 h/day (Paruthi et al., 2016). However, there has been a detrimental shift in adolescent sleeping habits in recent decades, which means that many adolescents are not getting enough sleep (Hysing et al., 2015;Twenge et al., 2017). Conklin et al. (2019) argue that young people are suffering from chronic sleep disturbance, rather than deprivation, including issues such as insomnia and difficulty initiating or maintaining sleep.

| The link between interactive electronic device (IEDs) use and sleep problems in adolescents
Over recent decades, IEDs such as smartphones and tablets have become essential to young people's development and ability to communicate with others (Vanden Abeele, 2016). Increases in IED use have run in parallel with the shift in adolescent sleeping habits, yet there is still relatively limited understanding of the implications that increased IED use might have on sleep (Lund et al., 2021).
For over two thirds of adolescents, the final activity engaged with before bed, at least three times per week, involves the use of an IED, and one-third report using these devices in darkness prior to sleep (Kubiszewski et al., 2014;Mireku et al., 2019). Crucially, night-time access to and use of IEDs has been associated with higher odds of poor-quality sleep amongst young people (Carter et al., 2016;Hysing et al., 2013;Mireku et al., 2019).
Additionally, 93% of 13-17-year-olds have at least one social media account, and adolescents report engaging with such accounts for nearly 3 h/day on average (Barry et al., 2017). Studies have consistently found associations between social media use and sleep quality (Simsek & Tekgül, 2019;Twenge et al., 2017;Woods & Scott, 2016).

| The link between sleep problems and mental health
Investigations into sleep problems in adolescents have highlighted their inextricable link with mental health outcomes (Owens et al., 2014;Scott & Woods, 2019). In the UK, the number of young people aged 5-15 years with a probable mental health disorder has increased over recent years to one in six (NHS Digital, 2020, NHS Digital, 2018. Going to bed later, spending less time asleep on weeknights, and oversleeping at weekends, has been shown to increase the likelihood of mood and anxiety disorders in adolescents (Zhang et al., 2017).
Some evidence points to a bi-directional relationship between sleep and mental health (Cortese et al., 2013), whereby mental health disorders, such as depression and anxiety, are also predictors of inadequate sleep amongst young people (Shochat et al., 2014). However, this is contested; a well-conducted meta-analysis suggests sleep problems exist only as a precursor to mental health disorders, and that current evidence for such disorders being predictive of poor sleep is weak (Lovato & Gradisar, 2014). It appears that the 'dominant pathway' is from poor sleep to the occurrence of mental health disorders; and importantly, sleep problems can be treated, with positive effects on mental health (Freeman et al., 2020).
Both sleep disturbance and IED use have been identified as risk factors for depression (Lemola et al., 2015;Mireku et al., 2019;Thomee, 2018). Notably, Lemola et al. (2015) found sleep disturbance to be a partial mediator in the relationship between IED use and depressive symptoms; and highlight that most other studies consider sleep as a covariate rather than a mediator. Therefore, there is a need for further exploration of the potential mediating role of sleep in the relationship between IED use and mental health, and to use this information to inform interventions to improve sleep.

| Rationale for the present study
Current advice on IED use is based on reviews of literature focused largely on television viewing, but nowadays adolescent IED use largely involves mobile devices (Scottish Government, 2017). Given the constant and rapid developments in IEDs and applications (apps), up-todate research on the effects of IED use on sleep and mental health and well-being is needed to keep up with the adoption of novel technologies. Literature reviews on this topic have also focused mainly on cross-sectional studies, which are unable to show causal or temporal relationships (Carter et al., 2016;Lund et al., 2021;Shochat et al., 2014;Thomee, 2018;Yang et al., 2020). Furthermore, the direct relationship between IED use and mental health has been recently investigated in a meta-analysis of 531 studies, demonstrating a small bi-directional association, but the indirect link involving sleep was not explored (Shin et al., 2022). As data based on up-to-date IED preferences is becoming increasingly available, an updated systematic review is necessary to consider implications for advice on adolescents' IED use.

| Research questions (RQs)
The aim of this systematic review was to answer the following RQs: RQ1. to what extent is adolescents' IED use associated with sleep outcomes?
RQ2. what is the potential role of sleep as a mediator or mechanism between screen time and mental health and well-being outcomes?

| METHODS
The reporting of this study was guided by the Preferred Reporting of Systematic Reviews and Meta-analysis (PRISMA) (Page et al., 2021).
This review is an updated and more focused version of a previous rapid review commissioned by the Scottish Government (Martin et al., 2019

| Study selection and data extraction
The study selection process was performed by a total of six reviewers (AJ, AM, GOD, JP, KS, JR, and CBS) using Endnote reference T A B L E 1 Inclusion/exclusion criteria

Criteria Inclusion Exclusion
Population Young people with a mean age of 10-19 years. Studies including children aged <10 years or adults aged >19 years alongside young people were only included when results were reported for young people separately.

Exposures
Engagement with mobile (i.e., portable) interactive electronic devices (IED; e.g., smartphones, tablets, laptops) and software accessible through mobile IEDs (e.g., social media, games, websites, messaging applications), including studies that assessed the effectiveness of mobile applications (apps) or websites designed to improve sleep or mental health outcomes. Interventions with parents as agents of change with reports of adolescents' mobile IED use were also included.
Studies which referred to screen time in general without specifying the device and/or specific use (e.g., an app). This was to ensure that we captured evidence on contemporary screen technology rather than older screen technology such as televisions. Outcomes

| Study quality assessment
The Cochrane Collaboration tool for assessing risk of bias was used (Higgins et al., 2011). Five quality domains were assessed: (i) selection bias, (ii) performance bias (i.e., bias in assessment of the exposure), (iii) detection bias (bias in assessment of the outcome), (iv) attrition bias, and (v) selective reporting bias. Studies were judged to be of 'high', 'unclear' or 'low' risk of bias. One reviewer appraised the quality of all included studies (of AJ, GOD, JR, or CBS), which was crosschecked for accuracy by another reviewer (of AM, GOD, or LM).

| Data synthesis
Given the substantial heterogeneity across studies in terms of exposure assessment tools, outcome types and measurement, follow-up, and effect estimates, a quantitative synthesis with meta-analysis was deemed inapplicable. Hence, a narrative synthesis of the study findings was conducted and presented in line with the Synthesis without meta-analysis guideline (Campbell et al., 2020).
Studies were thematically grouped into five main exposure categories based on the different type of IED use described in Table 2.
Summary tables for each RQ (RQ1: IED use and relevant sleep outcomes; and RQ2: IED use and mental health outcomes) where sleep is a mediator were generated, mapping individual study results to each IED-use category, and a vote-counting approach was used to compare the number of results showing significantly positive or negative associations, or no association.

| Characteristics of included studies
Tables 3 and 4 present an overview of study characteristics, study exposure, mediator, direction of effects and risk-of-bias scores of reviewed papers. Detailed characteristics and results of reviewed studies are available in supplementary Tables S3 and S4 in Appendix S1.
2. Social media use Any interaction with a social media platform, whether actively or passively, at any time of day.

IED screen time/ brightness
Time exposed to the screen light from any IED at any time of day.
Negative consequences following the use of IEDs. Addiction to, dependency to, or a maladaptive pattern of use of IEDs A psychological state in which the individual has a preoccupation with or urge to respond quickly to messagebased communications on IEDs. The experience of being victimised by others via the use of IEDs.

Smartphone applications as sleep aids
The use of smartphone applications to improve sleep and mental health outcomes.
Regarding exposures, 10 studies used a form of general 'IED use' (references 1-10), four studies used 'IED screen time or brightness' (ref-   Table S5 in Appendix S1). There was a and high in only one (4%) study (reference 7).  Table S6 in Appendix S1). There was a high risk of selection bias in one study (25%) due to use of a convenience sample with limited generalisability (reference 28). A high risk of performance bias was detected in two (50%) studies (references 8, 27), with two (50%) studies judged as unclear due to lack of exposure measurement information or psychometric properties (references 12, 28). Detection bias was low in three (75%) studies that used validated questionnaires or scales to measure mental health and wellbeing outcomes (references 8, 12, 27), and unclear in one (25%) study where a previously validated questionnaire was adapted by the authors (reference 28). There was a high risk of attrition bias in one (25%) study (reference 27), with high levels of drop out or exclusion from analysis. Selective reporting was generally adequate across studies with two (50%) low risk of bias (references 27, 28) and two (50%) unclear risk of bias studies (references 8, 12). use and sleep quality, where interactive screen time was in fact associated with better sleep quality in males.

Mobile phone use/ownership
One interventional study, and two longitudinal cohort studies examined the relationships between mobile phone use and sleep outcomes (references 7-9). Adolescents (n = 63) who participated in an intervention involving instructions to stop mobile phone use 1 h before bed stopped using mobile phones an average of 80 min earlier and participants also turned their bedroom lights off earlier and slept 21 min longer compared to baseline (reference 7). However, participants did not go to bed earlier and did not fall asleep any quicker (ref- . Similarly, reported relationships reported amongst the longitudinal studies were conflicting (significant negative associations: four of 11 [36%], no significant associations: four of 11 [64%]). It is worth noting that the risk of bias was high amongst these studies.
Assuming smartphone ownership implies use of the device, one study involving 591 adolescents showed that smartphone owners were significantly more likely to have shorter sleep duration than nonowners, and the prevalence of sleeping problems increased over time among new owners to reach the prevalence among more long-term smartphone owners (reference 10). However, this study was judged to have a high risk of bias in three out of the five domains.

| Social media use
Across five studies (n = 3014 participants) reporting associations between social media use and sleep, eight out of a total of 33 (24%) results demonstrated negative associations between social media use and sleep outcomes. This included more frequent use of Twitter, and shorter tweets on weekday late nights, and fearful tweets associated with poor quality sleep (reference 11), problematic use of social networking sites associated with increased sleep disruption (reference 12), and links between social media use or social media stress with later bedtimes (reference 1), sleep latency (reference 13), and daytime F I G U R E 3 Risk of bias across studies assessing the potential role of sleep as mediator between screen time and mental health and well-being outcomes.
sleepiness (reference 13). Conversely, tweeting more frequently and longer tweets on weekday evenings were associated with better quality sleep (reference 11), and 23 (70%) of the reported results showed no association between social media and sleep.

| Interactive electronic device screen time/ brightness
Four studies considered the effects of the duration of IED screen time or screen brightness on sleep outcomes, with conflicting results.
Whilst one study with 26,205 participants found no longitudinal effect on sleep when students increased their screen use of any type In a laboratory study comparing three pre-bedtime Apple iPad screen light conditions (bright, filtered short-wavelength, and dim) in 16 adolescents, there were no differences in any self-assessed or objectively measured sleep outcomes (reference 18).

| Adverse implications of IED use
Overuse or problematic use of IEDs Five studies, all conducted in Asia (n = 13,278 participants), assessed the relationships between sleep and IED overuse or problematic use with six out of eight reported results (75%) indicating significantly negative associations. Increased mobile phone or smartphone addiction or overuse were associated with poorer sleep quality (references 19, 22) and sleep quantity (references 17, 22), and problematic internet use was associated with having a later bedtime and increased sleepiness after awakening in the morning (reference 21), whereas mobile phone dependency was not found to be associated with sleep duration (reference 20).

Telepressure
One study (n = 241 participants) found that an increase in telepressure (experiencing pressure to socially engage using a mobile phone) was associated with poorer sleep hygiene (reference 23).

Cyber-victimisation
Four studies (n = 31,701 participants) examined the effects of cybervictimisation on sleep outcomes (references 14, 22, 24, 25  Other research further suggests that it may not simply be the 'screen time' but specific effects of IED use and social media that impacts sleep, compared to more passive screen-based activities such as television watching (Hale & Guan, 2015). Social media inherently has more personally relevant information than gaming or viewing content (Nesi et al., 2018). Additionally, social feedback and one's standing in a peer network are important in adolescence (Nelson et al., 2005). However, while there are rewarding components of social media use (e.g., positive comments, 'likes', new followers), there are associated negative impacts (e.g. negative comments, absence of feedback, cyber-bullying) (Shapiro & Margolin, 2014 4.1.2 | What is the potential role of sleep as mediator or mechanism between screen time and mental health and well-being outcomes?

| Mediating role of sleep on mental health and well-being
There is an abundance of evidence supporting associations between IED use or social media and adolescent mental health and well-being (Keles et al., 2019;Shannon et al., 2022;Shin et al., 2022); however, it has been argued that sleep is amongst one of the largest contributing factors to adolescent well-being (Gireesh et al., 2018). We believe this to be the first review to specifically consider the role of sleep as a mediator or mechanism between IED use and mental health and wellbeing outcomes. We identified a small, but consistent body of evidence suggesting that changes in sleep behaviour acts as a mediator in the relationship between changing IED use and both mental health and well-being outcomes.
Poor sleep quality has been associated with lower competencies in control and perception of one's emotions (Brand et al., 2016), alongside an impaired ability for behavioural and emotional regulation in youth (Clarke & Harvey, 2012). Coupled with changing peer dynamics within adolescents' social and peer group makeup (reference 8), it may well be that IED use provides the conduit to impact sleep as a gateway for problematic social media use and cyber-bullying, leading to ruminating thoughts and fear or anxiety throughout the night, especially when it occurs in the hours before bed (reference 28).
Interestingly, a small number of included studies observed gender differences in the relationships between IED use, sleep, and mental health. One study found that displacement of sleep mediated the relationship to a much smaller degree in boys compared to girls (reference 27). Similarly, in another included study the magnitude of the relationship between mobile phone addiction and sleep quality was greater in girls (reference 19), and in a third study gender was found to moderate the effect of IED screen time on sleep quality, where interactive IED time was associated with better sleep quality in males, but poorer sleep quality in females (reference 2). Boys and girls have been shown to engage in electronic media differently, with boys spending more time gaming, and girls spending more time on smartphones and social media (Twenge & Martin, 2020). Further suggesting that the type of IED use (e.g., passive versus active, communication-based) could be important, as well as gender differential media use and underlying mechanisms, which warrants further exploration.

| Future research recommendations
We identified three interventional studies that targeted pre-bedtime IED use, with modest but promising results, including improvements in sleep duration, sleep quality, and sleep onset latency. However, results of these studies are limited by their respective study designs, where none are based on the 'gold standard' of randomised controlled intervention. Therefore, we did not give them any more weight than other study designs. Future interventions including education programmes to increase awareness of the potential risks of problematic IED use, prevent, or increase resilience to cyber-victimisation, and ensure healthy sleep should be considered. Ideally interventions should be tested within a fully powered randomised controlled trial.
Additional considerations worthy of investigation include: (i) is there an impact from the nature of viewing screens up-close compared to traditional screen viewing practises associated with television; (ii) does the use of second or third screens (e.g., television, laptop, and smartphone together) have increased impact on sleep; and (iii) does the nature of the content being engaged with impact the effects on sleep?

| Strength and limitations of the review process
The strengths of this review include comprehensive and up-to-date literature searches across nine electronic databases, the inclusion of longitudinal and experimental studies, robust risk-of-bias assessments, and the categorisation of exposures to consider the different types and complexity of IED use. There are some limitations to this review including the exclusion of studies not published in English, which may have introduced language bias. Substantial heterogeneity across studies in terms of outcome and exposure measurement methods, different metrics reported, and effect estimates meant that we were unable to perform meta-analysis. To synthesise the results, we applied a vote counting approach; however, this methodology has limitations. The magnitude of effects and study sample sizes are not accounted for, and studies that report multiple results (e.g., some studies reported up to six different, but relevant, exposures and up to 10 different sleep outcomes) are ultimately given more weight than those that report a single result. We were also unable to examine the potential presence of publication bias.

| CONCLUSION
This systematic review summarises the current longitudinal and experimental evidence of the impact of IED use on adolescent's sleep and subsequent mental health. Across the 28 included studies results demonstrated that it may not be as simple to suggest that IED use alone is the key factor, and that adverse implications of IED use including overuse, problematic use, telepressure and cybervictimisation can have an impact on several sleep characteristics. Evidence suggests that sleep acts as a mediator in the relationship between IED use and mental health and well-being in adolescents.
The methodological quality of the studies was generally poor, so findings must be interpreted with caution. The risks of problematic IED use or overuse, especially before bed, and potential impacts on sleep and subsequent adolescent mental health and well-being are important areas of focus, and greater awareness and understanding would enable the development of future interventions and policies.