Sleep in young people: What works now and where to? A meta-review of behavioural and cognitive interventions and lifestyle factors

ABSTRACT

Objectives

1) Systematically review meta-analyses and systematic reviews that (a) explored health/lifestyle factors affecting sleep, and/or (b) investigated behavioral/psychological sleep interventions in young people (10–25-years); 2) Evaluate the quality of published literature, and, if an intervention; 3) Examine method and effectiveness of mode of delivery, to inform current clinical practice and research direction.

Method

A systematic search of Embase (n = 45), MEDLINE (n = 67), Web of Science (n = 375), Google Scholar (n = 138), and hand-searching was conducted. After full review, 12 papers were selected, 2 systematic reviews without, and 10 with, meta-analyses. Six examined associations between sleep and lifestyle/health, and six examined cognitive-behavioral (n = 4), or school education (n = 2), programs.

Results

Electronic media use, type of day (week/end), sex, age, culture/geographical location, substance use, family environment, and evening light exposure were negatively associated with sleep, in young people. Only cognitive and/or behavioral interventions of at least 2 × 1-hr sessions improved sleep.

Conclusion

This paper informs sleep recommendations for young people and advises that ≥ 2 × 1-hr sessions of cognitive behavioral or behavioral therapy is the minimum to improve sleep in young people. School-based sleep interventions do not produce long-term change.

Introduction

In young people (aged 10–24-years), poor sleep is associated with worse educational performance, cognitive and behavioral problems in school, and increased risk-taking behaviors (Astill et al., 2012). Possibly through reduced inhibitory control (Warren et al., 2017) and lower distress tolerance (Kechter & Leventhal, 2019), poor sleep has a demonstrated relationship with psychosocial difficulties (Verkooijen et al., 2018), criminal behaviors (Raine & Venables, 2017), and substance use (Marmorstein, 2017).

Sleep problems in young people are highly prevalent (Liang et al., 2021; Roberts et al., 2009), 25% report insomnia symptoms (Ohayan et al., 2000), 27% are at risk of a sleep disorder (Gaultney, 2010), and approximately one in four reports obtaining fewer than 6-h sleep per night: fewer than the 7–10 h recommended (Roberts et al., 2009). Less consensus exists around: 1) What health and lifestyle factors predict poorer sleep outcomes, including light exposure, electronic media use, type of day (weekday/weekend), geography/culture, substance use, exercise, and family environment, and; 2) What psychological and/or behavioral treatments work best in young people, the number of sessions required, and/or whether individualized, face-to-face, treatments are necessary or if group and/or internet-based therapies can be used (Gradisar & Richardson, 2015).

This paper provides: 1) A review of systematic reviews and meta-analyses that have (a) explored health and lifestyle factors that affect sleep, and/or (b) examined behavioral and/or psychological sleep intervention in young people (aged 10–24-years); 2) Reports on the quality of the published literature, and; 3) If a review of interventions, examines the method and effectiveness of mode of delivery (number of sessions; face-to-face vs digital-health; peer vs health specialist vs non-health specialist delivery).

This paper summarizes findings across the reviews and meta-analyses within this field in order to direct current sleep recommendations for young people and to inform future research.

Method

Registration

This review was pre-registered with PROSPERO (Reg. ID: 186906) and reported in accordance with the PRISMA guidelines (21).

Search strategy

The database search terms common to all databases were “Cognit*” OR “Behav*” AND “Intervention” AND “Sleep” (where * indicates a wildcard to allow for multiple word endings). For a database-specific set of search terms, search parameter restrictions and the number of reviews returned see the Online Supplement A (Table S1-S4).

In scoping searches, specific sleep-wake disorder terms from the DSM-5 (e.g., parasomnias, delayed sleep-phase syndrome, etc.) and modes of intervention (e.g., telehealth, digital health, face-to-face) were included however, this did not increase the number of relevant reviews returned. Leaving the term “Intervention” out of the search terms to find more lifestyle/environment reviews, broadened the number of returned papers expansively but did not return more pertinent reviews.

Search limits

The search was limited to journal articles, published in English, human research, full articles, focusing on adolescents (10–19-years) or young people (10–24-years), which were systematic reviews and/or meta-analyses. While the start date of the search was not limited to a specific date, no studies were returned published before 1990. These searches were started on the 18.01.20 and ended on the 21.06.2021. This review was rerun on the 05.05.2022, before submission for publication, however no new papers were returned. Papers that reviewed studies across the lifespan, including adults, infants, and children were kept provided data pertaining to the adolescent and young people subgroups could be separated from the overall results. Reviews were retained if they examined correlational associations between behaviors or cognitions and sleep quality or quantity. Special populations, e.g., those with Down’s syndrome were excluded. For more details on these searches, see the Online Supplement, Tables S1-S4.

Results

Study selection and data extraction

The selection procedure followed the order: 1) Abstract and title screening; 2) Full paper review; 3) Inclusion cross-checked by coauthors (MO; CR; KH), and; 4) Data extraction (see Figure S1, in Online Supplement A).

Data extracted, included: The number of papers examined; total sample size; the mean population demographics; type of studies included; the type of sleep disruption/disorders identified; description of the sleep intervention or factors that impact sleep in young people, and; the average number/type/duration of sessions. Where reviews did not summarize this information, individual studies were found, the data obtained, and summarized.

The acronym C/BT is used to describe all intervention studies that used cognitive behavioral techniques or behavioral techniques. CBT denotes cognitive behavioral interventions and behavioral denotes interventions of a behavioral nature alone, whilst CBT-I denotes CBT adapted specifically for insomnia.

Study characteristics

Table 1 provides an overview of the 12 reviews including: First author, publication year, and basic characteristics, describing 1,248,494 young people (Age M = 15.2 (range 11–30 years); female 61%). The research reviews included populations from many countries, representing each continent of the world. This included 18% from the US, 12% from Australia, 10% from Japan, 6% from the UK, 5% from Germany, 4% from countries including Korea, Switzerland, Belgium, and China, and other countries providing 0.4–3% of the total sample (Sweden, Taiwan, the Netherlands, New Zealand, France, Brazil, Finland, Israel, Canada, Spain, Italy, Norway, Iceland, India, Turkey, Saudi Arabia, Portugal, Iran, Lebanon, Russian Federation, and Kuwait).

Table 1. Description of the reviews examining factors affecting sleep and sleep interventions in young people, included in this meta-review.

	Review paper	n = papers	n = total sample	Sample details	Type of studies included	Type of disorder/disruption	Factors associated with sleep & Tx components	Length of interventions	Mode of interventions	Sleep outcomes
	Systematic reviews
1	Hale & Guan (2015). Sleep Med Rev., 21, 50–58	67 (35 in adolescents)	549,454	Subsample (n = 298,733) of adolescents and young adults (11–25 years)	Cohort studies; Cross-sectional studies	Poor sleep quality and/or quantity	Screen time and devices	No intervention	No intervention	Sleep timing; Sleep duration; Sleep quality; SOL; Daytime tiredness or sleepiness; Insomnia symptoms
4	Hall & Nethery (2019). Paediatr. Respir. Rev., 31, 64–74	44 (10 in adolescents)	282,615	Subsample (n = 109,081) of adolescents (13–18 years).	Exploratory trials (7); Intervention (3) studies	Poor sleep; Sleep hygiene	Sleep education; modifying technology exposure and/or environmental factors (e.g., passive smoking, school commute times); milk drinks to promote sleep	≥1 session	Group; Individual; Individual with parent; Individual with parents and teachers	Sleep duration; Sleep quality; Sleep onset
	Meta-analyses
5	Aslund et al. (2018). Clin Sleep Med., 14, (11), 1937–47	6	528	School aged children and adolescents; M age = 14.6 yrs; female = 63%)	RCT	Poor sleep	Cognitive and behavioral sleep interventions	≥4 × 1 hr Tx sessions	Group; Individual; Internet	TST; SOL
3	Mei et al ((2018)). Sleep Sci. Prac.; 2:9	19	253,904	School aged young adults, M age, 14.82 years; 51.1% male)	Clinical trials; Cohort studies; Cross-sectional studies	Poor sleep quality and/or quantity	Electronic media use	No intervention	No intervention	SOL; WASO; TST
2	Carter et al. (2016). JAMA Pediatr, E1–7.	20	125,198	School aged children and adolescents (6–19 years); M age = 14.6 yrs	Clinical trials; Cohort studies; Cross-sectional studies	Poor sleep quality and/or quantity	Portable screen-based media devices	No intervention	No intervention	Sleep quantity; Sleep quality; EDS.
6	McLay et al. (2020). J. Behav. Educ.	10 (3 in adolescents)	893	Subsample of adolescents (n = 162; M age = 17.7 yrs; female = 70.4%)	Intervention; RCT (1)	Sleep problems; Insomnia	Telehealth Interventions; CBT based	2 × 1 hr group/individual workshops + follow-up telehealth contact	Individual; Group; Parent led/supported	SOL; Night wakenings; TST; other-nighttime events
7	Blake et al. (2017). Clin. Child Fam. Psychol. Rev., 20, 227–49	9	357	Adolescents (M age = 15.0 yrs; female = 61.7%)	Intervention; RCTs (4)	Insomnia; poor sleep.	CBT based	≥4 × 1 hr treatment sessions	Group; Individual; Individual with parent; Parent education sessions	Subjective quality; TST; SE; WASO; SOL.
8	Chung et al., (2017). J Sch Health, 87, 401–8	7	1,876	Adolescents (M age = 14.7 yrs; female = 61.0%)	Intervention: RCT; Cluster RT	Short sleep	School-based sleep education programs	≥4 × 50-minute classes	Group	TST; SOL; Mood
9	Bartel et al. (2015). Sleep Med. Rev., 21, 72e85	41	85,561	Young adults 11–20 yrs	Correlational; cohort studies	Sleep disruption	Protective and risk factors for adolescent sleep;	No intervention	No intervention	SOL, TST; Self-reported sleep.
10	Olds et al. (2010). Sleep Med. Rev., 14, 371–78.	30	92,977	Sample of adolescents (11–18 yrs)	Correlational; cohort studies	Sleep patterns	Factors associated with time in bed	No intervention	No intervention	Sex, age, geographical region and day type interact and predict sleep patterns.
11	Gradisar et al. (2011). Sleep Med., 12, 110–18	41	371,758	Sample of adolescents (11–18 yrs)	Correlational; cohort studies	Sleep patterns	Factors associated with time in bed	No intervention	No intervention	Cultural factors, day type, and age interact to impact sleep pattern.
12	Niu et al. (2021). Sleep Med. Rev., 58, 101443.	17 (7 in adolescents M = 14.9 & 10 in young adults M = 21.7)	605	Sample of adolescents-young adults (13–30 yrs)	Intervention	Low sleep quantity	Behavioral sleep intervention	5–49 nights (41% of studies conducted over 7-days)	Extend sleep opportunity; reduce screen access; Prevent naps; Reduce caffeine consumption, and; Provide sleep education.	SE; Sleep quality; EDS

Note: Tx = Treatment; yrs = years’ SOL = Sleep Onset Latency; WASO = Wake After Sleep Onset; TST = Total Sleep Time; EDS = Excessive Daytime Sleepiness.

This sample of review papers includes 2 systematic reviews and 10 meta-analyses, of which 6 examined associations between factors that impact sleep in young people (1 review/5 meta-analyses) and 6 examined interventions for sleep disruption in young people (1 review/5 meta-analyses).

Of these intervention reviews, 2 addressed school-based education programs, 3 focussed on CBT-based interventions for poor sleep, sleep problems generally, and insomnia specifically, and the final meta-analysis examined behavioral interventions (BT) that aimed to extend sleep opportunity. These intervention reviews included a mix of formats, examining group, individual, individual with parent, and parent education sessions. While most reviews examined interventions of duration≥4 × 50–60-minute sessions, one included studies of ≥ 2 × 60-minute sessions. All studies reported on total sleep time (TST), and each reported a mix of sleep efficiency (SE), wake after sleep onset (WASO), sleep quality, excessive daytime sleepiness (EDS), and/or sleep onset latency (SOL).

Study quality

All 12 studies were systematic reviews and 10 included a meta-analysis of the selected studies. According to the Oxford Centre for Evidence-based Medicine (CEBM) (W, 2011) all of these articles are Level 1 evidence, indicating that this meta-review summarizes high research quality.

Quality was further assessed using quality rating criteria for systematic reviews (O’neil et al., 2011; Thompson et al., 2012). All papers were rated≥60% quality, with a mean rating of 78% quality. Further detail is provided in Online Supplement A, Table S5.

Findings

Pearson’s r was reported as an effect size, with effect sizes of .10–.30 considered a small effect, .31–.50 as a medium effect, and .51–1.00 as a large effect (Gignac & Szodorai, 2016). The results are summarized in Table 2. Where age ranges of the reviews examined did not cover the full range (10–25 years), we have provided a specific age range for the factor examined.

Table 2. List of lifestyle/associated factors affecting sleep quality and quantity as grouped across the reviews and the average effect size across reviews.

Factor affecting	# Reviews (article #)	Description and possible mechanisms	Key findings	Recommendations and future directions resulting from the included paper/s*
Electronic media use	5 (1;2;3;4;9)	Impact of use of electronic devices with screens, particularly in the evening, including television, computers, gaming, mobile devices, and unspecified screens on sleep. Possible causal pathways: Exposure to blue wave length light emissions. Exciting or stressful information creating cognitive and/or emotional arousal. Lack of parental or personal boundaries around social media/device use established, or could be indicative of family boundaries more extensively. Difficulty sleeping may lead to increased evening screen use.	Screen time is adversely associated with sleep outcomes: Primarily shortened duration, delayed timing, and reports of excessive daytime somnolence. It is unclear what the causal pathway is and the direction of the relationship between screens and sleep problems.	Youth should be advised to limit or reduce screen time use, particularly during the hours before bed Whether light from screen devices impacts sleep is equivocal. Research into causal pathways needs to be conducted.
Type of day	2 (9;10)	The type of day (work or school day vs. weekend/holiday) impacts sleep. Possible causal pathways: Societal pressures and expectations, e.g., meeting deadlines, joining events at predefined times, and school start times curtailing sleep duration.	The type of day impacts regularizing wake and bed times: Generally, individuals sleep longer on weekends/holidays.	Youth should be advised to regularize their sleep schedule, particularly keeping a regular wake time. Education explaining that: Daytime pressures may impact sleep regularity, and; That school start times are incompatible with young adult circadian drives, May be important information for sleep hygiene in young adults.
Sex	1 (10)	Biological sex or gender identity impact sleep. Possible causal pathways: Hormones. Societal expectations.	Females appear to have greater TST than males.	Information on how gendered norms or biological sex factors impact sleep could be delivered in the course of sleep hygiene. Research into causal pathways needs to be conducted.
Age	2 (10;11)	Age impacts sleep. Possible causal pathways: Biological need for sleep declines with age. Ability to maintain sleep declines with age. Social and work/education demands increase with age. Circadian timing delays in adolescence and young adults.	Sleep appears to decline with age: -14 mins/school day per year of age increase. -7 mins/non-school day per year of age increase.	Youth should be provided information on how age can impact sleep routines. Research into causal pathways needed.
Culture/Geographical location	4 (4;9;10;11)	Cultural practices impact sleep and the view of importance of sleep vs. other activities. Possible causal pathways: Societal norms and expectations, e.g., prioritizing socializing and education over sleep. Geographically different day and night lengths may impact the environmental cues for sleep and wake.	Cultural practices such as norms and expectations, for example around achievement, substance use, impact sleep length, e.g., Turkish adolescents exposed to passive smoking and who drink caffeinated drinks report shorter sleep than peers who abstain. Individuals living in Asiatic countries sleep on average 40-60 min less than individuals living in predominantly European countries.	Youth and parents/housemates should be informed that cultural practices and environmental cues (e.g., amount of sunlight) may impact sleep routines. Research into causal pathways is needed.
Substance use	2 (4;9)	Substance use by self or others in living environment impacts sleep. Possible causal pathways: Substances impact sleep through chemical pathways Substance use may be a marker of prioritizing other activities over sleep, e.g., use of alcohol, coffee, or smoking may indicate a group prioritization of socializing late at night.	Use of substances impacts sleep duration. Use of caffeine is associated with shorter sleep duration. Sleeping in environments where passive smoking occurs shortens sleep duration.	Youth and parents/housemates should be informed that substance use and timing of use can impact sleep duration. Research into causal pathways needs to be conducted.
Family environment and sleep boundaries	2 (7;9)	Establishing sleep boundaries and family routines impacts sleep. Possible causal pathways: Normalization of establishing family/personal boundaries for sleep and sleep routines. Understanding of the importance of sleep and the need for regularity. “Permission” from a perceived expert to establish sleep boundaries for self or family.	Establishing a safe sleep environment, with regularized routines, appears to be associated with better sleep.	Youth should be informed that regularizing sleep time is important for sleep quality. Families should be educated in how to create “safe sleep environments”, establish sleep routines, and set sleep boundaries, e.g., bedtimes, family time together in the evening, and/or limiting technology use in the hours before bed. Research into causal pathways is necessary.
Evening light exposure	1 (9)	Determining if evening light exposure from technological devices impacts sleep. Possible causal pathways: Light in the blue wavelength alters internal clocks and sleep timing.	Evening light exposure does appear to: Delay bedtime, and is; Associated with a shorter sleep duration. It does not appear to be associated with: Delayed sleep onset.	Youth should be informed that while light exposure can impact sleep, it is as yet unknown if the amount and type of light from devices, or the cognitive and emotional arousal of the types of activities, e.g., social media and gaming, impact sleep. Research into causal pathways is necessary.
Exercise and extracurricular activities	1 (9)	Generally increasing the amount and type, and/or timing of exercise to promote greater sleep amount and quality. Possible causal pathways: Physical and cognitive exertion increase demand for sleep through biological (buildup of metabolites) or psychological (sense of satisfaction with ones’ day) means. Exercise may act as an external zeitgeber for sleep and wake timings.	Exercise is: Beneficial for bedtimes, but; Does not appear to impact sleep onset latency or total sleep time.	Youth should be advised that exercise is associated with keeping a regular bedtime. Research into causal pathways is necessary.
Sleep hygiene education programs	2 (4;8)	Education is provided on the importance and purpose of sleep, what healthy sleep looks and feels like, and the average guidelines of “good sleep” for certain age ranges and genders. This also includes definitions, signs, and symptoms of sleep disorders. Possible causal pathways: Information allows an individual to notice areas for improvement, challenge social norms, and/or understand the biological need for sleep.	Sleep education programs yield equivocal results as to their impact on sleep onset and/or duration.	It is unclear if sleep-based education programs are useful: Utility may be increased if school-based programs follow-up with telehealth sessions that individualize the program (as done in brief CBT group interventions), or if school-wide changes are made that match student sleep timings, e.g., delayed-school start times. Mood may be improved through sleep education programs. However, further research is required on how to modify school-based programs to improve effectiveness.
Cognitive behavioral interventions	3 (5;6;7)	Young people are encouraged to identify and challenge unhelpful thoughts, beliefs, and behaviors concerning sleep. Ultimately the aim is to reduce negative emotion and cognitive arousal, and strengthen the association between bed and sleep by restricting the types of activities that can be conducted in bed and standardizing routines and sleep times around bed in order to create a consolidated sleep period. Possible causal pathways: Greater quantity of sleep Better quality of sleep Greater consistency of sleep timing Perceived better sleep	CBT interventions of 2+ face-to-face sessions, with phone or telehealth follow-up, had a positive impact on sleep: Where objective TST did increase (Range: +11.47 to 16.15 min) Where SOL did decrease (−9.31 to 16.15 min) Where objective SE did improve (+2.82%) Typically, individuals overestimated how much their sleep objectively improved, e.g., 16 minutes on actigraphy vs 30 minutes by self-report. Improvements for SE and TST appear to be maintained over time. Not all studies where participants noted improved sleep also recorded improvements in measurable sleep measures, e.g., TST/SOL. Improvements in self- and/or parent-reported sleep quality Where the first 2 sessions were delivered in a face-to-face and group format, the follow-up phone sessions aimed to personalize the general sleep advice provided in the group sessions.	CBT techniques can be useful in a youth population for improving objective sleep amount and self-perceived sleep quality and regularity. Improvements in timing or amount of sleep appear to be small (11–16 minutes), suggesting quality or self-perception of sleep, not quantity, may be more important when examining sleep satisfaction and daytime achievement. Improvements in sleep appear to also result in improvements in mood and anxiety symptoms. Research in causal pathways is necessary.
Sleep opportunity	1 (12)	Young people are provided with at-home sleep education and agreed upon bedtimes that increase the sleep opportunity by 1–1.5 hours. Greater quantity of sleep Sleep quality improves Greater consistency of sleep timing Perceived better sleep Sleep education and rules are established Reduced access to technology at bedtime reduces bedtime procrastination Imposed bedtimes may have brought circadian rhythms in to line with daytime activities.	Behavioral intervention that aims to increase time in bed so as to increase sleep opportunity and, inadvertently, regularize bed/wake times. Screen access was reduced before bed, naps were prevented, and caffeine consumption reduced. In some studies, sleep education was also provided. This led to: Improved self-reported sleep quality but poorer sleep efficiency Decreased excessive daytime sleepiness Increased sleep opportunity and increased sleep duration (26–115 minutes). Sleep education did not appear to impact the outcomes	Regularizing and prescribing bed and wake times so as to increase sleep opportunity appears to be an effective intervention, with and without sleep education. Such an approach may be particularly useful where longer, more costly, more individualized approaches, are not possible or not suitable, such as in the very young or in those with cognitive impairments. Further research into the suitability of this approach for those with comorbidity, including mental health difficulty, needs to be assessed. Data from this review suggest that psychosocial distress did improve. Further study into suitability for telehealth delivery.

Note: * These recommendations and future directions were elicited from the discussion section of the papers included and from discussions of the current authors, and not necessarily substantiated in the review. Further research is required, in particular, research that isolates components of potential change.

Environmental and individual factors

The reviews suggest factors that impact sleep, particular to young people (11–25 years), include electronic media use, the type of day (week/end), biological sex, age, culture and geographical location, substance use, family environment and sleep boundaries, exercise, and evening light exposure.

Electronic media use

Five reviews (Two systematic reviews: 1; 4, and; three meta-analyses: 2; 3; 9) examined the association between screen use, particularly in the evening, and sleep quality, quantity, and disorder symptoms. Reviews provided evidence that young people (11–25 years) who report more screen time before bed also show shortened sleep duration (1; 2; 3; 4 9), delayed sleep timing (1), poorer sleep quality (2) and excessive daytime somnolence (2). Studies that reported shorter sleep time, ranging from 21- to 51-minutes, or reported a total shorter sleep time of 5–10-minutes of sleep per hour of screen time. There were mixed findings for sleep onset latency, with one meta-analysis reporting no effect (9) and another reporting longer latency with more screen time (3). Effect sizes reported were small (r = .10 to .30) and odds ratios for this effect were in the range of 1.33–2.52, in the three meta-analyses (2; 3; 9) on this topic.

There were some differences in the studies reporting on different media. In one meta-analysis (9), total sleep time was negatively associated by computer and phone use but not by video games or television use. This same study found that computer, phone, the Internet, and video-gaming were associated with later bed-times. However, a systematic review (1) and a meta-analysis (3), reported that, no matter the device, television, computer, phone, tablet, and gaming were all associated with a delayed bedtime and shorter total sleep time. A greater risk of self-evaluated poor sleep and excessive daytime somnolence was reported for young people with access to a device when compared to young people without access (2; 3).

The authors proposed that screen-use may be engaging and distract the young person or make them feel apathetic to bed-time, and/or young people engage in phone use during the night, after sleep onset, reducing total sleep time. However, causality has not been established and the relationship may be the other way around or represent something different, i.e., young people may use devices at night because of a different underlying variable, such as waking from pain or ruminative thinking. This makes device use a symptom, not problematic in and of itself.

Biological sex

One systematic review (10) reported on the impact of biological sex on sleep in young people (11–18 years). This review demonstrated that girls, on average, sleep more regardless of the day of the week (11 more minutes on a week day and 29 more minutes on the weekend), than boys.

Age

Two papers, one systematic review (10) and one meta-analysis (11) reported on the impact of age (11–18 years) on sleep outcomes. In both papers, sleep appears to decline with greater age, with a decrease of 14 min on a school day, per year of age increase, and a decrease of 7 min on a non-school day per year of age increase (10). This may result in a total decrease in sleep time of 35–70 min over the high school years (grade 8–12).

Type of day

Two meta-analyses and one systematic review examined the impact of type of day (week/end) on sleep outcomes in young people (11–20-years). One meta-analysis (9) reported on the type of day by examining daily activities, and showed no effect of activities such as homework, work, or extracurricular activities on sleep outcomes. In contrast, when type of day considered week vs. weekend day, one systematic review (10) and one meta-analysis (11) showed that young people spend less total time asleep on a week day compared to the weekend (10; 11) and have earlier morning wake times on week days (11). From the papers, the amount of time spent asleep on weekends was between 29- to 122.3-min greater than for weekdays. However, there were interactions between type of day, biological sex, and age. This resulted in girls spending 29 min more asleep on weekends, and an increasing difference between week and weekend sleep with greater age. This relationship is not straight forward. The review authors posit that greater demands during the week lead to a reduction in sleep time to meet these demands, followed by greater homeostatic sleep pressure on the weekend. Additionally, young people demonstrate a circadian drive for greater eveningness until early adulthood, as such, this relationship may represent a mismatch between the biological circadian drive for young people and societal timing of events during the week, such as school start times, (Randler et al., 2017; Randler, 2011), whereas on weekends, they can sleep in longer, matching their biological circadian rhythm. However, again, this relationship is not straight forward (Randler et al., 2017; Randler, 2011).

Culture/Geographical location

Four papers, two systematic reviews (4; 10) and two meta-analyses (9; 11) reported on the impact of geographical location and/or culture on sleep outcomes in young people (11–20-years). One meta-analysis (11) found that type of day was associated sleep hours, worldwide, with young people going to sleep 2+ hours later on the weekend than on a week day. This same paper showed that European young people sleep, on average, 60-min more than young people from North America or Asian countries, no matter the day of the week, and that individuals living in Asian countries demonstrated the latest bedtimes (also shown in systematic review 4). However, the other meta-analysis (9), which reported on geographical location by longitude and latitude, showed no impact on bedtime, sleep onset latency, or total sleep time (r = −.02 to −.10) (9). Taken together, this may indicate that cultural factors, rather than geographical location, impact sleep habits.

Substance use

One meta-analysis (9) and one systematic review (4) examined substance use, examining, tobacco, alcohol, and caffeine. The meta-analysis found that caffeine use in the evenings (r = −.18) (also supported by the systematic review), tobacco (r = −.18), and alcohol use (r = −.12), were associated with shorter total sleep time and time in bed (r = −.17; r = −.18; r = −.16), but were not associated with sleep latency.

Family environment and sleep boundaries

Two meta-analyses (2; 9) reported on family environment and/or sleep boundaries (“rules”) in young people (6–20 years). Both papers demonstrated that parent set bed-times and restriction of electronic media at bed-times were positively associated with sleep, being associated with longer total sleep time (r = .22) and compared to young people without parent established bedtime boundaries (OR = 1.53–2.27).

Of the environmental and individual factors, family factors and electronic media use appeared to have the most impact on sleep in adolescents, demonstrating the largest effect sizes (.22 and .10–.30, respectively), although these were in the small range (.10–.30). A topic we will expand upon in the Discussion.

Exercise

One meta-analysis (9) reported that exercise was beneficial for bedtimes for athletes (r = .14), showing that athletes fall asleep faster than non-athletes (11–20 years). However, exercise in general, that is, regardless of the level of exercise in which participants engaged, did not impact sleep onset latency or total sleep time. Sleep quality was not commented on, which is an important future direction for reviews.

Evening light exposure

One meta-analysis (9) examined evening light exposure and/or bright-light (9) exposure before bed, finding that evening light exposure was associated with delayed bed-time (r = .17), and shorter total sleep time (r = -.14), but with no effect on sleep onset latency (r = -.03) (11–20 years). The authors proposed that bright light may alter circadian timing, but caution that the types of devices that emit light (e.g., screens and electronic media devices) also frequently have stimulating content, (e.g., social media and gaming).

Interventions for improving sleep among young people

Cognitive behavioral sleep interventions

Three papers (two meta-analyses: 5 & 7, and; one systematic review: 6) examined cognitive behavioral sleep interventions (6–18 years). These reviews suggest that cognitive behavioral based interventions are effective at increasing total sleep time, reducing wake after sleep onset and sleep onset latency, and improving sleep efficiency. The reviews found some disparate and some similar results post intervention. Regarding objective sleep data, the two meta-analyses (5, 7) found that objective sleep onset latency (SOL) improved by 16.15- to 19.48 min and one paper (5) reported that sleep efficiency (SE) improved by 2.82% and that total sleep time (TST) was improved by 11.47 min. Some of the improvements in sleep were subjectively perceived but not detected in objective sleep assessments: At post intervention, one paper (7) found that subjective TST had increased by 29.47 min, another (5) found no impact on subjective TST; both reported improvements in SOL by 9.31 to 21.44 min (5, 7), and; one paper (7) reported improvements in SE by 5.34% and wake after sleep onset (WASO) improved by an effect of d = .59, while the other review (5) found no impact on WASO. Effects were small to large across individual papers within the reviews. Overall, these results are comparative to those found in meta-analysis of adult CBT-I research. (Trauer et al., 2015) Impacts on SOL were maintained over time.

These interventions were effective when delivered in a face-to-face format or via telehealth platforms. These interventions could be delivered individually, parent mediated, and/or in a group setting. In a group setting, follow-up telehealth appointments to individualize treatment were utilized. While most treatments examined here conducted 4 or more 1-h appointments, one study reported effectiveness with just 2, 1-h appointments. Furthermore, results suggest that the programs are well accepted by young people and their families with ratings above 70%. The most helpful components were: sleep scheduling, personalized bedtimes, relaxation, and mindfulness (see Blake et al. (paper 6) for more detail).

Behavioral programs that increase sleep opportunity

One meta-analysis (12) examined behavioral interventions designed to increase sleep opportunity by 1–1.5 h, concluding that such interventions are effective at increasing total sleep time (26–115 min), improving self-rated sleep quality, and decreasing daytime sleepiness (11–25 years). These interventions decreased sleep efficiency; however, this may be a treatment artifact due to the way that sleep efficiency is calculated (Sleep efficiency = hours asleep/hours in bed) and as individuals were required to spend longer in bed to promote greater sleep opportunity, reduced sleep efficiency may be an artifact of the treatment advice rather than a genuine change in sleep quality. Alternative measures of sleep quality, such as objective electroencephalographic measures of percent in sleep states may better capture positive or negative changes in sleep quality in these studies.

These interventions were effective if delivered to the young person alone, or with a parent. All studies were completed at home, in a face-to-face format. The interventions appeared to be effective whether or not they were delivered with or without psychoeducation about sleep.

Sleep hygiene education programs

Two papers, one systematic review (4) and one meta-analysis (8) examined school-based sleep hygiene education programs (13–18 years). Both concluded that, while there appeared to be a small impact of increased total sleep time in the short term, there was no effect on any sleep outcome at follow-up.

Discussion

This paper summarizes the literature on behavioral and environment factors and sleep interventions that impact sleep in young people by conducting a meta-review of systematic reviews and meta-analyses, reporting the balance of evidence from these reviews, and providing a summary of sleep-health information relevant to young people.

What works now and where to from here?

Environmental and individual factors

Individual and environmental factors that appear to impact sleep in young people include electronic media use, the type of day, biological sex, age, culture, substance use, family environment and sleep boundaries, and evening light exposure. Athletes appear to benefit from exercise, with faster sleep onset, however this is a level of exercise the average population do not attain.

However, we advise caution in imposing a black and white interpretation on these findings. Regardless of the environmental or individual factor, effect sizes across the papers were small (.10–.30), and relationships were correlational. It is likely there are alternative, unexamined, or individually difference variables creating the relationships seen (Cain & Gradisar, 2010). For example, electronic media use may in and of itself not impact sleep in young people, and may instead be a symptom of other causal mechanism/s, such as nighttime pain, rumination, high sedentary activity during the day, or poor choice of media content (i.e., opting to watch a horror film vs. engage with digital fidget apps). Instead, the use of electronic media may be a self-soothing mechanism for distress or boredom and removing it without replacement will not address the issues impacting sleep, or could even make sleep problems worse (Richardson et al., 2021). Therapists working in this space must carefully examine the reasons an individual young person is not getting sufficient sleep and the possible coping behaviors they have adopted.

Of the environmental and individual factors, family factors and electronic media have the most impact on sleep in adolescents, demonstrating the largest effect sizes (.22 and .10–.30, respectively), although these also remain in the small range (.10–.30). Electronic media use was associated with a later bed-time and less sleep in total, but not associated with sleep onset latency, which may suggest the engaging nature of the activities on electronic media, rather than light waves impacting biological sleep pressure, is the causal process: a conclusion other authors have also drawn (Bartel & Gradisar, 2017). The sleep promoting environment had parent set bed-times and restriction of electronic media at bed-times and was associated with longer sleep times. This indicates, somewhat unsurprisingly, that these two factors, family boundaries/environment and electronic media may be associated.

However, the relationship between family and electronic media is not as straightforward as parental control of young adult technology use. It may be that parents who establish co-operative and shared family guidelines, rather than rules, around bed-times and electronic media use, also generally have more effective communication with the young people in their homes, indicating that factors such as warmth and communication are more important than control. Indeed, a paper by Richardson et al. (2021), demonstrated that parental control does not predict technology use in young people, and that the relationship between technology and sleep was more complex and bi-directional. It may be that parental warmth and communication can uncover the reasons young people use electronic media at night, to which parents can act as collaborators and find sleep supporting ways to address underlying issues. Indeed, findings in the wider community and specialty mental health settings show that sleep problems in young people are associated with poor family functioning, depressive symptoms, and coping through distraction (Reigstad et al., 2010). A family system approach where parents scaffold the childs’ routines, and invest time with their kids, is possibly an effective way forward; however, the interplay between family environment, electronic media, and family communication needs further investigation. Such information will better guide intervention through understanding causal links, rather than working in a space guided by associations.

This review demonstrated that exercise for non-athletes did not modify sleep onset latency or total sleep time. This finding is not surprising as the wider literature also suggests mixed findings for the impact of exercise on sleep in children and young people (Dolezal et al., 2107). However, the paper examining exercise [9] did not examine reviews identifying changes in percentage of sleep stage or objective assessment of sleep. Research in adults using objective sleep monitoring has demonstrated that those who practice aerobic exercise exhibit greater deep sleep (Aritake-Okada et al., 2109). These findings need to be replicated in young people. Future reviews may examine studies that have taken objective measurements of sleep in young athletes compared to those who engage in more sedentary activity and provide more detail on what type, timing, and amount of exercise may lead to better sleep quality and/or is associated with better daytime functioning (self-report or lab-assessed) (Rosipal et al., 2013). These details remain to be elucidated. As young people have reported wanting strategies to improve sleep quality (Paterson et al., 2019) and as exercise increases the percentage of deep sleep across the night in adults, this area may be fruitful yet.

However, it may not be exercise per se that impacts sleep, but rather sedentary time that has greater impact on sleep variables, a factor not examined in these reviews. Certainly, in other populations, sedentary time plays a significant role in impacting sleep quality and quantity (Yang et al., 2016).

However, it remains that these individual and environmental factors may be altered to assist with better sleep, at the person level (electronic media use, substance use, and evening light exposure), the familial level (family environment and sleep boundaries, e.g., prescribed bed-times, restriction of electronic media before bed, and healthy sleep role modeling), and the societal level (later school start times to suit greater eveningness). We theorize that improved sleep would be longer lasting if approached on all levels, concomitantly; but to the best of our knowledge, no intervention studies have taken this approach. Past research indicates willingness and desire in young people to make healthy sleep changes but difficulty with long-term change (Paterson et al., 2019).

Young people have reported willingness and openness to making positive changes to their sleep, and report that they have attempted strategies including advancing their bedtime and reducing screen time before and in bed (Paterson et al., 2019). Some of the difficulties they report include competing demands on their time and shifting routines, technology use and difficulty switching off.

Finally, sleep traits, to a certain degree appear to be heritable, and some of these environmental and individual factors are largely not modifiable (e.g., sex, culture), as such responses to interventions may be impacted by how much a persons’ sleep traits are influenced by unchangeable inherited factors, or changeable environmental factors (Kocevska et al., 2021). For example, young adults experience biological changes that create both delayed sleep timing and increased time to build homeostatic sleep pressure (Orchard et al., 2020).

From the variety of factors identified in this review, qualitative research with young people, and from health behavior change theories specific to youth (Perry, 1999), it follows that interventions to change sleep behavior will need to be multifaceted, individualized, and flexible. This approach is crucial to address the unique problematic aspects of sleep for each young person.

Interventions for improving sleep among young people

Reviews and meta-analyses examined here support the proposal that cognitive behavioral therapy interventions (CBT) of at least 2 h, in any format (deBruin et al., 2015), or behavioral interventions (not school interventions) are effective in this population. The results from these reviews suggest that school-based sleep hygiene education programs have a small impact on increasing total sleep time in the short-term but do not have long-term efficacy. These results are from 3 reviews (2 meta-analysis; 1 systematic review) that examined 18 papers, pooling data from 1,778 young people of cognitive behavioral interventions, 1 meta-analysis that examined 17 papers, pooling data from 605 young people of behavioral interventions aimed to extend sleep opportunity, and 2 reviews (1 meta-analysis; 1 systematic review) that examined 17 studies, pooling data from 284,491 young people of school-based programs.

The success of the C/BT programs may be due to increased knowledge, as in education programs, and via directly addressing unhelpful behaviors (and cognitions, in the case of CBT). Long term benefits may be gained with the aid of goal setting, motivational interviewing, practice, problem solving and feedback. Furthermore, C/BT approaches are often individualized. Behavior change, individualized medicine, and the opportunity to review challenges and successes with a professional may be key to the effectiveness of these programs. In the school-based education programs, those few studies that also practiced behavior change and reviewed progress appeared to have the most gains, at least in the short-term. These studies were in the minority but provide direction on how best to modify brief interventions for longer-term effect, that may be taken into educational settings.

CBT prevention programs are, in general, longer and more costly than school-based education programs. However, investing in CBT-based programs for young people in schools and universities may also produce the added benefit of reducing somatic complaints and anxiety, and improving physical health and quality of life (Friedrich et al., 2018; Schlarb et al., 2017). Indeed, the research summarized in this review suggests that just 2 × 1-h sessions are needed to see improvements in sleep outcomes (Paavonen et al., 2016). Furthermore, while effects for factors remain small, directionality between factors unclear, and evidence that appears individual young person and family factors may be important, an individualized approach may be best. CBT-I incorporates these individualized treatment factors while also modeling a warm communicative environment.

A summary of sleep recommendations for young people, advocacy suggestions, and future research directions resulting from findings consistent across reviews are presented in Table 2.

Conclusions

Sleep in young people is negatively associated, though with small effect sizes, with electronic media use, the type of week day, biological sex, age, culture, substance use, and evening light exposure, and positively associated with a certain type of family environment and sleep boundaries. Some of these factors interact, particularly sex, age, and week/weekend day. While the direction of the relationship between these factors and sleep is unclear, this may not be important in so far as healthy change is concerned, as use of CBT to impact these behavioral factors creates effective, long-lasting change in sleep for young people.

Data statement

The full set of papers used in this analysis can be found in Table 1, and further detail on these papers can be found in Online Supplement A.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The work for this paper was supported by monies from the Alcohol and Drug Foundation, Local Drug Action Group (CAP#2: 300061).