Research Letter
Insufficient Sleep Among US Adolescents Across Behavioral Risk Groups
Tanner J. Bommersbach, Mark Olfson, Taeho Greg Rhee
JAMA 2026;335;(13):1173-1176. doi:10.1001/jama.2026.1417
Insufficient sleep among US adolescents is a major public health problem. Although studies documented recent declines in sleep duration across demographic groups,1 it is unknown whether sleep loss has disproportionately affected adolescents with behavioral health risks associated with insufficient sleep, including electronic media use, mental health symptoms, substance use, sedentary activity, or victimization.2 This study examined national trends in insufficient sleep from 2007 to 2023 and whether trends differed by demographic or behavioral risk subgroups.
Methods
We analyzed cross-sectional data from the 2007-2023 national Youth Risk Behavior Survey, a biennial, school-based survey, to produce representative estimates of US high school students. Informed consent was obtained from the student’s parent/guardian. The Centers for Disease Control and Prevention Institutional Review Board approved the Youth Risk Behavior Survey protocol. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
The analytic sample included 120 950 students who answered a question about sleep duration on an average school night. Insufficient sleep was defined as less than or equal to 7 hours per night based on American Academy of Sleep Medicine guidelines3 ; very short sleep was defined as less than or equal to 5 hours per night. Self-reported demographic characteristics were included given known demographic variation in sleep duration.4 Fifteen health risk behaviors associated with insufficient sleep were grouped into the following 5 categories: mental health, victimization, physical/sedentary, weight-related behaviors, and substance use. Additional details are shown in Supplement 1.
Trends in insufficient sleep, stratified by demographic and behavioral risk subgroups, were estimated using mean annual percent change (APC), adjusting for grade, sex, and race and ethnicity, using a modified Poisson model. Interactions between risk factors and time assessed whether trends differed by demographic and behavioral subgroups. Analyses were 2 sided, and P <.05 was considered significant. Analyses used Stata version 19.1 (StataCorp).
Results
The percentage of students reporting insufficient sleep increased from 68.9% in 2007 to 76.8% in 2023 (mean APC, 0.7% [95% CI, 0.6%-0.8%]). This increase was driven by a rise in very short sleep (≤5 hours), which increased from 15.8% to 23.0% (mean APC, 3.1% [95% CI, 2.5%-3.6%]), while the percentage reporting 6 to 7 hours of sleep did not change (Figure).
Figure. Sleep Patterns Among High School Students in the US
Data are from the Youth Risk Behavior Survey. Analyses were adjusted for complex survey designs. Bars represent 95% CIs.
Insufficient sleep (≤7 hours) increased across all demographic subgroups by sex, grade, and race and ethnicity (Table). Increases were significantly larger among non-Hispanic Black students than non-Hispanic White students. Similarly, insufficient sleep increased across nearly all behavioral subgroups, occurring as much or more among adolescents without behavioral risk factors as among those with them. Although insufficient sleep was consistently more common among adolescents reporting behavioral risks, the rate of increase was similar or greater among students without these risks. A few behaviors showed differential trends: insufficient sleep increased more among students reporting depressive symptoms or suicidal thoughts and increased less among students who watched television daily, used social media or video games at least 4 hours per day, or used alcohol or cigarettes (Table).
Table. National Trends of Insufficient Sleep (≤7 Hours) Among High School Students in the US, 2007-2023
| Unweighted sample | Insufficient sleep, % | Mean APC (95% CI), %a | P value | ||
|---|---|---|---|---|---|
| 2007 | 2023 | ||||
| Unweighted sample, No. (weighted %) | 12 073 (9.4) | 17 309 (14.0) | |||
| Total | 120 950 | 68.9 | 76.8 | 0.7 (0.6 to 0.8) | <.001 |
| Grade | |||||
| Freshman | 31 130 | 57.7 | 69.0 | 0.9 (0.7 to 1.1) | <.001 |
| Sophomore | 30 452 | 67.5 | 78.0 | 0.8 (0.6 to 0.9) | <.001 |
| Junior | 30 269 | 75.1 | 79.3 | 0.6 (0.4 to 0.7) | <.001 |
| Senior | 28 660 | 78.2 | 81.8 | 0.5 (0.3 to 0.6) | <.001 |
| Female | 60 398 | 71.3 | 78.3 | 0.6 (0.5 to 0.8) | <.001 |
| Male | 60 122 | 66.6 | 75.4 | 0.8 (0.6 to 0.9) | <.001 |
| Sexual orientation | 2015b | ||||
| Heterosexual | 54 318 | 71.8 | 75.1 | 0.6 (0.3 to 0.8) | <.001 |
| Gay or lesbian | 1904 | 79.3 | 82.6 | 0.3 (−0.6 to 1.1) | .51 |
| Bisexual | 6234 | 76.2 | 83.8 | 0.7 (0.2 to 1.1) | .003 |
| Not sure | 4746 | 79.3 | 81.0 | 0.4 (−0.3 to 1.1) | .23 |
| Race and ethnicityc | |||||
| Non-Hispanic American Indian or Alaska Native | 2234 | 54.3 | 64.2 | 1.3 (0.2 to 2.4) | .03 |
| Non-Hispanic Asian American, Native Hawaiian, or Other Pacific Islander | 5802 | 74.1 | 80.3 | 0.6 (0.3 to 0.9) | <.001 |
| Non-Hispanic Black | 18 190 | 71.1 | 80.7 | 0.9 (0.6 to 1.1) | <.001d |
| Hispanic | 30 284 | 65.5 | 76.3 | 0.8 (0.6 to 1.0) | <.001 |
| Non-Hispanic White | 56 371 | 69.2 | 75.5 | 0.6 (0.5 to 0.8) | <.001 |
| Multiple races | 6178 | 73.5 | 83.1 | 0.9 (0.5 to 1.2) | <.001 |
| Depressive symptoms | |||||
| No | 80 004 | 66.1 | 71.4 | 0.6 (0.4 to 0.7) | <.001 |
| Yes | 39 951 | 76.2 | 84.8 | 0.6 (0.5 to 0.8) | <.001d |
| Suicidal thoughts | |||||
| No | 98 359 | 67.2 | 74.5 | 0.7 (0.5 to 0.8) | <.001 |
| Yes | 21 390 | 79.2 | 87.3 | 0.6 (0.5 to 0.8) | <.001d |
| Threatened with a weapon in the past year | |||||
| Never | 111 150 | 68.4 | 76.3 | 0.7 (0.6 to 0.8) | <.001 |
| 1-3 Times | 6269 | 75.9 | 84.0 | 0.8 (0.5 to 1.1) | <.001 |
| ≥4 Times | 2483 | 72.9 | 79.5 | 0.4 (−0.0 to 0.9) | .10 |
| Missed school due to safety issues in the past month | |||||
| Never | 111 534 | 68.8 | 76.3 | 0.7 (0.6 to 0.8) | <.001 |
| 1-3 Times | 7151 | 70.6 | 80.9 | 0.7 (0.4 to 1.0) | <.001 |
| ≥4 Times | 1952 | 69.5 | 78.8 | 0.3 (−0.2 to 0.8) | .210 |
| Bullied on a school property in the past year | 2009b | ||||
| No | 87 454 | 68.1 | 75.7 | 0.8 (0.7 to 0.9) | <.001 |
| Yes | 20 279 | 73.2 | 81.9 | 0.8 (0.6 to 1.0) | <.001 |
| Bullied electronically in the past year | 2011b | ||||
| No | 78 854 | 67.2 | 75.1 | 0.9 (0.7 to 1.0) | <.001 |
| Yes | 14 353 | 75.4 | 84.9 | 0.8 (0.5 to 1.0) | <.001 |
| Watching television on a daily basis | 2019b | ||||
| Not at all | 14 546 | 67.1 | 81.1 | 1.1 (0.8 to 1.4) | <.001 |
| ≤1 h | 28 743 | 71.1 | 77.2 | 0.5 (0.3 to 0.8) | <.001d |
| 2-3 h | 30 809 | 67.1 | 77.2 | 0.7 (0.4 to 0.9) | <.001d |
| ≥4 h | 14 696 | 69.7 | 76.3 | 0.7 (0.4 to 1.0) | <.001d |
| Social media and video game use on a daily basis | |||||
| Not at all | 17 313 | 65.6 | 69.2 | 0.8 (0.5 to 1.1) | <.001 |
| ≤1 h | 31 081 | 67.2 | 74.7 | 0.6 (0.4 to 0.8) | <.001 |
| 2-3 h | 36 043 | 69.9 | 78.2 | 0.5 (0.4 to 0.7) | <.001 |
| ≥4 h | 31 747 | 76.1 | 78.6 | 0.4 (0.3 to 0.6) | <.001d |
| Physical activity on a weekly basis | |||||
| Not at all | 21 454 | 71.3 | 81.9 | 0.7 (0.6 to 0.9) | <.001 |
| 1-2 d | 21 033 | 70.0 | 80.5 | 0.7 (0.6 to 0.9) | <.001 |
| 3-4 d | 24 786 | 70.1 | 79.0 | 0.8 (0.6 to 0.9) | <.001 |
| ≥5 d | 52 757 | 65.9 | 73.1 | 0.8 (0.6 to 0.9) | <.001 |
| Soda consumption on a weekly basis | |||||
| Never | 26 579 | 69.6 | 75.7 | 0.7 (0.5 to 0.9) | <.001 |
| 1-3 Times | 39 151 | 67.4 | 76.4 | 0.9 (0.7 to 1.1) | <.001 |
| 4-6 Times | 18 720 | 68.1 | 79.0 | 0.9 (0.7 to 1.1) | <.001 |
| Every day | 26 465 | 70.4 | 77.9 | 0.7 (0.5 to 0.8) | <.001 |
| Underweight or normal (BMI <85th percentile) | 76 215 | 69.2 | 76.9 | 0.7 (0.6 to 0.8) | <.001 |
| Overweight (BMI 85th percentile to <95th percentile) | 17 719 | 67.9 | 78.4 | 0.9 (0.7 to 1.0) | <.001 |
| Obese (BMI ≥95th percentile) | 16 788 | 70.0 | 80.5 | 0.8 (0.6 to 1.0) | <.001 |
| Alcohol use in the past month | |||||
| Never | 75 124 | 63.5 | 75.2 | 0.9 (0.8 to 1.1) | <.001 |
| <10 d | 32 346 | 74.4 | 81.2 | 0.6 (0.5 to 0.8) | <.001 |
| ≥10 d | 4069 | 81.3 | 78.7 | 0.1 (−0.3 to 0.4) | .67d |
| Cigarette use in the past month | |||||
| Never | 103 290 | 66.4 | 76.3 | 0.8 (0.7 to 1.0) | <.001 |
| <10 d | 7641 | 75.7 | 85.3 | 0.7 (0.4 to 1.0) | <.001 |
| ≥10 d | 5506 | 80.3 | 83.9 | 0.0 (−0.4 to 0.3) | .93d |
| Cannabis use in the past month | |||||
| Never | 94 587 | 66.9 | 75.1 | 0.8 (0.6 to 0.9) | <.001 |
| <10 Times | 13 930 | 78.1 | 83.9 | 0.6 (0.4 to 0.8) | <.001 |
| ≥10 Times | 10 578 | 74.9 | 86.1 | 0.7 (0.4 to 0.9) | <.001 |
Discussion
These findings show a broad increase in insufficient sleep across all demographic groups, driven largely by increasing reports of very short sleep (≤5 hours). Notably, insufficient sleep increased as much or more among students without behavioral risks, suggesting that structural and environmental factors affecting most adolescents, rather than specific behaviors such as electronic media use, substance use, or sedentary activity, are contributing to widespread sleep loss. These trends highlight the need for population-level rather than targeted interventions. Later school start times, for example, have been associated with longer sleep, greater academic engagement, and improved mental health.5 Despite the public health significance of declining adolescent sleep, intervention research remains limited.
This study has limitations, including that multiple hypotheses were tested without α adjustment; P values should be interpreted with caution. Other limitations include the inability to assess some potential behavioral contributors, such as extracurricular demands and family conflict, the cross-sectional design that precludes causality or directionality of associations, and reliance on single-item measures of behavioral risks and self-reported sleep duration rather than actigraphy. In addition, because adolescents tend to overestimate sleep, true levels of insufficient sleep may be even higher.6
