Pediatrics · Medical Literature
Insufficient Sleep Among US Adolescents
Abstract
To the Editor A recent Research Letter examined trends in adolescent sleep using biennial Youth Risk Behavior Survey (YRBS) data from 2007 to 2023. For most of that period, the relevant YRBS question items are nearly identical, with 1 notable exception. The screen time question in the 2023 YRBS is substantially different from the screen time questions used in prior years. From 2007 through 2019, the YRBS asked “On the average school day, how many hours do you play video or computer games or use a computer for something that is not school work? (Include activities such as Nintendo, Game Boy, PlayStation, Xbox, computer games, and the internet).” The examples provided in parentheses changed over time, with later iterations of the YRBS citing examples such as social media, texting, and tablet use. However, the underlying construct remained largely unchanged. In contrast, the only screen-related question item in the 2023 YRBS was “How often do you use social media?” In this case, respondents were asked to give a frequency-based response (eg, “A few times a month,” “Several times a day,” etc) rather than an estimate of hours.
Clinical relevance
This letter to the editor raises a critical methodological concern about the comparability of screen time data used in a JAMA Research Letter (Bommersbach, Olfson, Rhee, 2026) examining insufficient sleep trends among U.S. adolescents from 2007–2023 via the Youth Risk Behavior Survey (YRBS). The affected population encompasses all U.S. high school students — a group in whom approximately 75% already report insufficient sleep (<8 hours/night) as of 2023. Clinicians and researchers relying on YRBS-derived screen time trends to counsel patients or design interventions must understand that the 2023 screen time variable is not directly comparable to prior years, potentially confounding conclusions about screen use as a driver of adolescent sleep loss.
Key findings
1. YRBS screen time measurement changed substantially in 2023: from 2007–2019, the question assessed hours/day of recreational video/computer use on school days; the 2023 YRBS replaced this with a frequency-based social media use question ('How often do you use social media?'), which is neither conceptually nor metrically equivalent. 2. The response format shift is fundamental: prior years elicited a quantitative hours-per-day estimate; the 2023 instrument elicits an ordinal frequency response (e.g., 'A few times a month,' 'Several times a day'), precluding direct longitudinal comparison of screen exposure dose. 3. The underlying construct changed: earlier iterations measured broad recreational screen time (gaming, internet, social media combined); 2023 measured only social media frequency, omitting other screen modalities entirely. 4. The original Research Letter (Bommersbach et al., JAMA 2026) analyzed ~121,000 U.S. high school students and found insufficient sleep (defined as <8 hours/night) rose approximately 8 percentage points from 2007 to 2023, reaching ~75% prevalence. 5. The Research Letter reported increases in insufficient sleep were broad and not confined to high behavioral-risk subgroups; notably, students with minimal screen time showed equal or greater increases in insufficient sleep than heavy screen users. 6. No specific effect sizes, confidence intervals, p-values, or NNT/NNH for the screen time–sleep association are reported in this letter (a methodological critique), as the letter's purpose is to flag measurement non-comparability rather than present new quantitative data.
Practice implications
Clinicians should exercise caution when citing YRBS-derived screen time trends as evidence that screen use specifically drives the contemporary rise in adolescent sleep insufficiency, given that the 2023 YRBS screen time variable is not longitudinally comparable to prior survey years. The finding that insufficient sleep increased broadly across all demographic and behavioral subgroups — including low-screen-time youth — suggests that structural and environmental factors (e.g., early school start times, academic and extracurricular burden, reduced parental monitoring) may be primary contributors, warranting system-level interventions such as advocating for delayed school start times rather than screen counseling alone. Future epidemiological studies and clinical quality-improvement initiatives measuring adolescent screen time should use validated, standardized instruments that capture both duration and modality of screen use consistently across time points to enable meaningful trend analysis.