The Gale readResearch

Phone use after 10 pm: 0.43 minutes of teen sleep lost per minute

A phone-sensor and wearable study of 425 teenagers over 7,591 nights found that heavier-than-usual use after 10 pm went with later sleep and less of it, and says how far to trust that.

By Gale Staff · October 22, 2026 · JAMA Pediatrics

Not reviewed by a clinician.

Illustration: Kids, teens & screens

The short answer

In 425 US teenagers followed with phone sensors and a wrist-worn sleep tracker, nights with more phone use after 10 pm than the teenager's own usual were associated with later sleep onset and shorter sleep: each extra minute above their own average went with sleep starting 0.87 minutes later (95% CI 0.84 to 0.90) and total sleep time 0.43 minutes shorter (95% CI 0.39 to 0.47 shorter). This is an observational cohort of Android users, so it shows a pattern across nights rather than proof that the phone caused the lost sleep, and the abstract does not rule out that a teenager who is already awake simply reaches for the phone.

Who to see

Pediatrician

Nights with more phone use after 10 pm went with later, shorter sleep in teenagers; a pediatrician can talk through your teen's sleep.

Find care →

The question that arrives after the headline

It is a quarter past eleven and the light under a bedroom door is the wrong colour, the cold blue of a screen. A parent knows the familiar story, that phones keep teenagers up, and also knows a harder version of it: the teenager who is awake anyway, scrolling because sleep will not come. The question that follows any headline on this subject is which way the arrow points, and how much sleep is actually at stake on a given night.

A paper in JAMA Pediatrics, published online on 19 October 2026 by Yihong Zhao, Lauren Hale and Dimitri A. Christakis, goes at that question with a design that most earlier work on phones and sleep did not have. The abstract opens by noting that smartphone use has been linked to poor sleep, but that the evidence mostly relies on self-reported, cross-sectional data, meaning a single snapshot in time rather than repeated measurement. This study replaces both of those with measurement taken night after night.

What the researchers did

The data come from the Adolescent Brain Cognitive Development Study, a large US cohort of young people. The authors drew on participants who used Android phones and who had at least 14 days of overlapping data from two sources: an application that passively logged smartphone use, and a Fitbit that measured sleep. Data were collected between 30 November 2020 and 5 February 2024. The analysis, according to the abstract, ran from 2 February to 1 June 2026.

The final sample was 425 teenagers, with a mean age of 14.8 years (standard deviation 1.1), 216 of them female (50.8%). Together they contributed 7,591 person-nights. The exposure was how much the phone was used after each hourly cutoff from 8 pm to midnight, and separately in the 5 to 30 minutes before sleep began. The outcomes were Fitbit-derived: total sleep time, sleep onset, sleep offset, and time spent awake after falling asleep.

The statistical move that gives the study its value is called within-between decomposition in multilevel models. In plain terms, it separates two different questions. One is whether teenagers who use phones heavily in general sleep differently from those who use them lightly, which could reflect anything from household routines to temperament. The other is whether the same teenager sleeps differently on a night when phone use runs above his or her own usual. The second comparison holds the person fixed, which removes a great deal of what makes two teenagers different.

What they found

On nights when teenagers used their phones more than their own evening average after 10 pm, they slept a mean of 20.05 minutes less (standard deviation 49.42) and fell asleep a mean of about 64.73 minutes later (standard deviation 50.95). The standard deviations are large relative to the means, which says that nights varied widely, and that the averages describe a tendency across thousands of nights rather than what any one night looked like.

The adjusted within-person estimates are smaller per unit and easier to read. Each additional minute of use after 10 pm above a teenager's own average was associated with sleep onset 0.87 minutes later (95% confidence interval 0.84 to 0.90), sleep offset 0.37 minutes later (0.33 to 0.41), and a change in total sleep time of minus 0.43 minutes (minus 0.47 to minus 0.39). All three had P values below .001. If that relationship were a straight line, which the abstract does not state, 30 extra minutes would work out to roughly 26 minutes of delay in falling asleep and 13 minutes of lost sleep. That figure is this page's arithmetic, not a result from the paper.

Two further results are worth carrying. The estimates grew steadily larger as the cutoff moved from 8 pm toward midnight. And use in the final minutes before sleep was linked to more wakefulness after sleep began: 0.74 additional minutes awake for use in the final 5 minutes (95% CI 0.50 to 0.98, P below .001), a link the abstract calls negligible when the window was the final 20 minutes. That extra wakefulness was concentrated in the first 30 minutes after sleep onset.

The caveat that does the most work

The paper is an observational cohort study, and its own conclusion is worded as an association: greater late-evening use was associated with later onset, later offset and shorter sleep. The within-person design is strong against one kind of confounding, the fixed differences between people. It cannot settle reverse causation. A teenager who cannot fall asleep may pick up the phone because of that, and the data would look the same as a teenager who stays up because of the phone. The design cannot tell the two apart from the sensor record alone.

There are also limits in who was measured. Only Android users were included, so the findings do not speak for other phones, and 425 teenagers with at least two weeks of usable paired data may differ from the cohort at large. The abstract measures use in minutes after a clock time; it does not say what was being done on the phone, whether messaging, video or homework, and the sleep measures come from a consumer wearable rather than a sleep laboratory.

Why it matters, and what the authors conclude

The practical value is in the size and the shape. A relationship of under one minute of sleep timing per minute of use is modest on a single night and could add up across a school week, but the abstract does not test that accumulation. The rising estimates across later cutoffs and the shape of the final-five-minutes result suggest that timing matters, not only total screen time.

The authors conclude that reducing evening phone use, particularly after 10 pm and right before bedtime, may help improve sleep health in adolescents. That is a hypothesis for an intervention study rather than something a cohort of this kind can show. The PubMed record lists conflict-of-interest disclosures for two of the authors that include expert roles in social media litigation; the abstract does not name a funder for the study, and none is supplied here.

What this study can't tell you

  • Whether late phone use causes the lost sleep: the design is observational and the abstract does not rule out that teenagers who are already awake pick up the phone, so the arrow may point either way.
  • Whether the findings hold for iPhone users or for teenagers outside this cohort: only Android users with at least 14 days of paired data were included, 425 young people in all.
  • What teenagers were doing on the phone: the exposure is logged minutes of use after a clock time, and the abstract gives no breakdown by activity such as messaging, video or schoolwork.
  • How well a wrist-worn tracker matches true sleep: sleep onset, offset and wakefulness are Fitbit-derived estimates, and the abstract reports no comparison against a laboratory sleep study.
  • Whether a straight-line reading of 0.43 minutes per minute is right: the abstract reports per-minute estimates and says they grew across cutoffs, but does not say that the relationship is linear at larger amounts of use.
  • Whether cutting evening phone use would improve sleep: no intervention was tested, so the authors' suggestion that reduction may help is a prediction and not a finding.
  • Who paid for the work: the abstract and the PubMed record read for this story do not name a funder.

The Gale read

This paper is worth reading for its measurement more than its headline. Replacing questionnaires with phone sensors and a wearable, and comparing each teenager with himself or herself, answers a real weakness in earlier work on phones and sleep, and the estimates are tight enough that chance is an unlikely explanation. The honest limit sits where the design stops: a within-person association across nights cannot say whether the phone shortened the sleep or the sleeplessness drew the phone. The authors word their conclusion as an association and a possibility, and a reader can do the same. Gale Staff read the abstract through Crossref and PubMed, not the full text, so details such as the covariates adjusted for, how a night was defined, and any sensitivity analyses are not covered here.

Common questions

does phone use at night make teens sleep less

In this study, nights with more phone use after 10 pm than a teenager's own usual went with less sleep: each extra minute was associated with total sleep time 0.43 minutes shorter (95% CI 0.39 to 0.47 shorter) and sleep onset 0.87 minutes later. The study was observational, so it shows an association across 7,591 nights in 425 teenagers and does not prove the phone was the cause.

how much does late phone use delay teenage sleep

On nights with above-average use after 10 pm, teenagers fell asleep a mean of about 64.73 minutes later (standard deviation 50.95) and slept a mean of 20.05 minutes less (standard deviation 49.42). Per minute of extra use the estimate was 0.87 minutes of delay in sleep onset. The large standard deviations mean individual nights varied widely.

is phone use right before bed worse for sleep than earlier in the evening

The study looked at two things. Cumulative use after hourly cutoffs from 8 pm to midnight showed estimates that grew steadily toward the later cutoffs. Use in the final 5 minutes before sleep was linked to 0.74 additional minutes awake after falling asleep (95% CI 0.50 to 0.98), while the abstract calls the link negligible for the final 20 minutes.

were iPhone users included in the teen sleep and phone study

No. Participants were Android users with at least 14 days of paired phone-use and Fitbit sleep data. The abstract gives no reason for that restriction and does not say how results might differ on other phones, so it is not known whether the findings carry over.

Sources

  1. 1.Zhao Y, Hale L, Christakis DA. Evening Phone Use and Sleep Outcomes Among Youth. JAMA Pediatrics. 2026 (published online 19 October 2026), e264902. doi:10.1001/jamapediatrics.2026.4902 link

Read the study →

1 sources, numbered by first appearance. General health information, not medical advice. AI-assisted editorial content — every citation independently verified. Editorial policy

More from Gale Staff

The Latest Health Research →

Published October 22, 2026 · Gale Staff

All Gale Staff analyses →