Smartwatches and Attention: How Wearables Affect Sleep, Anxiety, and Focus

Editorial Note: The people in this piece are composites. Names, details, and timelines have been blended from patterns we see in the research literature and in everyday use, so we can walk through the mechanisms without pointing at any individual. The evidence cited is real and listed in the sources at the end; the characters are not.

Smartwatches and Attention: How a Wrist-Size Screen Took Over Our Interruptions

When Your Sleep Score Wakes Up Before You Do

Mara Ellison is thirty-four, works as a data analyst, and has worn a sleep tracker for two years. The alarm goes off. Before she sits up, before her feet find the floor, she lifts her wrist. Three numbers are waiting: sleep score 68, deep sleep 41 minutes, HRV down roughly ten percent from her seven-day average.

She remembers sleeping through the night. She did not feel wrecked in the afternoon yesterday. None of that survives contact with the 68. The number gets to her mood first, and then she starts reconstructing the night backwards: what time she actually fell asleep, whether she had wine, whether the curtains were closed, whether last week's deadline is still sitting in her system. By the time she stands up, she has spent several minutes auditing something that already happened and cannot be checked.

One detail is worth pausing on. Nobody sent her a notification. Her inbox was quiet. She raised her wrist on her own, opened the app on her own, and started doubting herself on her own. That is the part of the story the usual "notifications are distracting" framing misses.

Smartwatch Notifications vs. Self-Checking: Two Kinds of Interruption

The familiar account of smartwatches and attention goes like this: notifications moved to the wrist. Messages used to require picking up a phone; now they buzz against your skin. That is a real change in friction. Unlocking, lifting, and looking at a screen are the costs that made an interruption expensive, and a vibration on the wrist removes most of them.

There is a second kind of interruption, though, and it is much quieter. Sleep scores, HRV, resting heart rate, blood oxygen: these sit there waiting, and no push notification is required to make a person look. You can turn off alerts. You cannot turn off the urge to check how you slept. This one behaves less like a notification and more like a background process installed in daily life. It takes up no screen time and a great deal of attention.

The two types of interruption pull attention in different directions. An external interruption draws you away from a task and toward the outside world. A self-initiated check redirects attention toward your own body and its reported signals. This distinction is an analytical framework rather than a formal clinical classification, but it helps explain why disabling notifications does not necessarily eliminate wearable-related distraction. The second is harder to notice, because it arrives dressed as self-care, and it carries a built-in reward loop: check once, feel a little less anxious; feel a little less anxious, want to check again sooner.

Mara's drain happens in the morning, inward and silent. Another kind happens during the day, and it is considerably noisier.

What Is Orthosomnia (And Who Does It Actually Affect)

In 2017, Kelly Glazer Baron and colleagues introduced the term orthosomnia in the Journal of Clinical Sleep Medicine, describing three patients who used sleep-tracking devices. All three became preoccupied with their device-reported sleep data and spent excessive time in bed attempting to increase the sleep duration shown by their trackers. The authors noted that they could not determine whether the devices had caused the patients' sleep concerns or whether pre-existing concerns had led them to rely on tracking data. In any case, the effort to improve the numbers may have reinforced behaviors that could worsen insomnia.

The label never became a formal diagnosis. It functions as a clinical observation with a name, useful for describing a pattern rather than for billing an insurance code.

Scale has been explored, cautiously. A 2024 cross-sectional study of 523 adults used a four-part algorithm combining sleep-tracker ownership, insomnia symptoms, anxiety screening, and sleep-related preoccupation. Depending on the threshold applied to the sleep-preoccupation scale, the algorithm identified 3.0 percent, 8.6 percent, or 14.0 percent of participants as potential cases. These were algorithm-based classifications, not clinical diagnoses. The sample was also relatively young, predominantly female, and excluded participants with certain chronic medical or psychiatric conditions and those with severe anxiety scores. The findings therefore offer an initial estimate within this particular sample, not a confirmed prevalence rate for the general population or smartwatch users.

The nature of the data makes the pattern stickier. Wrist-worn consumer devices do not measure sleep stages directly through the brain activity signals used in polysomnography. Instead, they estimate sleep and wake states using movement, heart rate, and related physiological signals. Validation studies comparing these devices with polysomnography have found meaningful discrepancies in several sleep measures, although the direction and magnitude of error vary across devices and metrics. Sleep-stage estimates should be treated particularly cautiously, and a vendor-generated sleep score is an algorithmic summary rather than a standardized clinical measurement.

There is also a cleaner experiment on what a number can do to a person. In a study published in the Journal of Sleep Research, 63 participants meeting DSM-5 criteria for insomnia received sham feedback from an actigraphy device. One group was told they had slept well, another that they had slept poorly, while their actual sleep was statistically indistinguishable. The group told they had slept badly showed measurably worse daytime alertness and higher fatigue and sleepiness that same day. An estimate changed how participants experienced and reported their day. The feedback did not merely describe sleep; it influenced the way they interpreted their waking state.

The Next Buzz: A Day of Interruptions on the Wrist

Dana Whitfield is forty-one, a project manager with two kids, three years into wearing a watch. His morning runs meeting into commute into lunch at his desk; the afternoon runs pickup into dinner into bedtime. The watch has absorbed a lot: group chats, marketing mail, threaded noise. It has also added a lot: sedentary reminders, heart-rate alerts, step-challenge progress, family activity shares.

His own verdict is that he filters better now, and one metric supports him. He picks up his phone less often. Another metric moves the other way. His wrist gets tapped far more often than his phone used to buzz. He is interrupted by any specific message less; he is interrupted by something arriving more.

Later that evening he glances at his activity rings, sees one still open, and goes for an extra lap around the block. Nothing about his day made him more relaxed. The completion metric simply moved from his job onto his body.

Why Wrist Interruptions Cost More Than Phone Interruptions

Phones and smartwatches create different patterns of attention capture. Phones often draw users into longer periods of screen engagement, while wrist-based notifications can make brief interruptions easier to initiate and repeat. The distinction is not absolute: both devices can produce both kinds of distraction. But a smartwatch can impose a less visible cost through repeated attention shifts. A single glance may take only seconds, yet returning to a complex task requires reconstructing the context that was interrupted.

A driving-simulator experiment involving 31 participants offers a more direct comparison: gaze-tracking results showed greater distraction in the smartwatch notification condition than in the mobile-phone condition. The finding is specific to driving, not workplace productivity, but it illustrates how wrist-based notifications can redirect attention even when the interaction is brief.

A single vibration is cheap. The cost is in repetition and in what follows each interruption, the re-entry period where you rebuild the thread you were holding. Raise the frequency and those recovery periods stack up. By the end of the day the genuinely continuous blocks are short and scattered, and the felt experience is just "I could not focus today," which is hard to trace back to forty vibrations you barely registered.

Gamified metrics extend the same logic. Step challenges, corporate wellness points, insurance discounts, family leaderboards: these convert bodily states into graded performance. The clinical case reports described people changing behavior to improve a number. Here that behavior change has infrastructure around it. The question quietly stops being "did I sleep well" and becomes "does my number look good."

Same device, same category of alert, and for a third person the meaning flips entirely.

Smartwatch Health Benefits vs. Health Anxiety: What the Evidence Shows

Hal Brennan is sixty-eight, has hypertension, and carries an elevated stroke-risk profile. His doctor suggested a period of rhythm monitoring. For Hal, an irregular-rhythm notification is a different kind of event than it is for Mara or Dana.

The Dutch EQUAL randomized trial assigned 437 adults aged sixty-five and older with elevated stroke risk to six months of monitoring with a watch offering photoplethysmography and single-lead ECG, or to usual care. New-onset atrial fibrillation was detected in 9.6 percent of the monitored group versus 2.3 percent in usual care, a hazard ratio of 4.40, with a number needed to screen of fourteen. A substantial share of those detected had no symptoms beforehand; they surfaced because a device was watching.

The other side of the same coin has data too. In a 2024 study of 172 patients with atrial fibrillation published in JAHA, wearable users scored higher on symptom monitoring and cardiac preoccupation and reported more treatment worry. Among wearable users, 20 percent reported intense fear and anxiety in response to irregular-rhythm notifications, and a similar proportion said they always contacted their doctor after receiving a notification for possible atrial fibrillation. Separately, 15 percent reported always feeling anxious, scared, or concerned in response to automated high- or low-heart-rate alerts. Atrial fibrillation-related healthcare use was also higher among users. Because this was a retrospective observational study, these associations do not establish that wearable use caused the higher levels of concern or healthcare utilization.

Put both results on the table and the picture stops being one-sided. These devices can detect previously unrecognized conditions, while their alerts may also be associated with reassurance-seeking and renewed fear in some users. The dividing line runs through how the data gets used. Treated as a trend, it is a tool. Treated as a verdict, it becomes a fresh source of anxiety. Where that line sits depends on who is wearing the watch.

Who Is at Risk From Wearable Health Data (Four Profiles)

Start with a rough self-assessment across four dimensions. These profiles are overlapping patterns rather than mutually exclusive personality types or clinical categories. When a number looks bad, does your mood go with it? How many times a day do you open the health app on purpose? Has a clinician asked you to monitor heart rate, rhythm, oxygen, or sleep? Have you ever changed your behavior to make a number better, staying in bed longer, adding a workout, adjusting the room before sleep?

High on the first two points toward the anxiety-sensitive side. A yes to the third points toward chronic-condition monitoring. A yes to the fourth points toward the KPI side. These are not lifelong labels. A new diagnosis, a new job, a training block, or a pregnancy moves people between cells.

Profile A: low anxiety sensitivity, no chronic condition. This group shrugs numbers off, so orthosomnia risk is low and the risk sits elsewhere. The watch has degraded into a second notification channel and into a liability waiver: the score was high last night, so staying up is fine; the app says seven hours, so the afternoon grogginess must be imagined. The fix is to get the data off the wrist. Clear the complications, keep only notifications a human sent, and turn off algorithmic insights, streaks, and badges. The test is simple: did a person send this, or did a model? Decisions get made on daytime function, not on a single night.

Profile B: anxiety-sensitive or perfectionistic, no chronic condition. This is orthosomnia's home turf, and "just don't overthink it" fails, because the problem is at the input. Change the switches once so you are not relying on willpower afterward: disable sleep-score and sleep-stage notifications, hide sleep and stress complications from the watch face, turn off achievements and leaderboards, turn off non-essential threshold alerts. Set a fixed data window, once a week for trends, and do not open the sleep app in the first hour after waking; score your daytime state first. Swap the metric being graded: replace the sleep score with a reasonably consistent wake time, plus three daytime questions about energy, mood, and focus. A two-week no-score reset helps: log only wake time and daytime state, then evaluate on daytime function. Red lines are explicit. If anxiety about sleep data becomes persistent, begins to shape your behavior, or interferes with daytime functioning, consider seeking a professional assessment. For people with chronic insomnia, cognitive behavioral therapy for insomnia (CBT-I) is an established treatment option. Its sleep-restriction component is not appropriate for everyone and should be used under professional guidance.

Profile C: existing chronic condition under monitoring. This group gets hurt most easily by Profile B's advice, because "turn off every threshold alert" can genuinely drop a meaningful signal. The principle is to reduce noise without going dark, and to convert every alert into a pre-agreed action. Walk through Hal's chain: an irregular-rhythm notification arrives. He does not self-diagnose or change medication based on the alert. If his device supports ECG recording and he is well enough to use it, he can save a tracing for clinical review, but this does not replace medical assessment. He follows the monitoring and follow-up plan agreed with his clinician, who can determine whether a standard twelve-lead ECG or ambulatory monitoring is needed. The notification strategy should preserve medically relevant alerts while reducing repetition and unrelated health prompts. Rather than inventing universal thresholds for when to seek care, patients should have a written action plan covering which alerts require routine follow-up, which symptoms warrant prompt medical advice, and which situations require emergency care. A monthly trend export may help organize information for a scheduled appointment, but it should never replace timely evaluation when symptoms or clinically significant alerts arise. Hard red lines: fainting, persistent or severe chest pain, significant shortness of breath, or sudden neurological symptoms such as facial drooping, weakness, or slurred speech require emergency medical attention. A normal smartwatch reading never rules out an emergency.

Profile D: uses the device as a KPI tool. Not necessarily anxious, most prone to goal displacement. Swap outcome metrics for behavior metrics: stop grading deep-sleep minutes and sleep scores, start grading a fixed wake time, workouts per week, screens off an hour before bed. Turn off leaderboards and team comparisons; where the device plugs into wellness points or insurance discounts, ask for monitoring without ranking. Write the plateau rules in advance: no extra training volume because of one day's data, no cancelling the day's plan because "last night was bad," no repaying a late night with extra time in bed. Once every four to eight weeks, go a few days without wearing it, or wear it without looking, and see whether behavior still holds up without the numbers propping it up.

The switch Profile B turns off is the switch Profile C keeps on. Same watch, same heart-rate alert: noise for one person, signal for another. Any single universal guide to "healthy wearable use" fails here, which is the whole reason for splitting the population in the first place.

When a Family Member's Health Data Becomes Your Anxiety

One additional layer deserves attention: health-data sharing can extend the burden of monitoring across a household. Adult children may follow a parent's heart-rate or sleep data remotely, while parents may receive stress or activity alerts from a child's profile. An alert sent to one person can prompt concern in another, potentially turning individual health monitoring into a shared source of vigilance. The extent of this effect will vary across families, but the possibility matters when deciding how much data to share and how frequently others should be notified.

Three practical moves. Let the person whose data it is keep the alert log, with family holding a backup and coming to appointments. Respond to the emotion first and stay out of symptom detail; asking "what was your heart rate just now" repeatedly feeds the checking behavior. Narrow the sharing scope, disable real-time pushes where possible, and keep review asynchronous.

A Smartwatch Boundary Audit You Can Run This Week

This audit uses only the three patterns above and no extra questionnaire.

Step one, locate yourself. Is your drain more like Mara's morning, where nobody pushed anything and you looked anyway? More like Dana's afternoon, where external taps keep arriving? Or more like Hal's, where there is a medical indication and the alert carries a job? Two cells at once is common.

Step two, three questions. Who initiated the last check, the device or you? Are you treating this number as a trend or as a verdict? Have you changed a behavior to make a number look better?

Step three, take one action card. Morning type: hide the sleep score for two weeks, log only wake time and daytime state, then evaluate on energy, emotional reactivity, and focus. Afternoon type: cut notifications down to human-sent only, clear the watch face, keep the watch on for commuting and cooking, take it off for the first half hour after waking and the hour before bed. Medical type: tape an if/then card to the bedside stating when to record, when to call the clinic, when to go to emergency care, and agree on a monthly trend export for the appointment. KPI type: turn off leaderboards, replace graded outcomes with three completable behavior metrics, and schedule an offline check in four to eight weeks.

The point of the audit is to put the data back in the tool slot. None of the four cells requires taking the watch off for good. What changes is its relationship to you.

Keeping the Data in Its Place

None of the three threw the device away. Mara moved what she grades to wake time; the score still exists, it just no longer decides her morning. Dana cut his notifications down to messages and calls from people, reducing the number of wrist interruptions while keeping his kids' calls available. Hal kept his rhythm alerts, but now he has a card: when the alert arrives, run the action first, then deal with the feeling.

The hardware stayed the same. What changed is the relationship between the number and the person. Where your attention goes depends partly on whether you treat the data as a verdict or as a reference.

Read More

• Why Your Feed Looks Different: The Algorithmic Gaze & Gendered Feeds Explained

• The Six Doors Between Seeing and Doing: Why You Forget What You Read

• The Two Ledgers of a Scroll: What Free Feeds Actually Cost

References

1. Baron, K. G., Abbott, S., Jao, N., Manalo, N., & Mullen, R. (2017). Orthosomnia: Are some patients taking the quantified self too far? Journal of Clinical Sleep Medicine, 13(2), 351–354. https://doi.org/10.5664/jcsm.6472

2. Jahrami, H., et al. (2024). Prevalence of Orthosomnia in a General Population Sample: A Cross-Sectional Study. Brain Sciences, 14(11), 1123. https://doi.org/10.3390/brainsci14111123

3. Gavriloff, D., Sheaves, B., Juss, A., Espie, C. A., Miller, C. B., & Kyle, S. D. (2018). Sham sleep feedback delivered via actigraphy biases daytime symptom reports in people with insomnia: Implications for insomnia disorder and wearable devices. Journal of Sleep Research, 27(6), e12726. https://doi.org/10.1111/jsr.12726

4. van Steijn, N. J., et al. (2026). Enhanced Detection and Prompt Diagnosis of Atrial Fibrillation Using Apple Watch: A Randomized Controlled Trial. Journal of the American College of Cardiology. https://doi.org/10.1016/j.jacc.2025.11.032

5. Rosman, L., et al. (2024). Wearable Devices, Health Care Use, and Psychological Well-Being in Patients With Atrial Fibrillation. Journal of the American Heart Association, 13(15), e033750. https://doi.org/10.1161/JAHA.123.033750

6. Lee, Y. J., Lee, J. Y., Cho, J. H., Kang, Y. J., & Choi, J. H. (2025). Performance of consumer wrist-worn sleep tracking devices compared to polysomnography: A meta-analysis. Journal of Clinical Sleep Medicine, 21(3), 573–582. https://doi.org/10.5664/jcsm.11460

7. Brodeur, M., Ruer, P., Léger, P.-M., & Sénécal, S. (2021). Smartwatches are more distracting than mobile phones while driving: Results from an experimental study. Accident Analysis & Prevention, 149, 105846. https://doi.org/10.1016/j.aap.2020.105846