When tracking makes it worse: orthosomnia and chronic illness

Short answer

Orthosomnia is the name for sleep that gets worse because you are trying to make a tracker's number better. Sleep clinicians coined it in 2017 after seeing patients whose main complaint was their sleep score rather than their sleep, and whose belief that they slept badly was hard to shift even when shown the device was not accurate (1). The pattern reaches well beyond sleep: resting heart rate, heart rate variability, body battery and readiness scores can all turn from information into a verdict you live by.

If you have caught yourself doing this, it is not your fault, and it is not a failure of willpower. Numbers are persuasive by design, and most persuasive when your body has been unreliable and other people have doubted your symptoms. For someone with a chronic illness, a number can feel like the first objective ally they have had. Leaning on it too hard is an understandable response to being disbelieved.

Note: this article is for education and is not medical advice. Insomnia lasting three months or more, loud snoring with pauses in breathing, falling asleep without meaning to during the day, or sleep problems tangled up with low mood or anxiety all deserve assessment by a clinician.

What exactly is orthosomnia?

The term was introduced in 2017 in the Journal of Clinical Sleep Medicine. It is a case series — a write-up of individual patients a clinic actually saw — not a survey, not a trial, not a prevalence study (1). A case series is the weakest kind of evidence for how common something is, and the strongest for noticing that something exists.

The authors described people seeking treatment for sleep problems they had diagnosed themselves from tracker data — stretches of "light" or "restless" sleep on a screen, or a nightly total that looked too small. Their inference, that the data explained their daytime tiredness, became what the paper calls a perfectionistic quest for ideal sleep. And the tracker data felt more true to them than polysomnography, the laboratory recording that is the reference standard (1).

The name joins ortho — correct, proper — to somnia — sleep. A 2023 editorial in Nature and Science of Sleep summarised it as the obsessive pursuit of ideal sleep and listed the behaviours that go with it: preoccupation with tracker data, frequent checking, and anxiety about being separated from the device. It also named the behaviour that does most of the damage — assuming the tracker is perfect, then spending an unusual amount of time in bed trying to improve what it reports. The same editorial confirms the original paper described three patients, and draws a line worth keeping: insomnia is a diagnosable sleep disorder, orthosomnia a societal phenomenon (2).

So orthosomnia is not a formal diagnosis. There are no agreed criteria and no validated scale in routine use. One cross-sectional study of 523 people has tried to put a number on it, with an unvalidated algorithm combining wearable ownership, insomnia symptoms and a questionnaire on anxiety and preoccupation with sleep (3) — a first attempt at measurement, not a prevalence you can quote.

Why would a sleep tracker make sleep worse?

What you are told about your sleep changes your day. Sixty-three people meeting DSM-5 criteria for insomnia disorder wore an actigraphy-diary watch and, at their usual rise time, were randomised to sham feedback: "positive" or "negative" sleep efficiency, regardless of what actually happened. By evening, the thirty-two given negative feedback had worse daytime function than the thirty-one given positive feedback — lower alert cognition (d = 0.79), more sleepiness and fatigue (d = 0.55) (4). Same nights, different number, different day.

An earlier experiment found the same with 22 people with primary insomnia given positive or negative feedback on three consecutive mornings. Actigraphy showed their objective sleep did not differ across nights or conditions, yet negative feedback was followed by more negative thoughts, more sleepiness, more monitoring for sleep-related threat and more safety behaviours during the day (5).

It is not only people with insomnia. Across two studies with 164 participants, people were told they had spent an above-average or below-average proportion of the night in REM sleep — figures assigned at random. Assigned sleep quality, but not their own reported sleep quality, predicted scores on two cognitive tests (6). A made-up percentage moved real performance.

Chasing the number drives the behaviours that cause insomnia. Spending longer in bed to "get more sleep", going to bed before you are sleepy, lying awake, sleeping in to compensate — these are the habits cognitive behavioural therapy for insomnia works to undo, because they weaken the link between bed and sleep. A study recording fourteen days of home sleep with a lumbar-worn actigraph found people with chronic insomnia went to bed earlier and spent significantly longer in bed than volunteers without insomnia, and underestimated their total sleep time most when bedtimes were earlier and time lying down was longer (8). The obvious response to a disappointing score is the one that makes the next score worse.

And the tracker is less certain than it looks. That is the next section, and the part most people have never been told.

What does a sleep tracker actually measure, and what does it estimate?

Consumer devices do not measure sleep. They measure movement, heart rate and sometimes temperature, and an algorithm converts those signals into sleep. Some of that conversion is good; some is a guess printed in the same typeface as a fact.

The most useful evidence is a laboratory study in which 34 healthy young adults slept three consecutive nights — one deliberately disrupted — wired to polysomnography while also wearing research actigraphy and consumer devices. Most did as well as or better than actigraphy at telling sleep from wake. The detail matters: epoch-by-epoch sensitivity was high for every device (all ≥ 0.93), so they are good at calling sleep "sleep"; specificity was low to medium (0.18 to 0.54), so they are much worse at calling wake "wake". Sleep stage results were inconsistent, and every device did worse on the nights with poorer, more disrupted sleep (9). The nights you most want a trustworthy number are the nights the device is least trustworthy.

A World Sleep Society task force recommended in 2025 that manufacturers standardise a set of fundamental sleep measures, separate from the proprietary, exploratory metrics devices also display (10). Your screen does not make that distinction: total sleep time and a sleep score sit side by side, same font, same authority, and are not the same kind of thing.

A score is the least checkable thing on the screen. It is a composite: an algorithm takes heart rate, HRV, movement, sometimes temperature, plus its own estimate of your sleep stages, weights them in a way the manufacturer chose and does not publish, and returns one clean number out of 100. Every step adds uncertainty, and the number hides all of it. Put the error bars back yourself — read your seven-day spread rather than today's value, and keep fundamental measures separate from composites.

Devices also get less reliable the further they travel from the people they were tested on, and that study used 34 healthy young adults (9). With a chronic illness — long periods lying still, fragmented nights, a high resting heart rate — you are outside the group the algorithm was tuned on, and nobody can say by how much.

Why does orthosomnia hit people with chronic illness harder?

Because tracking is not a hobby here. It is a coping tool, sometimes close to a treatment strategy, and the advice that works for everyone else — stop looking — is not available.

In ME/CFS and long COVID, heart rate and HRV are used for pacing: staying under a personal ceiling, or holding back on a day when the morning reading is unusually low. In POTS, resting and standing heart rate are how the condition is monitored — our guide to what a normal resting heart rate means with chronic illness covers reading it against your own baseline, and our guides to heart rate pacing and morning HRV for pacing cover how those readings get used.

The problem is when a number adopted as a tool becomes the authority.

• It turns into a permission slip. "My score is low, so I cannot go" — on a day you felt able. Or worse, the reverse: "My score is fine, so I must be exaggerating."

• It displaces your own judgement. People told for years that their symptoms are not real can end up trusting the device more than the body it is strapped to.

• It adds a daily evaluation to a condition that already costs enormous mental energy.

• The data is noisiest on the days that matter. Fever, pain, a flare, a medication change and alcohol all distort heart rate and HRV. The morning you most want a clear answer is often the morning the signal is least clean.

There is also a specific trap in sleep data here, and it has been measured. Unrefreshing sleep is a core symptom of ME/CFS — yet a meta-analysis of 24 case-control studies with objective recordings (20 adult studies, 426 patients and 375 controls) found differences that are real but modest: longer time in bed, longer to fall asleep, longer awake after falling asleep, lower sleep efficiency (17). Nothing on that list looks like the devastation people describe. Fibromyalgia has the same shape: a meta-analysis of 25 case-control studies with 2,086 participants concluded that sleep difficulties there are larger when reported subjectively than when assessed objectively (18).

If the laboratory reference standard already understates how bad these nights feel, a wrist-worn estimate of it understates them further. A reassuring sleep score is not evidence that you slept well. It is evidence that an algorithm built around movement and heart rate did not detect what is wrong with your sleep.

When a score and a feeling disagree, who is right?

Neither, automatically — because they measure different things, and the research is more interesting than "people are bad at estimating".

A systematic review of the sleep perception literature found that most people estimate their polysomnography-measured sleep duration reasonably accurately. In good sleepers, feeling awake while asleep is the norm at sleep onset, common through the first non-REM cycle, and almost never happens in REM. People with insomnia differ specifically: they consistently underestimate how long they slept, however long that was, and keep feeling awake after the first cycle and during REM. Newer techniques have found metabolic and microstructural EEG changes in these patients that look like a shift toward greater cortical activation during sleep, correlating with feeling awake (16). The authors propose reframing "misperception" as partly mismeasurement — the standard recording may miss wake-like brain activity the sleeper can feel.

That is worth sitting with if you have been told your sleep looks fine. "Fine on the recording" and "restorative" are not the same claim, and the gap is a research question, not a character assessment.

So when the score and the body disagree, do not pick a winner. Write both down and read the pattern over weeks — which is what an activity and symptom diary is for, and what we cover in when your metrics disagree with how you feel.

How do you know if sleep tracking has become a problem?

Signs worth taking seriously:

• You check your sleep or recovery score before you have noticed how you feel.

• A bad score changes your mood or plans even on a day you felt fine.

• A good score makes you doubt yourself on a day you feel unwell.

• You have changed your bedtime, time in bed or naps specifically to improve a number.

• You feel anxious when the device is charging, broken or forgotten — one of the behaviours clinicians associate with orthosomnia (2).

• You spend more time looking at graphs than the graphs could plausibly be worth.

• Your sleep got worse after you started tracking, not better.

Two or three of those, and it is worth changing how you use the data rather than how much — which, for anyone pacing by heart rate, is usually the only change available.

What helps if sleep tracking has started to backfire?

Feel first, check second. Rate how you feel from 1 to 5 before you open anything. The order keeps your own read primary and makes the device a second opinion.

Read weeks, not nights. Single-night scores are noisy and emotional. A seven-day average of total sleep time, time in bed or resting heart rate carries more information and far less drama.

Decide in advance which decisions a number may make. For example: "if my resting heart rate is more than five beats above my normal for two days running, I cancel non-essential plans." A rule set when calm is not a verdict handed down at 7am.

Turn off what you do not act on. If a metric never changes anything you do, hiding it removes a daily evaluation for free.

Take a holiday from the screen, not the device. Keep wearing it for the record; stop opening the app for a week. If you feel better, that is information.

Do not fix a score by spending longer in bed. This is the behaviour the orthosomnia literature points at most directly (2), and home actigraphy found earlier bedtimes and longer lying-down time tracking with worse underestimation of sleep in chronic insomnia (8).

Use the treatment that exists. The American Academy of Sleep Medicine gives multicomponent cognitive behavioural therapy for insomnia (CBT-I) its only STRONG recommendation for chronic insomnia in adults, and suggests clinicians not use sleep hygiene as a stand-alone therapy (15). CBT-I targets the behaviours orthosomnia encourages.

Work through the data with someone rather than avoiding it. In a trial, 48 people with insomnia wore an actigraph and kept a diary for two nights. Half were shown the discrepancy through a hands-on behavioural experiment; half were told about it verbally. The experiment produced far larger reductions in self-reported sleep impairment, insomnia symptoms and sleep-related anxiety and distress (effect sizes 0.79 to 1.25) than verbal feedback (−0.06 to 0.31) (7). Data corrects a misperception when someone walks through it with you; it rarely does when it arrives alone as a grade.

Does this mean sleep tracking is bad for you?

No. The evidence says something more specific: the difference between tracking that helps and tracking that harms is mostly interpretation.

The strongest counterweight is a randomised trial built for this question. 113 adults with significant insomnia symptoms were randomised to five weeks of sensor-based sleep feedback with guidance on interpreting it, or to sleep education and hygiene; both arms had one session and two check-in calls. The feedback group ended with lower insomnia severity (d = 0.51) and less sleep disturbance (d = 0.42), though the gap between self-reported and sensor-measured sleep did not change (11). Guided feedback beat sleep hygiene; it did not reconcile the number with the feeling.

A second randomised study points the same way: 100 university employees were assigned to two weeks of actigraphy plus a 45-minute appointment with sleep data feedback and education, to feedback alone, or to a waiting-list control. Both active groups improved on sleep and well-being measures; the inactive control improved on nothing (12).

A small controlled study designed to see whether daily tracker feedback worsens sleep worry did not find that it did. Twenty-six primary-care patients were split between wearing a consumer tracker for four weeks and keeping a handwritten diary only; sleep quality, sleep reactivity to stress and quality of life improved in both groups, with no significant difference (13). The sample is tiny, but the conclusion is stated carefully: using a wearable does not necessarily make sleep worries worse.

Population data is more mixed. In a nationally representative survey of 1,200 Canadians, 19.3% used a wearable to monitor sleep; close to 45% of those users felt it had a positive effect on their sleep and stress, and 4.5% a negative one. Users also reported shorter sleep and more severe insomnia symptoms than non-users, and wearable use moderated the anxiety–sleep relationship: as anxiety rose, total sleep time fell more steeply in wearable users (14). This is cross-sectional, so it cannot say which way causation runs — a reason to be careful about anxiety and data, not a reason to throw the device away.

Data paired with explanation, context and a purpose tends to help. Data delivered as a daily score, with no guidance, treated as more authoritative than the body it came from — that is where orthosomnia lives.

What does the evidence on orthosomnia not show?

How common it is. There is no validated measure. That one prevalence attempt produced estimates from 3.0% to 14.0% of its sample depending on the cut-off, in a cross-sectional, 81%-female sample (3). The founding paper is a case series of three patients (1, 2).

That owning a tracker worsens sleep on average. The sham-feedback experiments show that what you are told about your sleep changes your next day (4, 5). That is a different claim, and guided feedback improved insomnia severity in two randomised trials (11, 12).

That a tracker can predict a crash, a flare or a bad night. No consumer device has been shown to do this. A record gives you a look back at what the days before something looked like, and associations to raise with your doctor.

Who is at risk. Nothing in the literature identifies who is likely to be harmed by tracking and who is likely to be helped.

Living with sleep tracking and chronic illness: the questions people ask next

I need heart rate data to pace. Can I stop tracking sleep without losing that?

Yes, and for many people with a chronic illness this is the practical answer. Pacing runs on heart rate and, for some people, a morning HRV reading — not on a sleep score. Most apps let you hide individual cards, so you can keep resting heart rate and HRV visible and remove sleep stages, sleep scores and readiness scores from the screen you open each morning. The device keeps recording either way. Removing one daily grade while keeping the measurement you use is the lowest-cost change available.

Can I get through a work day when the sleep score has already ruined my morning?

Often, and it helps to know the score itself is part of what makes the morning hard: in a randomised experiment, people told they had slept badly had lower alert cognition and more fatigue by evening than people told the opposite, on identical nights (4). A practical approach is to check the number after the first task rather than before it, and judge the day on what you can actually do. If colleagues treat an invisible condition as something you are imagining, the record you keep is for you and your clinician — not evidence you owe anyone at work.

Should I keep exercising if my sleep score says I should not?

A sleep score is not an exercise prescription, and the answer depends on your condition, not the number. With ME/CFS or long COVID with post-exertional malaise, NICE's 2021 guideline (NG206) advises against graded exercise therapy and against fixed incremental increases; any activity change stays inside your energy limit and is reversed if symptoms worsen a day or two later. With POTS, fibromyalgia, MS, lupus, rheumatoid arthritis or cancer-related fatigue, structured exercise is evidence-based and recommended by those conditions' own guidelines, and a low sleep score is not a reason to skip a programme your clinician set. Ask which applies to you.

What do I do with my tracker when I travel or have a big event?

Travel and events wreck sleep metrics in ways that say nothing about your health — different bed, different temperature, late meals, alcohol, a wearable charging at the wrong time. Keep wearing the device for the record but stop reading scores for the duration, then look at the whole week once you are home. If you pace by heart rate, keep that one visible, and plan the rest day after the event in advance rather than deciding on the morning from whatever the app says.

Does alcohol really explain a bad sleep score?

Alcohol raises resting heart rate and lowers HRV overnight, and those are two of the inputs most sleep and readiness scores are built from, so a low score the morning after drinking is expected. What the score cannot tell you is how much of that night was alcohol and how much was something else. More useful is logging drinking nights for a few weeks alongside how you felt next day, and reading the pattern rather than the single morning. Alcohol tolerance changes in some chronic conditions — a conversation for your clinician, not your app.

I wake up at 3am and check my score. How do I stop?

The checking is the part to change first, because a number in the night supplies something to be anxious about at the moment you are least able to be reasonable about it. A practical approach is to charge the phone and the watch out of reach of the bed, so the check costs standing up. The deeper fix is cognitive behavioural therapy for insomnia, which the American Academy of Sleep Medicine recommends most strongly for chronic insomnia in adults (15) and which works directly on night-time checking.

How do I explain to my family that the score is not the same as how I feel?

The two measure different things. A tracker measures movement and heart rate and infers sleep from them; it has no access to whether sleep restored you. There is a research basis for this: in ME/CFS, objective recordings show differences from healthy controls that are real but far smaller than the symptom (17), and in fibromyalgia, sleep difficulties are larger when reported subjectively than when measured objectively (18). "The recording looks normal" has never meant "you slept well" — and being doubted on that is exhausting in a way the data cannot show either.

Will I have to track forever?

Not necessarily, and not at the same intensity. Tracking is most valuable in two situations: working out what your limits are, and building a record for an appointment. Between those, a lighter version often does — a weekly glance, or one measurement a day. If your pacing depends on heart rate, that part may stay as long as the condition does. The thing to retire first is the daily score that grades you without changing any decision you make.

How to bring this up with your doctor

Open with one sentence. "I think I have started trying to fix my sleep score rather than my sleep — I have been going to bed earlier and staying in bed longer, and it has got worse." Most sleep clinicians will recognise that immediately.

Bring two weeks, averaged by week. Total sleep time, time in bed, bedtime and wake time as weekly averages, plus how many nights you felt unrested. A clinician can use that; twenty screenshots of sleep-stage charts usually cannot, because stage estimates are the least reliable thing the device produces (9).

Ask these specifically. Would cognitive behavioural therapy for insomnia be appropriate for me? Is there a digital or group CBT-I programme I can be referred to? Should I stop looking at my sleep data during treatment? Do I need testing for sleep apnoea or another sleep disorder?

If you live with a chronic illness, ask about the numbers you pace with. "I use heart rate to pace. Which reading should I base decisions on, and what should override it?" One rule agreed with your clinician beats improvising one every morning.

If you are dismissed, ask: "can you note in my record that I asked about CBT-I for insomnia?" Then ask for a referral to a sleep service or behavioural sleep medicine specialist if one is available.

And do not put these down to tracking anxiety. Loud snoring with witnessed pauses in breathing, falling asleep without meaning to during the day, new chest pain, fainting, or sleep problems alongside thoughts of harming yourself need proper assessment and sometimes urgent care — call 911 in an emergency.

How Welltory helps — and what it cannot do

The limits first. Welltory is a general wellness product, not a medical device. It does not diagnose, predict, monitor, prevent, treat or mitigate insomnia, orthosomnia or any other condition. It cannot tell you whether your sleep restored you, and it cannot tell you in advance that a bad day is coming.

And the obvious point, said plainly: Welltory is itself a tracker. Everything here about sleep scores and daily grades applies to Welltory's numbers exactly as it applies to anyone else's. If opening this app has become a morning verdict rather than a source of information, the fix is the same — change what you look at, or look less often. What a tracker is good for is a long, consistent record measured against your own baseline, so here is how to get that without the daily grade.

1. Decide what you look at, and hide the rest. Resting heart rate, HRV, sleep analysis, stress minutes, recovery and Battery are all available. If a card never changes a decision you make, you do not have to look at it. Rate how you feel before you open anything.

2. Log what happened. On iOS with an Apple Watch or Oura, the Today screen flags stress stretches and asks "What happened?". Tap a suggested tag, type a few words, or just talk: "bad night", "woke at 3", "checked score", "late caffeine", "alcohol", "flare", "pain night", "travel". You can also add a note any time with the plus icon; it goes into your Journal.

3. Look at the one to three days before a bad night. Put sleep analysis, stress minutes, Battery, resting heart rate and your morning HRV reading next to your own notes for the 72 hours beforehand. Stress minutes and sleep analysis need a supported wearable; with the phone camera alone you have your morning spot readings. If the data looks fine and you feel wrecked, believe your body.

4. Check My Patterns after two to three weeks of tagging. My Patterns collects tags you add to stress and rest stretches, so it needs iOS with an Apple Watch or Oura. Patterns appear at around 7 tagged events, and insights typically need at least 7 occurrences of a tag in the current month plus history from the month before. It shows which tagged situations tend to come with stressful stretches, trends by day of week, and every time a tag occurred.

5. Build a log for the appointment. The Journal (Premium) shows HRV measurements, tags, mood, notes and workouts, and you can export a CSV from the web app (Dashboard → choose a chart → Export). Morning readings use the phone camera (PPG) via Heartbeat Report or a compatible wearable; how accurate HRV from a phone camera is covers what it can and cannot do. Any pattern you find is an association to discuss with your doctor — not proof of a cause, and not a warning system.

Where to find other people in the same situation. Welltory runs a paid, moderated community called Energy Lab for women aged 18 to 65 living with energy-limiting conditions — ME/CFS, long COVID, fibromyalgia, POTS, MCAS and similar. It runs alongside the app: you keep collecting your own data, and the Lab is where people learn to read it together, with a medical board answering the science. It is education and peer support, not medical care, it does not replace your clinician, and there is a 14-day money-back guarantee on a first purchase.

How we made it

Made with AI tools, then edited and fact-checked by the Welltory team. See our Editorial & AI policy.

Data analysis by Jane Smorodnikova, co-founder of Welltory and the person who built the methodology behind how we read physiological data.

Written by Tatsiana Yashyna.

Discounts for blog readers: up to 36% off

See what affects your energy, stress, sleep, and daily state with Welltory

Explore your data